USRE49105E1 - Self-calibrated, remote imaging and data processing system - Google Patents

Abstract

An imaging sensor system, having a view of a target area comprising: a rigid mount unit having at least two imaging sensors disposed within the mount unit, wherein a first imaging and a second imaging sensor each has a focal axis passing through an aperture in the mount unit, wherein the first imaging sensor generates a first image area comprising a first data array of pixels and the second imaging sensor generates a second image area comprising a second data array of pixels, wherein the first and second imaging sensors are offset to have a first image overlap area in the target area, wherein the first sensors image data bisects the second sensors image data in the first image overlap area.

Description

CROSS-REFERENCE TO RELATED APPLICATION

This application is a reissue of U.S. patent application Ser. No. 15/200,883 (now issued as U.S. Pat. No. 9,797,980), filed on Jul. 1, 2016, which was a continuation of U.S. patent application Ser. No. 13/772,994 (to issue now issued as U.S. Pat. No. 9,389,298), filed on Feb. 21, 2013, which was a continuation of U.S. patent application Ser. No. 12/798,8999 12/798,899 (now issued as U.S. Pat. No. 8,483,960), filed on Apr. 13, 2010, which was a continuation-in-part of U.S. patent application Ser. No. 11/581,235 (now issued as U.S. Pat. No. 7,725,258), filed on Oct. 11, 2006, which was a continuation-in-part of and claimed priority to U.S. patent application Ser. No. 10/664,737 (now issued as U.S. Pat. No. 7,127,348), filed on Sep. 18, 2003, which claimed priority to U.S. Provisional Patent Application Ser. No. 60/412,504, filed on Sep. 20, 2002 for “Vehicle Based Data Collection and Processing System.”

TECHNICAL FIELD OF THE INVENTION

The present invention relates, generally, to the field of remote imaging techniques and, more particularly, to a system for rendering high-resolution, high accuracy, low distortion digital images over very large fields of view.

BACKGROUND OF THE INVENTION

Remote sensing and imaging are broad-based technologies having a number of diverse and extremely important practical applications—such as geological mapping and analysis, and meteorological forecasting. Aerial and satellite-based photography and imaging are especially useful remote imaging techniques that have, over recent years, become heavily reliant on the collection and processing of data for digital images, including spectral, spatial, elevation, and vehicle location and orientation parameters. Spatial data—characterizing real estate improvements and locations, roads and highways, environmental hazards and conditions, utilities infrastructures (e.g., phone lines, pipelines), and geophysical features—can now be collected, processed, and communicated in a digital format to conveniently provide highly accurate mapping and surveillance data for various applications (e.g., dynamic GPS mapping). Elevation data may be used to improve the overall system's spatial and positional accuracy and may be acquired from either existing Digital Elevation Model (DEM) data sets or collected with the spectral sensor data from an active, radiation measuring Doppler based devices, or passive, stereographic calculations.

Major challenges facing remote sensing and imaging applications are spatial resolution and spectral fidelity. Photographic issues, such as spherical aberrations, astigmatism, field curvature, distortion, and chromatic aberrations are well-known problems that must be dealt with in any sensor/imaging application. Certain applications require very high image resolution—often with tolerances of inches. Depending upon the particular system used (e.g., aircraft, satellite, or space vehicle), an actual digital imaging device may be located anywhere from several feet to miles from its target, resulting in a very large scale factor. Providing images with very large scale factors, that also have resolution tolerances of inches, poses a challenge to even the most robust imaging system. Thus, conventional systems usually must make some trade-off between resolution quality and the size of a target area that can be imaged. If the system is designed to provide high-resolution digital images, then the field of view (FOV) of the imaging device is typically small. If the system provides a larger FOV, then usually the resolution of the spectral and spatial data is decreased and distortions are increased.

Ortho-imaging is an approach that has been used in an attempt to address this problem. In general, ortho-imaging renders a composite image of a target by compiling varying sub-images of the target. Typically, in aerial imaging applications, a digital imaging device that has a finite range and resolution records images of fixed subsections of a target area sequentially. Those images are then aligned according to some sequence to render a composite of a target area.

Often, such rendering processes are very time-consuming and labor intensive. In many cases, those processes require iterative processing that measurably degrades image quality and resolution—especially in cases where thousands of sub-images are being rendered. In cases where the imaging data can be processed automatically, that data is often repetitively transformed and sampled—reducing color fidelity and image sharpness with each successive manipulation. If automated correction or balancing systems are employed, such systems may be susceptible to image anomalies (e.g., unusually bright or dark objects)—leading to over or under-corrections and unreliable interpretations of image data. In cases where manual rendering of images is required or desired, time and labor costs are immense.

There is, therefore, a need for an ortho-image rendering system that provides efficient and versatile imaging for very large FOVs and associated data sets, while maintaining image quality, accuracy, positional accuracy and clarity. Additionally, automation algorithms are applied extensively in every phase of the planning, collecting, navigating, and processing all related operations.

SUMMARY OF THE INVENTION

The present invention relates to remote data collection and processing system using a variety of sensors. The system may include computer console units that control vehicle and system operations in real-time. The system may also include global positioning systems that are linked to and communicate with the computer consoles. Additionally, cameras and/or camera array assemblies can be employed for producing an image of a target viewed through an aperture. The camera array assemblies are communicatively connected to the computer consoles. The camera array assembly has a mount housing, a first imaging sensor centrally coupled to the housing having a first focal axis passing through the aperture. The camera array assembly also has a second imaging sensor coupled to the housing and offset from the first imaging sensor along an axis, that has a second focal axis passing through the aperture and intersecting the first focal axis within an intersection area. The camera array assembly has a third imaging sensor, coupled to the housing and offset from the first imaging sensor along the axis, opposite the second imaging sensor, that has a third focal axis passing through the aperture and intersecting the first focal axis within the intersection area. Any number of one-to-n cameras may be used in this manner, where “n” can be any odd or even number.

The system may also include an Attitude Measurement Unit (AMU) such as inertial, optical, or similar measurement units communicatively connected to the computer consoles and the camera array assemblies. The AMU may determine the yaw, pitch, and/or roll of the aircraft at any instant in time and successive DGPS positions may be used to measure the vehicle heading with relation to geodesic north. The AMU data is integrated with the precision DGPS data to produce a robust, real-time AMU system. The system may further include a mosaicing module housed within the computer consoles. The mosaicing module includes a first component for performing initial processing on an input image. The mosaicing module also includes a second component for determining geographical boundaries of an input image with the second component being cooperatively engaged with the first component. The mosaicing module further includes a third component for mapping an input image into the composite image with accurate geographical position. The third component being cooperatively engaged with the first and second components. A fourth component is also included in the mosaicing module for balancing color of the input images mapped into the composite image. The fourth component can be cooperatively engaged with the first, second and third components. Additionally, the mosaicing module can include a fifth component for blending borders between adjacent input images mapped into the composite image. The fifth component being cooperatively engaged with the first, second, third and fourth components.

A sixth component, an optional forward oblique and/or optional rear oblique camera array system may be implemented that collects oblique image data and merges the image data with attitude and positional measurements in order to create a digital elevation model using stereographic techniques. Creation of which may be performed in real-time onboard the vehicle or post processed later. This sixth component works cooperatively with the other components. All components may be mounted to a rigid platform for the purpose of providing co-registration of sensor data. Vibrations, turbulence, and other forces may act on the vehicle in such a way as to create errors in the alignment relationship between sensors. Utilization of common, rigid platform mount for the sensors provides a significant advantage over other systems that do not use this co-registration architecture.

Further, the present invention may employ a certain degree of lateral oversampling to improve output quality and/or co-mounted, co-registered oversampling to overcome physical pixel resolution limits.

BRIEF DESCRIPTION OF THE DRAWINGS

For a better understanding of the invention, and to show by way of example how the same may be carried into effect, reference is now made to the detailed description of the invention along with the accompanying figures in which corresponding numerals in the different figures refer to corresponding parts and in which:

FIG. 1 illustrates a vehicle based data collection and processing system of the present invention;

FIG. 1A illustrates a portion of the vehicle based data collection and processing system ofFIG. 1;

FIG. 1B illustrates a portion of the vehicle based data collection and processing system ofFIG. 1;

FIG. 2 illustrates a vehicle based data collection and processing system ofFIG. 1 with the camera array assembly of the present invention shown in more detail;

FIG. 3 illustrates a camera array assembly in accordance with certain aspects of the present invention;

FIG. 4 illustrates one embodiment of an imaging pattern retrieved by the camera array assembly ofFIG. 1;

FIG. 5 depicts an imaging pattern illustrating certain aspects of the present invention;

FIG. 6 illustrates an image strip in accordance with the present invention;

FIG. 7 illustrates another embodiment of an image strip in accordance with the present invention;

FIG. 8 illustrates one embodiment of an imaging process in accordance with the present invention;

FIG. 9 illustrates diagrammatically how photos taken with the camera array assembly can be aligned to make an individual frame;

FIG. 10 is a block diagram of the processing logic according to certain embodiments of the present invention;

FIG. 11 is an illustration of lateral oversampling looking down from a vehicle according to certain embodiments of the present invention;

FIG. 12 is an illustration of lateral oversampling looking down from a vehicle according to certain embodiments of the present invention;

FIG. 13 is an illustration of flight line oversampling looking down from a vehicle according to certain embodiments of the present invention;

FIG. 14 is an illustration of flight line oversampling looking down from a vehicle according to certain embodiments of the present invention;

FIG. 15 is an illustration of progressive magnification looking down from a vehicle according to certain embodiments of the present invention;

FIG. 16 is an illustration of progressive magnification looking down from a vehicle according to certain embodiments of the present invention;

FIG. 17 is an illustration of progressive magnification looking down from a vehicle according to certain embodiments of the present invention;

FIG. 18 is a schematic of the system architecture according to certain embodiments of the present invention;

FIG. 19 is an illustration of lateral co-mounted, co-registered oversampling in a sidelap sub-pixel area for a single camera array looking down from a vehicle according to certain embodiments of the present invention;

FIG. 20 is an illustration of lateral co-mounted, co-registered oversampling in a sidelap sub-pixel area for two overlapping camera arrays looking down from a vehicle according to certain embodiments of the present invention; and

FIG. 21 is an illustration of fore and lateral co-mounted, co-registered oversampling in sidelap sub-pixel areas for two stereo camera arrays looking down from a vehicle according to certain embodiments of the present invention.

DETAILED DESCRIPTION OF THE INVENTION

While the making and using of various embodiments of the present invention are discussed in detail below, it should be appreciated that the present invention provides many applicable inventive concepts, which can be embodied in a wide variety of specific contexts. The specific embodiments discussed herein are merely illustrative of specific ways to make and use the invention and do not limit the scope of the invention.

A vehicle based data collection andprocessing system100 of the present invention is shown inFIGS. 1, 1A, and 1B. Additional aspects and embodiments of the present invention are shown inFIGS. 2 and 18.System100 includes one or more computer consoles102. The computer consoles contain one ormore computers104 for controlling both vehicle and system operations. Examples of the functions of the computer console are the controlling digital color sensor systems that can be associated with the data collection and processing system, providing the display data to a pilot, coordinating the satellite generated GPS pulse-per-second (PPS) event trigger (which may be 20 or more pulses per second), data logging, sensor control and adjustment, checking and alarming for error events, recording and indexing photos, storing and processing data, flight planning capability that automates the navigation of the vehicle, data, and providing a real-time display of pertinent information. A communications interface between the control computer console and the vehicle autopilot control provides the ability to actually control the flight path of the vehicle in real-time. This results in a more precise control of the vehicle's path than is possible by a human being. All of these functions can be accomplished by the use of various computer programs that are synchronized to the GPS PPS signals and take into account the various electrical latencies of the measurement devices. In an embodiment, the computer is embedded within the sensor.

One or more differentialglobal positioning systems106 are incorporated into thesystem100. Theglobal positioning systems106 are used to navigate and determine precise flight paths during vehicle and system operations. To accomplish this, theglobal positioning systems106 are communicatively linked to thecomputer console102 such that the information from theglobal positioning systems106 can be acquired and processed without flight interruption. Zero or more GPS units may be located at known survey points in order to provide a record of each sub-seconds' GPS satellite-based errors in order to be able to back correct the accuracy of thesystem100. GPS and/or ground based positioning services may be used that eliminate the need for ground control points altogether. This technique results in greatly improved, sub-second by sub-second positional accuracy of the data capture vehicle.

One or more AMUs108 that provide real-time yaw, pitch, and roll information that is used to accurately determine the attitude of the vehicle at the instant of data capture are also communicatively linked to thecomputer console102. The present attitude measurement unit (AMU) (e.g., Applanix POS AV), uses three high performance fiber optic gyros, one gyro each for yaw, pitch, and roll measurement. AMUs from other manufacturers, and AMUs that use other inertial measurement devices can be used as well. Additionally, an AMU may be employed to determine the instantaneous attitude of the vehicle and make the system more fault tolerant to statistical errors in AMU readings. Connected to the AMU can be one or moremulti-frequency DGPS receivers110. Themulti-frequency DGPS receivers110 can be integrated with the AMU's yaw, pitch, and roll attitude data in order to more accurately determine the location of the remote sensor platform in three dimensional space. Additionally, the direction of geodesic North may be determined by the vector created by successive DGPS positions, recorded in a synchronized manner with the GPS PPS signals.

One or morecamera array assemblies112 for producing an image of a target viewed through an aperture are also communicatively connected to the one or more computer consoles102. Thecamera array assemblies112, which will be described in greater detail below, provide the data collection and processing system with the ability to capture high resolution, high precision progressive scan or line scan, color digital photography.

The system may also include DC power andconditioning equipment114 to condition DC power and to invert DC power to AC power in order to provide electrical power for the system. The system may further include anavigational display116, which graphically renders the position of the vehicle versus the flight plan for use by the pilot (either onboard or remote) of the vehicle to enable precision flight paths in horizontal and vertical planes. The system may also include an EMU module comprised of LIDAR,SAR118 or a forward and rear oblique camera array for capturing three dimensional elevation/relief data. TheEMU module118 can include alaser unit120, anEMU control unit122, and anEMU control computer124. Temperature controlling devices, such as solid state cooling modules, can also be deployed as needed in order to provide the proper thermal environment for the system.

The system also includes a mosaicing module, not depicted, housed with thecomputer console102. The mosaicing module, which will be described in further detail below, provides the system the ability to gather data acquired by theglobal positioning system106, theAMU108, and thecamera system112 and process that data into useable orthomaps.

Thesystem100 also can include a self-locking flight path technique that provides the ability to micro-correct the positional accuracy of adjacent flight paths in order to realize precision that exceeds the native precision of the AMU and DGPS sensors alone.

A complete flight planning methodology is used to micro plan all aspects of missions. The inputs are the various mission parameters (latitude/longitude, resolution, color, accuracy, etc.) and the outputs are detailed on-line digital maps and data files that are stored onboard the data collection vehicle and used for real-time navigation and alarms. The ability to interface the flight planning data directly into the autopilot is an additional integrated capability. A computer program may be used that automatically controls the flight path, attitude adjustments, graphical display, moving maps of the vehicle path, checks for alarm conditions and corrective actions, notifies the pilot and/or crew of overall system status, and provides for fail-safe operations and controls. Safe operations parameters may be constantly monitored and reported. Whereas the current system uses a manned crew, the system is designed to perform equally well in an unmanned vehicle.

FIG. 2 shows another depiction of the present invention. InFIG. 2, thecamera array assembly112 is shown in more detail. As is shown, thecamera array assembly112 allows for images to be acquired from the rear oblique, the forward obliques and the nadir positions.FIG. 3 describes in more detail a camera array assembly of the present invention.FIG. 3 provides acamera array assembly300 airborne over target302 (e.g., terrain). For illustrative purposes, the relative size ofassembly300, and the relative distance between it andterrain302, are not depicted to scale inFIG. 3. Thecamera array assembly300 comprises ahousing304 within whichimaging sensors306,308,310,312 and314 are disposed along a concavecurvilinear axis316. The radius of curvature ofaxis316 may vary or be altered dramatically, providing the ability to effect very subtle or very drastic degrees of concavity inaxis316. Alternatively,axis316 may be completely linear—having no curvature at all. Theimaging sensors306,308,310,312 and314 couple to thehousing304, either directly or indirectly, byattachment members318.Attachment members318 may comprise a number of fixed or dynamic, permanent or temporary, connective apparatus. For example, theattachment members318 may comprise simple welds, removable clamping devices, or electro-mechanically controlled universal joints.

Additionally, thesystem100 may have a real-time, onboard navigation system to provide a visual, bio-feedback display to the vehicle pilot, or remote display in the case of operations in an unmanned vehicle. The pilot is able to adjust the position of the vehicle in real-time in order to provide a more accurate flight path. The pilot may be onboard the vehicle or remotely located and using the flight display to control the vehicle through a communication link.

Thesystem100 may also use highly fault-tolerant methods that have been developed to provide a software interleaved disk storage methodology that allows one or two hard drives to fail and still not lose target data that is stored on the drives. This software inter-leaved disk storage methodology provides superior fault-tolerance and portability versus other, hardware methodologies, such as RAID-5.

Thesystem100 may also incorporate a methodology that has been developed that allows for a short calibration step just before mission data capture. The calibration methodology step adjusts the camera settings, mainly exposure time, based on sampling the ambient light intensity and setting near optimal values just before reaching the region of interest. A moving average algorithm is then used to make second-by-second camera adjustments in order to deliver improved, consistent photo results. This improves the color processing of the orthomaps. Additionally, the calibration may be used to check or to establish the exact spatial position of each sensor device (cameras, DPG, AMU, EMU, etc.). In this manner, changes that may happen in the spatial location of these devices may be accounted for and maintain overall system precision metrics.

Additionally, thesystem100 may incorporate a methodology that has been developed that allows for calibrating the precision position and attitude of each sensor device (cameras, DPG, AMU, EMU, etc.) on the vehicle by flying over an area that contains multiple known, visible, highly accurate geographic positions. A program takes this data as input and outputs the micro positional data that is then used to precisely process the orthomaps.

As depicted inFIG. 3,housing304 comprises a simple enclosure inside of whichimaging sensors306,308,310,312 and314 are disposed. WhereasFIG. 3 depicts a 5-camera array, the system works equally well when utilizing any number of camera sensors from 1 to any number.Sensors306 through314 couple, via theattachment members318, either collectively to a single transverse cross member, or individually to lateral cross members disposed between opposing walls of thehousing304. In alternative embodiments, thehousing304 may itself comprise only a supporting cross member of concave curvature to which theimaging sensors306 through314 couple, viamembers318. In other embodiments, thehousing304 may comprise a hybrid combination of enclosure and supporting cross member. Thehousing304 further comprises an aperture320 formed in its surface, between the imaging sensors andtarget302. Depending upon the specific type of host craft, the aperture320 may comprise only a void, or it may comprise a protective screen or window to maintain environmental integrity within thehousing304. In the event that a protective transparent plate is used for any sensor, special coatings may be applied to the plate to improve the quality of the sensor data. Optionally, the aperture320 may comprise a lens or other optical device to enhance or alter the nature of the images recorded by the sensors. The aperture320 is formed with a size and shape sufficient to provide theimaging sensors306 through314 proper lines of sight to atarget region322 onterrain302.

Theimaging sensors306 through314 are disposed within or alonghousing304 such that the focal axes of all sensors converge and intersect each other within an intersection area bounded by the aperture320. Depending upon the type of image data being collected, the specific imaging sensors used, and other optics or equipment employed, it may be necessary or desirable to offset the intersection area or point of convergence above or below the aperture320. Theimaging sensors306 through314 are separated from each other at angular intervals. The exact angle of displacement between the imaging sensors may vary widely depending upon the number of imaging sensors utilized and on the type of imaging data being collected. The angular displacement between the imaging sensors may also be unequal, if required, so as to provide a desired image offset or alignment. Depending upon the number of imaging sensors utilized, and the particular configuration of the array, the focal axes of all imaging sensors may intersect at exactly the same point, or may intersect at a plurality of points, all within close proximity to each other and within the intersection area defined by the aperture320.

As depicted inFIG. 3, the imaging sensor310 is centrally disposed within thehousing304 alongaxis316. The imaging sensor310 has afocal axis324, directed orthogonally from thehousing304 to align the line of sight of the imaging sensor with theimage area326 of theregion322. Theimaging sensor308 is disposed within thehousing304 along theaxis316, adjacent to the imaging sensor310. Theimaging sensor308 is aligned such that its line of sight coincides with theimage area328 of theregion322, and such that itsfocal axis330 converges with and intersects theaxis324 within the area bounded by the aperture320. Theimaging sensor312 is disposed within thehousing304 adjacent to the imaging sensor310, on the opposite side of theaxis316 as theimaging sensor308. Theimaging sensor312 is aligned such that its line of sight coincides with theimage area332 of theregion322, and such that itsfocal axis334 converges with and intersectsaxes324 and330 within the area bounded by the aperture320. Theimaging sensor306 is disposed within thehousing304 along theaxis316, adjacent to thesensor308. Theimaging sensor306 is aligned such that its line of sight coincides with theimage area336 ofregion322, and such that itsfocal axis338 converges with and intersects the other focal axes within the area bounded by aperture320. Theimaging sensor314 is disposed withinhousing304 adjacent tosensor312, on the opposite side ofaxis316 assensor306. Theimaging sensor314 is aligned such that its line of sight coincides withimage area340 ofregion322, and such that itsfocal axis344 converges with and intersects the other focal axes within the area bounded by aperture320.

Theimaging sensors306 through314 may comprise a number of digital imaging devices including, for example, individual area scan cameras, line scan cameras, infrared sensors, hyperspectral and/or seismic sensors. Each sensor may comprise an individual imaging device, or may itself comprise an imaging array. Theimaging sensors306 through314 may all be of a homogenous nature, or may comprise a combination of varied imaging devices. For ease of reference, theimaging sensors306 through314 are hereafter referred to ascameras306 through314, respectively.

In large-format film or digital cameras, lens distortion is typically a source of imaging problems. Each individual lens must be carefully calibrated to determine precise distortion factors. In one embodiment of this invention, small-format digital cameras having lens angle widths of 17 degrees or smaller are utilized. This alleviates noticeable distortion efficiently and affordably.

Cameras306 through314 are alternately disposed withinhousing304 alongaxis316 such that each camera's focal axis converges upon aperture320, crossesfocal axis324, and aligns its field of view with a target area opposite its respective position in the array resulting in a “cross-eyed”, retinal relationship between the cameras and the imaging target(s). Thecamera array assembly300 is configured such that adjoining borders ofimage areas326,328,332,336 and340 overlap slightly.

If theattachment members318 are of a permanent and fixed nature (e.g., welds), then the spatial relationship between the aperture320, the cameras, and their lines of sight remain fixed as will the spatial relationship betweenimage areas326,328,332,336 and340. Such a configuration may be desirable in, for example, a satellite surveillance application where thecamera array assembly300 will remain at an essentially fixed distance fromregion322. The position and alignment of the cameras is set such thatareas326,328,332,336 and340 provide full imaging coverage ofregion322. If theattachment members318 are of a temporary or adjustable nature, however, it may be desirable to selectively adjust, either manually or by remote automation, the position or alignment of the cameras so as to shift, narrow or widenareas326,328,332,336 and340—thereby enhancing or altering the quality of images collected by thecamera array assembly300.

In an embodiment, multiple, i.e., at least two, rigid mount units are affixed to the same rigid mount plate. The mount unit is any rigid structure to which at least one imaging sensor may be affixed. The mount unit is preferably a housing, which encloses the imaging sensor, but may be any rigid structure including a brace, tripod, or the like. For the purposes of this disclosure, an imaging sensor means any device capable of receiving and processing active or passive radiometric energy, i.e., light, sound, heat, gravity, and the like, from a target area. In particular, imaging sensors may include any number of digital cameras, including those that utilize a red-blue-green filter, a bushbroom filter, or a hyperspectral filter, LIDAR sensors, infrared sensors, heat-sensing sensors, gravitometers and the like. Imagining sensors do not include attitude measuring sensors such as gyroscopes, GPS devices, and the like devices, which serve to orient the vehicle with the aid of satellite data and/or inertial data. Preferably, the multiple sensors are different.

In the embodiment wherein the imaging sensor is a camera, LIDAR, or the like imaging sensor, the mount unit preferably has an aperture through which light and/or energy may pass. The mount plate is preferably planer, but may be non-planer. In the embodiment, wherein the imaging sensor is a camera, LIDAR, or the like imaging sensor, the mount plate preferably has aperture(s) in alignment with the aperture(s) of the mount unit(s) through which light and/or energy may pass.

A rigid structure is one that flexes less than about 100^thof a degree, preferably less than about 1,000^thof a degree, more preferably less than about 10,000^thof a degree while in use. Preferably, the rigid structure is one that flexes less than about 100^thof a degree, preferably less than about 1,000^thof a degree, more preferably less than about 10,000^thof a degree while secured to an aircraft during normal, i.e., non-turbulent, flight. Objects are rigidly affixed to one another if during normal operation they flex from each other less than about 100^thof a degree, preferably less than about 1,000^thof a degree, more preferably less than about 10,000^thof a degree.

Camera310 is designated as the principal camera. Theimage plane326 of camera310 serves as a plane of reference. The orientations of theother cameras306,308,312 and314 are measured relative to the plane of reference. The relative orientations of each camera are measured in terms of the yaw, pitch and roll angles required to rotate the image plane of the camera to become parallel to the plane of reference. The order of rotations is preferably yaw, pitch, and roll.

The imaging sensors affixed to the mount unit(s) may not be aligned in the same plane. Instead, the angle of their mount relative to the mount angle of a first sensor affixed to the first mount unit, preferably the principle nadir camera of the first mount unit, may be offset. Accordingly, the imaging sensors may be co-registered to calibrate the physical mount angle offset of each imaging sensor relative to each other. In an embodiment, multiple, i.e., at least two, rigid mount units are affixed to the same rigid mount plate and are co-registered. In an embodiment, thecameras306 through314 are affixed to a rigid mount unit and co-registered. In this embodiment, the geometric centerpoint of the AMU, preferably a gyroscope, is determined using GPS and inertial data. The physical position of the first sensor affixed to the first mount unit, preferably the principle nadir camera of the first mount unit, is calculated relative to a reference point, preferably the geometric centerpoint of the AMU. Likewise, the physical position of all remaining sensors within all mount units are calculated—directly or indirectly—relative to the same reference point.

The boresight angle of a sensor is defined as the angle from the geometric center of that sensor to a reference plane. Preferably the reference plane is orthogonal to the target area. The boresight angle of the first sensor may be determined using the ground target points. The boresight angles of subsequent sensors are preferably calculated with reference to the boresight angle of the first sensor. The sensors are preferably calibrated using known ground targets, which are preferably photo-identifiable, and alternatively calibrated using a self-locking flight path or any other method as disclosed in U.S. Publication No. 2004/0054488A1, now U.S. Pat. No. 7,212,938B2, the disclosure of which is hereby incorporated by reference in full.

The imaging sensor within the second mount unit may be any imaging sensor, and is preferably a LIDAR. Alternative, the second imaging sensor is a digital camera, or array of digital cameras. In an embodiment, the boresight angle of so the sensor(s) affixed to the second mount unit are calculated with reference to the boresight angle of the first sensor. The physical offset of the imaging sensor(s) within the second mount unit may be calibrated with reference to the boresight angle of the first sensor within the first mount unit.

In this manner, all of the sensors are calibrated at substantially the same epoch, using the same GPS signal, the same ground target(s), and under substantially the same atmospheric conditions. This substantially reduces compounded error realized when calibrating each sensor separately, using different GPS signals, against different ground targets, and under different atmospheric conditions.

Referring now toFIG. 4, images ofareas336,328,326,332 and340 taken bycameras306 through314, respectively, are illustrated from an overhead view. Again, because of the “cross-eyed” arrangement, the image ofarea336 is taken bycamera306, the image ofarea340 is taken bycamera314, and so on. In one embodiment of the present invention, images other than those taken by the center camera310 take on a trapezoidal shape after perspective transformation.Cameras306 through314 form an array alongaxis316 that is, in most applications, pointed down vertically. In an alternative embodiment, a second array of cameras, configured similar the array ofcameras306 through314, is aligned with respect to the first array of cameras to have an oblique view providing a “heads-up” perspective. The angle of declination from horizontal of the heads-up camera array assembly may vary due to mission objectives and parameters but angles of 25-45 degrees are typical. Other alternative embodiments, varying the mounting of camera arrays, are similarly comprehended by the present invention. In all such embodiments, the relative positions and attitudes of the cameras are precisely measured and calibrated so as to facilitate image processing in accordance with the present invention.

In one embodiment of the present invention, an external mechanism (e.g., a GPS timing signal) is used to trigger the cameras simultaneously thereby capturing an array of input images. A mosaicing module then renders the individual input images from such an array into an ortho-rectified compound image (or “mosaic”), without any visible seams between the adjacent images. The mosaicing module performs a set of tasks comprising: determining the geographical boundaries and dimensions of each input image; projecting each input image onto the mosaic with accurate geographical positioning; balancing the color of the images in the mosaic; and blending adjacent input images at their shared seams. The exact order of the tasks performed may vary, depending upon the size and nature of the input image data. In certain embodiments, the mosaicing module performs only a single transformation to an original input image during mosaicing. That transformation can be represented by a 4×4 matrix. By combining multiple transformation matrices into a single matrix, processing time is reduced and original input image sharpness is retained.

During mapping of the input images to the mosaic, especially when mosaicing is performed at high resolutions, pixels in the mosaic (i.e., output pixels) may not be mapped to by any pixels in the input images (i.e., input pixels). Warped lines could potentially result as artifacts in the mosaic. Certain embodiments of the present invention overcome this with a super-sampling system, where each input and output pixel is further divided into an n×m grid of sub-pixels. Transformation is performed from sub-pixels to sub-pixels. The final value of an output pixel is the average value of its sub-pixels for which there is a corresponding input sub-pixel. Larger n and m values produce mosaics of higher resolution, but do require extra processing time.

During its processing of image data, the mosaicing module may utilize the following information: the spatial position (e.g., x, y, z coordinates) of each camera's focal point at the time an input image is captured; the attitude (i.e., yaw, pitch, roll) of each camera's image plane relative to the target region's ground plane at the time an input image was captured; each camera's fields of view (i.e., along track and cross track); and the Digital Terrain Model (DTM) of the area. The attitude can be provided by the AMUs associated with the system. Digital terrain models (DTMs) or Digital surface models (DSMs) can be created from information obtained using aLIDAR module118. LIDAR is similar to the more familiar radar, and can be thought of as laser radar. In radar, radio waves are transmitted into the atmosphere that scatters some of the energy back to the radar's receiver. LIDAR also transmits and receives electromagnetic radiation, but at a higher frequency since it operates in the ultraviolet, visible and infrared region of the electromagnetic spectrum. In operation, LIDAR transmits light out to a target area. The transmitted light interacts with and is changed by the target area. Some of this light is reflected/scattered back to the LIDAR instrument where it can be analyzed. The change in the properties of the light enables some property of the target area to be determined. The time for the light to travel out to the target area and back to LIDAR device is used to determine the range to the target.

DTM and DSM data sets can also be captured from the camera array assembly. Traditional means of obtaining elevation data may also be used such as stereographic techniques.

There are presently three basic types of LIDAR: Range finders, Differential Absorption LIDAR (DIAL) and Doppler LIDAR. Range finder LIDAR is the simplest LIDAR and is used to measure the distance from the LIDAR device to a solid or hard target. DIAL LIDAR is used to measure chemical concentrations (such as ozone, water vapor, pollutants) in the atmosphere. A DIAL LIDAR uses two different laser wavelengths that are selected so that one of the wavelengths is absorbed by the molecule of interest while the other wavelength is not. The difference in intensity of the two return signals can be used to deduce the concentration of the molecule being investigated. Doppler LIDAR is used to measure the velocity of a target. When the light transmitted from the LIDAR hits a target moving towards or away from the LIDAR, the wavelength of the light reflected/scattered off the target will be changed slightly. This is known as a Doppler-shift and therefore Doppler LIDAR. If the target is moving away from the LIDAR, the return light will have a longer wavelength (sometimes referred to as a red shift), if moving towards the LIDAR the return light will be at a shorter wavelength (blue shifted). The target can be either a hard target or an atmospheric target (e.g. microscopic dust and aerosol particles that are carried by the wind.

A camera's focal point is preferably used as a perspective transformation center. Its position in space may be determined, for example, by a multi-frequency carrier phase post-processed GPS system mounted on the host craft. The offsets, in three dimensions, of a camera's focal point are preferably carefully measured against the center of the GPS antenna. These offsets may be combined with the position of the GPS antenna, and the orientation of the host craft, to determine the exact position of the camera's focal point. The position of the GPS antenna is preferably determined by processing of collected GPS data against similar ground-based GPS antennas deployed at precisely surveyed points.

One or more AMUs (e.g., the Applanix POS AV) are preferably mounted onboard for attitude determination. The attitude of the AMU reference plane relative to the target region's ground plane is preferably measured and recorded at short intervals, with accuracy better than one-hundredth of one degree. The attitude of the AMU reference plane may be defined as the series of rotations that can be performed on the axes of this plane to make it parallel to the ground plane. The term “align” could also be used to describe this operation.

The attitude of center camera310 (i.e. its image plane), relative to the AMU, is preferably precisely calibrated. The attitude of each of the other cameras, relative to center camera310, is preferably also be carefully calibrated. This dependent calibration is more efficient than directly calibrating each camera. When thecamera array assembly300 is remounted, only center camera310 needs to be recalibrated. Effectively, a series of two transformations is applied to an input image from center camera310. First, the center camera's image plane is aligned to the AMU plane. Then, the AMU plane is aligned again to the ground plane. These transformations, however, combine into a single operation by multiplying their respective transformation matrices. For images from each of the other cameras, an additional transformation is first performed to align it with the center camera's image plane.

The position of the focal point of center camera310 may be determined as described above. The x and y components of this position preferably determine the position of the mosaic'snadir point400 on the ground. Field of view (FOV) angles of each camera are known, thus the dimensions of each input image may be determined by the z component of that camera's focal point. An average elevation of the ground is preferably determined by computing the average elevation of points in the DTMs of the area, and then each input image is projected to an imaginary horizontal plane at this elevation. Relief displacement is then preferably applied using the DTMs of the area. The DTMs can be obtained from many sources including: the USGS 30-or 10-meter DTMs available for most of the US; commercial DTMs; or DTMs obtained by a LIDAR or SAR EMU device mounted on the host craft that captures data concurrently with the cameras.

Besides being geographically correctly placed, the resulting compound image also needs to have radiometric consistency throughout, and no visible seams at the joints between two adjacent images. The present invention provides a number of techniques for achieving this goal.

A characteristic of a conventional camera is the exposure time (i.e., the time the shutter is open to collect light onto the image plane). The longer the exposure time, the lighter the resultant image becomes. Exposure time must adapt to changes in ambient lighting caused by conditions such as: cloud coverage; the angle and position of the sun relative to the camera; and so forth. Optimal exposure time may also depend on a camera's orientation with respect to lighting sources (e.g., cameras pointing towards a sunlit object typically receive more ambient light than those pointing towards a shaded object). Exposure time is adjusted to keep the average intensity of an image within a certain desired range. For example, in 24-bit color images each Red, Green and Blue component can have intensity values from 0 to 255. In most instances, however, it is desirable to keep the average intensity at a mean value (i.e., 127).

In the present invention, an exposure control module controls exposure time for each of the cameras or imaging sensors. It examines each input image and calculates average image intensity. Based on a moving average (i.e., average intensity of the last X number of images), the exposure control module determines whether to increase or decrease exposure time. The module can use a longer running average to effect a slower reaction to changes in lighting conditions, with less susceptibility to unusually dark or light images (e.g., asphalt roads or water). The exposure control module controls exposure time for each camera separately.

In systems where cameras are mounted without forward-motion compensation mechanisms, there must be a maximum limit for exposure time. Setting exposure time to a value larger than the maximum may cause motion-induced blurriness. For example, assume cameras are mounted on an airplane traveling at 170 miles/hour (or about 3 inches/ms). Assume desired pixel resolution is 6 inches. Forward motion during image capture should be limited to half a pixel size—which in this case equals 3 inches. Thus, maximum exposure for example is 1 millisecond.

In controlling imaging quality, it is useful to be able to determine if changes in light intensity are caused either due to a change in ambient light or due to the presence of unusually light or dark objects (e.g., reflecting water body, metal roofs, asphalts, etc.). Certain applications of this invention involve aerial photography or surveillance. It is observed that aerial images of the ground usually contain plants and vegetation—which have more consistent reflectivity than water bodies or man-made structures such as roads and buildings. Of course, images of plants and vegetation are usually green-dominant (i.e., the green component is the greatest of the red, green and blue values). Therefore, intensity correlation can be made more accurate by focusing on the green-dominant pixels.

The exposure control module computes the average intensity of an image by selecting only green-dominant pixels. For example, if an image has 1 million pixels and 300,000 are green-dominant, only those 300,000 green-dominant pixels are included in the calculation of average intensity. This results in an imaging process that is less susceptible to biasing caused by man-made structures and water bodies, whose pixels are usually not green-dominant. As previously noted, it is desirable to maintain an intensity value of about 127. When intensity value is over 127 (i.e., over-exposed), exposure time is reduced so that less light is captured. Similarly, when intensity value is under 127 (i.e., under-exposed), exposure time is increased so that more light is captured. For example, consider a system flying over a target terrain area having many white roofs, whose intensities are very high. Average intensity for the images captured would tend to be high. In most conventional systems, exposure time would by reduced in order to compensate. In such an example, however, reducing exposure time is not proper, because the average intensity of the images has been biased by the bright roofs. Reducing exposure time would result in images where the ground is darker than it should be. In contrast, if only green-dominant pixels are processed in accordance with the present invention, then pixels representing the overly bright roofs do bias the average intensity and the exposure time is not changed.

Thus, the exposure control module reduces intensity differences between input images. Nonetheless, further processing is provided to enhance tonal balance. There are a number of factors (e.g., lens physics, atmospheric conditions, spatial/positional relationships of imaging devices) that cause an uneven reception of light from the image plane. More light is received in the center of a camera or sensor than at the edges.

The mosaicing module of the present invention addresses this with an anti-vignetting function, illustrated in reference now toFIG. 5. A number offocal columns500,502,504,506 and508 converge fromimage plane509 and cross throughfocal point510 as they range across imaging target area512 (e.g., ground terrain).Columns500 through508 may comprise individual resolution columns of a single camera or sensor, or may represent the focal axes of a number of independent cameras or sensors. For reference purposes,column504 serves as the axis andpoint513 at whichcolumn504 intersectsimage plane509 serves as a principal point. The exposure control module applies an anti-vignetting function multiplying the original intensity of an input pixel with a column-dependent anti-vignetting factor. Because the receiving surface is represented as a plane with a coordinate system, each column will have a number of resolution rows (not shown). This relationship may be expressed, for a pixel p at column x and row y, as follows:
<adjusted intensity>=<original intensity>*ƒ(x);
where ƒ(x) is a function of the form:
ƒ(x)=cos(off-axis angle)**4.
The off-axis angle514 is: zero forcenter column504; larger forcolumns502 and506; and larger still forcolumns500 and508. The overall field of view angle516 (FOVx angle) is depicted betweencolumns504 and508.

The function ƒ(x) can be approximated by a number of line segments between columns. For a point falling within a line segment between any given columns c1 and c2, an adjustment factor is computed as follows:
<adjustment factor for c>=ƒ(c1)+[ƒ(c2)−ƒ(c1)*(c−c1)/(c2−c1)];

where ƒ(c1) and ƒ(c2) are the ƒ function values of the off-axis angles at column c1 and c2, respectively.

Each set of input images needs to be stitched into a mosaic image. Even though the exposure control module regulates the amount of light each camera or sensor receives, the resulting input images may still differ in intensity. The present invention provides an intensity-balancing module that compares overlapping area between adjacent input images, to further balance the relative intensities. Because adjoining input images are taken simultaneously, the overlapping areas should, in theory, have identical intensity in both input images. However, due to various factors, the intensity values are usually not the same. Some such factors causing intensity difference could include, for example, the exposure control module being biased by unusually bright or dark objects present in the field of view of only a particular camera, or the boresight angles of cameras being different (i.e., cameras that are more slanted receive less light than those more vertical).

To balance two adjacent images, one is chosen as the reference image and the other is the secondary image. A correlation vector (fR, fG, FB) is determined using, for example, the following process. Let V be a 3×1 vector representing the values (R, G and B) of a pixel:

$V=\begin{array}{c}\begin{array}{c}R\\ G\end{array}\\ B\end{array}.$
A correlation matrix C may be derived as:

$C=\begin{array}{ccc}\mathrm{FR}& 0& 0\\ 0& \mathrm{FG}& 0\\ 0& 0& \mathrm{FB}\end{array};$
where FR=AvgIr/AvgIn; AvgIr=Red average intensity of overlapped region in reference image; AvgIn=Red average intensity of overlapped region in new image; and FG and FB are similarly derived.

The correlation matrix scales pixel values of the secondary image so that the average intensity of the overlapping area of the secondary image becomes identical to the average intensity of the overlapping area of the reference image. The second image can be balanced to the reference image by multiplying its pixel values by the correlation matrix.

Thus, in one embodiment of a balancing process according to the present invention, a center image is considered the reference image. The reference image is first copied to the compound image (or mosaic). Overlapping areas between the reference image and an adjoining image (e.g., the near left image) are correlated to compute a balancing correlation matrix (BCM). The BCM will be multiplied with vectors representing pixels of the adjoining image to make the intensity of the overlapping area identical in both images. One embodiment of this relationship may be expressed as:

Let I(center)=Average intensity of overlapping area in center image;

I(adjoining)=Average intensity of overlap in adjoining image; then
Balancing factor=I(center)/I(adjoining).

The balancing factor for each color channel (i.e., red, green and blue) is independently computed. These three values form the BCM. The now-balanced adjoining image is copied to the mosaic. Smooth transitioning at the border of the copied image is providing by “feathering” with a mask. This mask has the same dimension as the adjoining image and comprises a number of elements. Each element in the mask indicates the weight of the corresponding adjoining image pixel in the mosaic. The weight is zero for pixels at the boundary (i.e. the output value is taken from the reference image), and increases gradually in the direction of the adjoining image until it becomes unity—after a chosen blending width has been reached. Beyond the blending area, the mosaic will be entirely determined by the pixels of the adjoining image. Similarly, the overlaps between all the other constituent input images are analyzed and processed to compute the correlation vectors and to balance the intensities of the images.

A correlation matrix is determined using, for example, the following process with reference toFIG. 6.FIG. 6 depicts astrip600 being formed in accordance with the present invention. Abase mosaic602 and anew mosaic604, added along path (or track)606, overlap each other inregion608. Let V be a vector that represents the R, G and B values of a pixel:

$V=\begin{array}{c}\begin{array}{c}R\\ G\end{array}\\ B\end{array}$
Let h be the transition width ofregion608, and y be the along-track606 distance from theboundary610 of the overlapped region to a point A, whose pixel values are represented by V. Let C be the correlation matrix:

$C=\begin{array}{ccc}\mathrm{FR}& 0& 0\\ 0& \mathrm{FG}& 0\\ 0& 0& \mathrm{FB}\end{array}$
The balanced value of V, called V′ is:
V′=[y/h·I+(1−y/h)·C]×V, for 0<y<h;
V′=V, for y>=h;

Where I is the identity matrix

$I=\begin{array}{ccc}1& 0& 0\\ 0& 1& 0\\ 0& 0& 1\end{array}.$
Note that the “feathering” technique is also used in combination with the gradient to minimize seam visibility.

When mosaics are long, differences in intensity at the overlap may change from one end of the mosaic to the other. Computing a single correlation vector to avoid creating visible seams may not be possible. The mosaic can be divided into a number of segments corresponding to the position of the original input images that make up the mosaic. The process described above is applied to each segment separately to provide better local color consistency.

Under this refined algorithm, pixels at the border of two segments may create vertical seams (assuming north-south flight lines). To avoid this problem, balancing factors for pixels in this area have to be “transitioned” from that of one segment to the other. This is explained now with reference toFIG. 7.

FIG. 7 depicts astrip700 being formed in accordance with the present invention. Abase mosaic702 and anew segment704 overlap inarea706. Mosaic702 and anothernew segment708 overlap inarea710.Segments704 and708 overlap inarea712, andareas706,710 and712 all overlap and coincide atarea714. For explanation purposes,point716 serves as an origin for y-axis718 andx-axis720. Movement along y-axis718 represents movement along the flight path of the imaging system.Point716 is located at the lower left ofarea714.

According to the present invention, the dimensions of a strip are determined by the minimum and maximum x and y values of the constituent mosaics. An output strip is initialized to a background color. A first mosaic is transferred to the strip. The next mosaic (along the flight path) is processed next. Intensity values of the overlapping areas of the new mosaic and the first mosaic are correlated, separately for each color channel. The new mosaic is divided into a number of segments corresponding to the original input images that made up the mosaic. A mask matrix, comprising a number of mask elements, is created for the new mosaic. A mask element contains the correlation matrix for a corresponding pixel in the new mosaic. All elements in the mask are initialized to unity. The size of the mask can be limited to just the transition area of the new mosaic. The correlation matrix is calculated for the center segment. The mask area corresponding to the center segment is processed. The values of the elements at the edge of the overlap area are set to the correlation vector. Then, gradually moving away from the first mosaic along the strip, the components of the correlation matrix are either increased or decreased (whether they are less or more than unity, respectively) until they become unity at a predetermined transition distance. The area of the mask corresponding to a segment adjoining the center segment is then processed similarly. However, thearea714 formed by the first mosaic and the center and adjoining segments of the new image requires special treatment. Because the correlation matrix for the adjoining segment may not be identical to that of the center segment, a seam may appear at the border of the two segments in theoverlap area714 with the first mosaic. Therefore, the corner is influenced by the correlation matrices from both segments. For a mask cell A at distance x to the border with the center segment and distance y to the overlap edge, its correlation matrix is the distance-weighted average of the two segments, evaluated as follows:

• For pixel A(x, y) inarea714 at distance x to the border with the center segment, its balanced values are computed as the distance-weighted averages of the values computed using the two segments;
• V1 is the balanced RGB vector based onsegment704;
• V2 is the balanced RGB vector based onsegment708;
• V′ is the combined (final) balanced RGB vector
  V′=((d−x)/d)·V1+(x/d)·V2;
  Where
  • x-axis is the line going through bottom of overlapped region;
  • y-axis is the line going through the left side of the overlapped region betweensegments704 and708;
  • h is the transition width; and
  • d is the width of the overlapped region betweensegments704 and708.
    The mask areas corresponding to other adjoining segments are computed similarly.

Further according to the present invention, a color fidelity (i.e., white-balance) filter is applied. This multiplies R and B components with a determinable factor to enhance color fidelity. The factor may be determined by calibrating the cameras and lenses. The color fidelity filter ensures that the colors in an image retain their fidelity, as perceived directly by the human eye. Within the image capture apparatus, the Red, Green and Blue light receiving elements may have different sensitivities to the color they are supposed to capture. A “while-balance” process is applied—where image of a white object is captured. Theoretically, pixels in the image of that white object should have equivalent R, G and B values. In reality, however, due to different sensitivities and other factors, the average color values for each R, G and B may be avgR, avgG and avgB, respectively. To equalize the color components, the R, G and B values of the pixels are multiplied by the following ratios:

R values are multiplied by the ratio avgG/avgR; and

B values are multiplied by the ratio avgG/avgB.

The end result is that the image of the white object is set to have equal R G B components.

In most applications, a strip usually covers a large area of non-water surface. Thus, average intensity for the strip is unlikely to be skewed by anomalies such as highly reflecting surfaces. The present invention provides an intensity normalization module that normalizes the average intensity of each strip so that the mean and standard deviation are of a desired value. For example, a mean of 127 is the norm in photogrammetry. A standard deviation of 51 helps to spread the intensity value over an optimal range for visual perception of image features. Each strip may have been taken in different lighting conditions and, therefore, may have different imaging data profiles (i.e., mean intensity and standard deviation). This module normalizes the strips, such that all have the same mean and standard deviation. This enables the strips to be stitched together without visible seams.

This intensity normalization comprises a computation of the mean intensity for each channel R, G and B, and for all channels. The overall standard deviation is then computed. Each R, G and B value of each pixel is transformed to the new mean and standard deviation:
new value=new mean+(old value−old mean)*(new std/old std).

Next, multiple adjacent strips are combined to produce tiled mosaics for an area of interest. Finished tiles can correspond to the USGS quads or quarter-quads. Stitching strips into mosaics is similar to stitching mosaics together to generate strips, with strips now taking the role of the mosaics. At the seam line between two strips, problems may arise if the line crosses elevated structures such as buildings, bridges, etc. This classic problem in photogrammetry arises from the parallax caused by the same object being looked at from two different perspectives. During imaging of a building, for example, one strip may present a view from one side of the building while another strip presents a view from another side of the building. After the images are stitched together, the resulting mosaic may look like a tepee. In order to address this, a terrain-guided mosaicing process may be implemented to guide the placement of a seam line. For example, LIDAR or DEM data collected with, or analyzed from, image data may be processed to determine the configuration and shaping of images as they are mosaiced together. Thus, in some mosaiced images, a seam line may not be a straight line—instead comprising a seam line that shifts back and forth to snake through elevated structures.

Referring now toFIG. 8, one embodiment of animaging process800 is illustrated in accordance with the present invention as described above.Process800 begins with aseries802 of one, or more, raw collected images.Images802 are then processed through a white-balancingprocess804, transforming them into a series of intermediate images.Series802 is then processed throughanti-vignetting function806 before progressing to theorthorectification process808. As previously noted, orthorectification may rely on position andattitude data810 from the imaging sensor system or platform, and onDTM data812.DTM data812 may be developed fromposition data810 and from, for example,USGS DTM data814 orLIDAR data816.Series802 is now orthorectified and processing continues with color balancing818. After color balancing,series802 is converted bymosaicing module820 intocompound image822.Module820 performs the mosaicing and feathering processes during this conversion. Now, one ormore compound images822 are further combined instep824, by mosaicing with a gradient and feathering, intoimage strip826. Image strips are processed throughintensity normalization828. The now normalizedstrips828 are mosaiced together instep830, again by mosaicing with a gradient and feathering, rendering a finishingtiled mosaic832. The mosaicing performed instep830 may comprise a terrain-guided mosaicing, relying onDTM data812 orLIDAR data816.

FIG. 9 illustrates diagrammatically how photos taken with the camera array assembly may be aligned to make an individual frame. This embodiment shows a photo patter illustration looking down from a vehicle, using data ortho-rectified from five cameras.

FIG. 10 is a block diagram of the processing logic according to certain embodiments of the present invention. As shown in block diagram1000, the processing logic accepts one or more inputs, which may includeelevation measurements1002,attitude measurements1004 and/or photo andsensor imagery1006. Certain inputs may be passed through an initial processing step prior to analysis, as is shown inblock1008, wherein the attitude measurements are combined with data from ground control points.Elevation measurements1002 andattitude measurements1004 may be combined to generate processedelevation data1010.Processed elevation data1010 may then be used to generateelevation DEM1014 andDTM1016. Similarly,attitude measurements1006 may be combined with photo andsensor imagery1006 to generategeoreferenced images1012, which then undergoimage processing1018, which may include color balancing and gradient filtering.

Depending on the data set to be used (1020), eitherDTM1016 or aUSGS DEM1022 is combined with processedimages1018 to generateorthorectified imagery1024.Orthorectified imagery1024 then feeds into self-lockingflightlines1026. Balancingprojection mosaicing1028 then follows, to generatefinal photo output1030.

The present invention may employ a certain degree of lateral oversampling to improve output quality.FIG. 11 is an illustration of alateral oversampling pattern1100 looking down from a vehicle according to certain embodiments of the present invention showing minimal lateral oversampling. In this illustration, thecentral nadir region1102 assigned to the center camera overlaps only slightly with theleft nadir region1104 andright nadir region1106, so that overlap is minimizedFIG. 12 is an illustration of alateral oversampling pattern1200 looking down from a vehicle according to certain embodiments of the present invention showing a greater degree of lateral oversampling. In this illustration, thecentral nadir region1202 shows a high degree of overlap withleft nadir region1204 andright nadir region1206.

In addition to the use of lateral oversampling as shown inFIGS. 11 and 12, the present invention may employ flight line oversampling as well.FIG. 13 is an illustration of a flightline oversampling pattern1300 looking down from a vehicle according to certain embodiments of the present invention showing a certain degree of flight line oversampling but minimal lateral oversampling.Central nadir regions1302 and1304 are overlapped to one another along the flight line, but do not overlap laterally withleft nadir regions1306 and1308 or withright nadir regions1310 and1312.

FIG. 14 is an illustration of flight line oversampling looking down from a vehicle according to certain embodiments of the present invention showing significant flight line oversampling as well as significant lateral oversampling. It can be seen that each of thecentral nadir regions1402 through1406 are significantly overlapped with one another as well as withleft nadir regions1408 through1412 andright nadir regions1414 through1418.Left nadir regions1408 through1412 are overlapped with one another, as areright nadir regions1414 through1418. Accordingly, each point on the surface is sampled at least twice, and in some cases as many as four times. This technique uses the fact that in the area of an image that is covered twice, or more, by different camera sensors, a doubling of the image resolution is possible in both the lateral (across path) and flight line (along path) directions for an overall quadrupling of the resolution. In practice, the improvement in image/sensor resolution is somewhat less than doubled in each of the dimensions, approximately 40% in each dimension, or 1.4×1.4=˜2 times. This is due to the statistical variations of the sub-pixel alignment/orientation. In effect, the pixel grid is rarely exactly equidistant from the overlaid pixel grid. If extremely precise lateral camera sensor alignments were made at the sub-pixel level, a quadrupling of image resolution could be realized.

FIG. 15 is an illustration of aprogressive magnification pattern1500 looking down from a vehicle according to certain embodiments of the present invention.Central nadir region1502 is bounded on its left and right edges by innerleft nadir region1504 and innerright nadir region1506, respectively. Innerleft nadir region1504 is bounded on its left edge by outerleft nadir region1508, while innerright nadir region1506 is bounded on its right edge by outerright nadir region1510. Note that these regions exhibit a minimal degree of overlap and oversampling from one to another.

FIG. 16 is an illustration of aprogressive magnification pattern1600 looking down from a vehicle according to certain embodiments of the present invention.Central nadir region1602 is bounded on its left and right edges by innerleft nadir region1604 and innerright nadir region1606, respectively. Innerleft nadir region1604 is bounded on its left edge by outerleft nadir region1608, while innerright nadir region1606 is bounded on its right edge by outerright nadir region1610. Note that, as above, these regions exhibit a minimal degree of overlap and oversampling from one to another. Within each of thenadir regions1604 through1610, there is acentral image region1614 through1620 shown shaded in grey.

FIG. 17 is an illustration of aprogressive magnification pattern1700 looking down from a vehicle according to certain embodiments of the present invention. In the center ofpattern1700, a leftinner nadir region1702 and a rightinner nadir region1704 overlap in the center. A leftintermediate nadir region1706 and a rightintermediate nadir region1708 are disposed partly outside ofregions1702 and1704, respectively, each sharing an overlapping area with the respective adjacent area by approximately 50%. An outerleft nadir region1710 and an outerright nadir region1712 are disposed partly outside ofregions1706 and1708, respectively, each sharing an overlapping area with the respective adjacent area by approximately 50%. Acentral image region1714 is disposed in the center ofpattern1700, comprised of the central portions ofnadir regions1702 through1712.

FIG. 18 depicts a schematic of the architecture of asystem1800 according to certain embodiments of the present invention.System1800 may include one ormore GPS satellites1802 and one ormore SATCOM satellites1804. One or moreGPS location systems1806 may also be included, operably connected to one ormore modules1808 collecting LIDAR, GPS and/or X, Y, Z location data and feeding such information to one or more datacapture system applications1812. One or more datacapture system applications1812 may also receive spectral data from acamera array1822. ADGPS1810 may communicate with one ormore SATCOM satellites1804 via a wireless communications link1826. One ormore SATCOM satellites1804 may, in turn, communicate with one or more datacapture system applications1812.

One or more datacapture system applications1812 may interface with anautopilot1816, an SSD and/or aRealTime StitchG system1820, which may also interact with one another.SSD1814 may be operably connected toRealTime DEM1818. Finally,RealTime DEM1818 andRealTime StitchG1820 may be connected to a storage device, such asdisk array1824.

The present invention may employ a certain degree of co-mounted, co-registered oversampling to overcome physical pixel resolution limits.FIG. 19 is an illustration of a lateral co-mounted,co-registered oversampling configuration1900 for asingle camera array112 looking down from a vehicle according to certain embodiments of the present invention showing minimal lateral oversampling. The cameras overlap a few degrees in thevertical sidelap area1904 and1908. WhereasFIG. 19 depicts a 3-camera array, these subpixel calibration techniques work equally well when utilizing any number of camera sensors from 2 to any number of cameras being calibrated.

Similar to the imaging sensors inFIGS. 3 and 4, the camera sensors may be co-registered to calibrate the physical mount angle offset of each sensor relative to each other and/or to the nadir camera. This provides an initial, “close” calibration. These initial calibration parameters may be entered into anonboard computer system104 in thesystem100, and updated during flight using oversampling techniques.

Referring now toFIG. 19, the rectangles labeled A, B, and C representimage areas1902,1906 and1910 from a 3-camera array C-B-A (not shown). Images ofareas1902,1906 and1910 taken by cameras A through C (not shown), respectively, are illustrated from an overhead view. Again, similar toFIGS. 3 and 4, because of the “cross-eyed” arrangement, the image ofarea1902 is taken by right camera A, the image ofarea1906 is taken by center/nadir camera B, and the image ofarea1910 is taken by left camera C. Cameras A through C form an array (not shown) that is, in most applications, pointed down vertically.

InFIG. 19, the hatched areas labeled A/B and B/C sidelaps representimage overlap areas1904 and1908, respectively. The leftimage overlap area1904 is where right camera A overlaps with the center/nadir camera B, and the rightimage overlap area1908 is where the left camera C overlaps with the center/nadir camera B. In thesesidelap areas1904 and1908, the camera sensor grid bisects each pixel in theoverlap areas1904 and1908, which effectively quadruples the image resolution in theseareas1904 and1908 via the mechanism of co-mounted, co-registered over-sampling. In effect, the improvement in image/sensor resolution is doubled in each dimension, or 2×2=4 times. This quadrupling of the image resolution also quadruples the alignment precision between adjacent cameras.

Further, this quadrupling of alignment precision between adjacent cameras improves thesystems100 alignment precision for all sensors affixed to a rigid mount plate. The cameras and sensors are affixed to a rigid mount unit, which is affixed to the rigid mount plate, as discussed above. In particular, when the angular alignment of adjacent cameras affixed to the rigid mount unit is improved, the angular alignment of the other sensors is also enhanced. This enhancement of alignment precision for the other sensors affixed to the rigid mount plate also improves the image resolution for those sensors.

A lateral co-mounted,co-registered oversampling configuration2000 for two overlappingcamera arrays112 is illustrated inFIG. 20. In particular,FIG. 20 is an illustration of a lateral co-mounted,co-registered oversampling configuration2000 for two overlappingcamera arrays112 looking down from a vehicle according to certain embodiments of the present invention showing maximum lateral oversampling. The adjacent cameras overlap a few degrees in the vertical sidelap areas2006,2008,2014 and2016, and the corresponding cameras overlap completely in theimage areas2002,2010,2018 and2004,2012,2020. WhereasFIG. 20 depicts two 3-camera arrays, these subpixel calibration techniques work equally well when utilizing two overlapping camera arrays with any number of camera sensors from 2 to any number of cameras being calibrated.

Similar to the imaging sensors inFIGS. 3 and 4, the camera sensors may be co-registered to calibrate the physical mount angle offset of each sensor relative to each other and/or to the nadir camera. In this embodiment, multiple, i.e., at least two, rigid mount units are affixed to a rigid mount plate and are co-registered. This provides an initial, “close” calibration. These initial calibration parameters may be entered into anonboard computer system104 in thesystem100, and updated during flight.

Referring now toFIG. 20, the rectangles labeled A, B, and C representimage areas2002,2010,2018, and2004,2012,2020 from two overlapping 3-camera arrays C-B-A (not shown), respectively. Images ofareas2002,2010,2018, and2004,2012,2020 taken by cameras A through C (not shown) and overlapping cameras A′ through C′ (not shown), respectively, are illustrated from an overhead view. Again, similar toFIGS. 3 and 4, because of the “cross-eyed” arrangement, the image ofarea2002 is taken by right camera A, the image ofarea2010 is taken by center/nadir camera B, and the image ofarea2018 is taken by left camera C. Further, the image ofarea2004 is taken by right camera A′, the image ofarea2012 is taken by center camera B′, and the image ofarea2020 is taken by left camera C′. Cameras A through C and overlapping cameras A′ through C′ form arrays (not shown) that are, in most applications, pointed down vertically.

InFIG. 20, the hatched areas labeled A/B and B/C sidelaps represent two overlapping image overlap areas2006,2008 and2014,2016, respectively. The left image overlap areas2006,2008 is where right camera A overlaps with the center/nadir camera B, and where right camera A′ overlaps with the center camera B′, respectively. The right image overlap areas2014 and2016 is where the left camera C overlaps with the center/nadir camera B, and where the left camera C′ overlaps with the center camera B′. In these sidelap areas2006,2008 and2014,2016, respectively, the camera sensor grid bisects each pixel in the overlap areas2006,2008 and2014,2016, which effectively quadruples the image resolution in these areas2006,2008 and2014,2016 via the mechanism of co-mounted, co-registered oversampling. In effect, the improvement in image/sensor resolution is doubled in each dimension, or 2×2=4 times. This quadrupling of the image resolution quadruples the alignment precision between adjacent cameras, as discussed above.

By having two overlapping camera arrays, the image resolution is effectively quadrupled again for the overlapping sidelap overlap areas2006,2008 and2014,2016. This results in an astounding overall 64 times improvement insystem100 calibration and camera alignment.

In the overlapping sidelap areas2006 and2008, the overlapping camera sensor grids bisects each pixel in the sidelap areas2006 and2008, which effectively quadruples the image resolution in these areas2006 and2008 via the mechanism of co-mounted, co-registered oversampling. Similarly, in the overlapping sidelap areas2014 and2016, the overlapping camera sensor grids bisects each pixel in the sidelap areas2014 and2016, which effectively quadruples the image resolution in these areas2014 and2016. In effect, the improvement in image/sensor resolution is again doubled in each dimension, or 2×2×2×2×2×2=64 times. This overall 64 times improvement of the image resolution also enhances alignment precision by 64 times between adjacent cameras.

This 64 times improvement of alignment precision between adjacent and corresponding cameras enhances thesystems100 alignment precision for all sensors affixed to a rigid mount plate. Cameras A through C and, optionally, other sensors are affixed to a first rigid mount unit and cameras A′ through C′ and, optionally, other sensors are affixed to a second rigid mount unit, which are each affixed to a rigid mount plate. In particular, when the angular alignment of adjacent and/or corresponding cameras affixed to the first and/or second rigid mount units is improved, the angular alignment of the other sensors is also enhanced. This enhancement of alignment precision for the other sensors affixed to the rigid mount plate also improves the image resolution for those sensors.

By having two overlapping camera arrays, the image resolution is effectively quadrupled for the entire image, not just for the A/B and B/C sidelap overlap areas. Referring now toFIG. 20, the overlapping grid detail labeled “OVERLAPPING GRID 4X” represents overlapping areas2022 and2024 inright images areas2018 and2020, respectively. In the overlapping areas2022 and2024, the overlapping camera sensor grids bisects each pixel in the overlapping areas2022 and2024, which effectively quadruples the image resolution in these areas2022 and2024 via the mechanism of co-mounted, co-registered oversampling. In effect, the improvement in image resolution is doubled in each dimension, or 2×2=4 times.

In a preferred embodiment, one camera array is monochrome, and another camera array is red-green-blue. Even though each array covers different color bands, simple image processing techniques are used so that all color bands realize the benefit of this increased resolution. Another advantage provided by these techniques is that, in the case where one camera array is red-green-blue and the other, overlapping camera array is an infrared or near infrared (or some other bandwidth), which results in a superior multi-spectral image.

Accordingly, all of the improvements (i.e., 4 times) identified for the embodiment ofFIG. 19 discussed above apply to the embodiment ofFIG. 20, however, additional significant enhancements (i.e., 64 times) to thesystems100 calibration precision and overall image resolution may be realized through the two overlapping camera arrays.

FIG. 21 is an illustration of a fore and lateral co-mounted,co-registered oversampling configuration2100 for twocamera arrays112 looking down from a vehicle according to certain embodiments of the present invention. In particular,FIG. 21 is an illustration of a fore and lateral co-mounted,co-registered oversampling configuration2100 for two overlappingcamera arrays112 looking down from a vehicle according to certain embodiments of the present invention showing minimal fore and minimal lateral oversampling. The adjacent cameras overlap a few degrees in thevertical sidelap areas2104,2108,2124 and2128, and the corresponding cameras overlap a few degrees along thehorizontal forelap areas2112,2116 and2120. WhereasFIG. 21 depicts two 3-camera arrays, these subpixel calibration techniques work equally well when utilizing two overlapping camera arrays with any number of camera sensors from 2 to any number of cameras being calibrated.

Similar to the imaging sensors inFIGS. 3 and 4, the camera sensors may be co-registered to calibrate the physical mount angle offset of each sensor relative to each other and/or to the nadir camera. In this embodiment, multiple, i.e., at least two, rigid mount units are affixed to a rigid mount plate and are co-registered. This provides an initial, “close” calibration. These initial calibration parameters may be entered into anonboard computer system104 in thesystem100, and updated during flight.

Referring now toFIG. 21, the rectangles labeled A, B, and C representimage areas2102,2106 and2110 from a 3-camera array C-B-A (not shown), and the rectangles D, E, and F representimage areas2122,2126 and2130 from a 3-camera array F-E-D (not shown), respectively. Images ofareas2102,2106 and2110 taken by cameras A through C (not shown), and images ofareas2122,2126 and2130 taken by cameras D through F (not shown), respectively, are illustrated from an overhead view. Again, similar toFIGS. 3 and 4, because of the “cross-eyed” arrangement, the rear, left image ofarea2102 is taken by rear, right camera A, the rear, center image ofarea2106 is taken by rear, center/nadir camera B, and the rear, right image ofarea2110 is taken by rear, left camera C. Further, the forward, left image ofarea2122 is taken by forward, right camera D, the forward, center image ofarea2126 is taken by forward, center camera E, and the forward, right image ofarea2020 is taken by forward, left camera F. Cameras A through C and overlapping cameras D through F form arrays (not shown) that are, in most applications, pointed down vertically.

InFIG. 21, the vertical hatched areas represent fourimage overlap areas2104,2108,2124 and2128. The rear, leftimage overlap area2104 is where rear, right camera A overlaps with the center/nadir camera B, and the rear, rightimage overlap area2108 is where rear, left camera C overlaps with the center/nadir camera B. The forward, leftimage overlap area2124 is where forward, right camera D overlaps with the center/nadir camera E, and the forward, rightimage overlap area2128 is where forward, left camera F overlaps with the center camera E.

Referring now toFIG. 21, the overlapping grid detail labeled “SIDELAP AREA 4:1” represents overlaping sidelap overlapareas2104,2108 and2124,2128. In these sidelap overlapareas2104,2108,2124 and2128, the camera sensor grid bisects each pixel in theoverlap areas2104,2108,2124 and2128, which effectively quadruples the image resolution in theseareas2104,2108,2124 and2128 via the mechanism of co-mounted, co-registered oversampling. In effect, the improvement in image/sensor resolution is doubled in each dimension, or 2×2=4 times. This quadrupling of the image resolution quadruples the alignment precision between adjacent cameras, as discussed above.

This quadrupling of alignment precision between adjacent cameras improves thesystems100 alignment precision for all sensors affixed to a rigid mount plate. Cameras A through C and, optionally, other sensors are affixed to a first rigid mount unit and cameras D through F and, optionally, other sensors are affixed to a second rigid mount unit, which are each affixed to a rigid mount plate. In particular, when the angular alignment of adjacent cameras affixed to the first or second rigid mount units is improved, the angular alignment of the other sensors affixed to the mount unit is also enhanced. This enhancement of alignment precision for the other sensors affixed to the rigid mount plate also improves the image resolution for those sensors.

Similarly, the horizontal hatched areas represent threeimage overlap areas2112,2116 and2120. The forward, leftimage overlap area2112 is where rear, right camera A overlaps with the forward, right camera D, forward, center image overlap area2116 is where rear, center/nadir camera B overlaps with the forward, center camera E, and the rear, right image overlap area2120 is where rear, left camera C overlaps with forward, left camera F.

Referring now toFIG. 21, the overlapping grid detail labeled “FORELAP AREA 4:1” represents overlaping forelap overlapareas2112,2116 and2120. In these forelap overlapareas2112,2116 and2120, the camera sensor grid bisects each pixel in theoverlap areas2112,2116 and2120, which effectively quadruples the image resolution in theseareas2112,2116 and2120 via the mechanism of co-mounted, co-registered oversampling. In effect, the improvement in image/sensor resolution is doubled in each dimension, or 2×2=4 times. This quadrupling of the image resolution quadruples the alignment precision between corresponding cameras.

This quadrupling of alignment precision between corresponding cameras improves thesystems100 alignment precision for all sensors affixed to a rigid mount plate. Cameras A through C and, optionally, other sensors are affixed to a first rigid mount unit and cameras D through F and, optionally, other sensors are affixed to a second rigid mount unit, which are each affixed to a rigid mount plate. In particular, when the angular alignment of corresponding cameras affixed to the first or second rigid mount units is improved, the angular alignment of the other sensors is also enhanced. This enhancement of alignment precision for the other sensors affixed to the rigid mount plate also improves the image resolution for those sensors.

Similar to the overlapping sidelap overlap areas2006,2008 and2014,2016 inFIG. 20, the intersecting forelap and sidelap overlapareas2114 and2118 inFIG. 21 results in an astounding overall 64 times improvement in system calibration and camera alignment. Referring now toFIG. 21, the intersecting grid detail labeled “QUAD OVERLAP AREA 64:1” represents intersecting forelap andsidelap overlap area2118. In the intersecting forelap and sidelap overlapareas2114 and2118, the overlapping camera sensor grids bisects each pixel in theintersecting areas2114 and2118, which effectively quadruples the image resolution in theseareas2114 and2118 via the mechanism of co-mounted, co-registered oversampling. In effect, the improvement in image/sensor resolution is again doubled in each dimension, or 2×2×2×2×2×2=64 times. This overall 64 times improvement of the image resolution also enhances alignment precision by 64 times between adjacent cameras.

This 64 times improvement of alignment precision between adjacent and corresponding cameras enhances thesystems100 alignment precision for all sensors affixed to a rigid mount plate. Cameras A through C and, optionally, other sensors are affixed to a first rigid mount unit and cameras D through E and, optionally, other sensors are affixed to a second rigid mount unit, which are each affixed to a rigid mount plate. In particular, when the angular alignment of adjacent and/or corresponding cameras affixed to the first and/or second rigid mount units is improved, the angular alignment of the other sensors is also enhanced. This enhancement of alignment precision for the other sensors affixed to the rigid mount plate also improves the image resolution for those sensors.

In a preferred embodiment, one camera array is monochrome, and another camera array is red-green-blue. Even though each array covers different color bands, simple image processing techniques are used so that all color bands realize the benefit of this increased resolution. Another advantage provided by these techniques is that, in the case where one camera array is red-green-blue and the other, overlapping camera array is an infrared or near infrared (or some other bandwidth), which results in a superior multi-spectral image.

As shown inFIGS. 19-21, these techniques may be used to overcome the resolution limits imposed on camera systems due to the inability of optical glass to resolve “very small” objects. In particular, there are known physical limits to the ability of optical glass in camera lenses to resolve very small objects. This is often called “the resolving limit of glass”. For example, if 1 millimeter pixels are required from 10,000 feet of altitude, the use of an extremely high magnification telescopic lens would be required to obtain a ground swath of about 100 feet. This is because no matter how many pixels can be produced by a charged-coupled device sensor (e.g., 1 billion pixels), the resolving power of the purest glass would not permit image resolution to 1 millimeter pixels at 10,000 feet of altitude. This example is used to make the point that there are physical limits for pixel resolution in glass as well as pixel density limits for an imaging sensor.

Thesystems100 imaging sensor alignment in the rigid mount unit(s) affixed to the rigid mount plate and related calibration techniques provide a unique solution to this problem, as described above. By using these techniques, the resolving limitations of glass can effectively be overcome. For example, a single camera array results in 1 times (or no) oversampling benefits. However, two overlapping camera arrays results in 4 times overall improvement in both image resolution and overall geospatial horizontal and vertical accuracy. Further, three overlapping camera arrays results in 16 times overall improvement, four overlapping camera arrays results in 64 times overall improvement, and so on.

As can be deduced from these examples, the equation for overall improvement is as follows:
overall improvement=4^N
where N is the number of overlapping camera arrays.

If there are four camera arrays, then there are three overlapping camera arrays (i.e., N=3). Accordingly, four camera arrays provide a 64 times (i.e., 4³=64 times) overall improvements in both the image resolution and overall geospatial horizontal and vertical accuracy.

Further, these subpixel calibration techniques may be combined with the self-locking flight path techniques, as disclosed in U.S. Publication No. 2004/0054488A1, now U.S. Pat. No. 7,212,938B2, the disclosure of which is hereby incorporated by reference in full.

In addition to fore and/or lateral co-mounted, co-registered oversampling as shown inFIGS. 19-21, the present invention may also employ flight line oversampling as well to further improve the image resolution, as shown inFIGS. 13-17. As shown inFIGS. 13-17, the flight lines overlap each other in an image region because each flight line is parallel to one another. These overlapping image regions may be used to calibrate the sensors by along-track and cross-track parallax of images in adjacent flight lines using stereographic techniques.

In an embodiment, the self-locking flight path may comprise any pattern that produces at least three substantially parallel travel lines out of a group of three or more travel lines. Further, at least one of the travel lines should be in an opposing direction to the other substantially parallel travel lines. In a preferred embodiment, the travel pattern comprises at least one pair of travel lines in a matching direction and at least one pair of travel lines in an opposing direction.

When using the self-locking flight path in opposite directions, the observable positional error may be doubled in some image regions. According, the self-locking flight path technique includes an algorithm to significantly reduce these positional errors. This reduction in positional errors is especially important in the outside, or far left and far right “wing” image areas where the greatest positional errors occur.

In an embodiment, these positional improvements may be realized by using a pattern matching technique to automatically match a pixel pattern area obtained from a flight line (e.g., North/South) with the same pixel pattern area obtained from an adjacent flight line (e.g., North/South). In a preferred embodiment, the latitude/longitude coordinates from one or more GPS location systems may be used to accelerate this pattern matching process.

Similarly, these subpixel calibration and self-locking flight path techniques may be combined with stereographic techniques because stereographic techniques rely heavily on the positional accuracy of each pixel relative to all other pixels. In particular, these techniques improve the stereographic image resolution and overall geospatial horizontal and vertical accuracy, especially, in the far left and far right “wing” image areas, where the greatest positional errors occur. Further, stereographic techniques are used to match known elevation data with the improved stereographic datasets. Accordingly, the combined subpixel calibration, self-locking flight path and stereographic techniques provide a greatly improved Digital Elevation Model, which results in superior image.

Further, these subpixel calibration and self-locking flight path techniques may be used to provide a dynamic, RealTime calibration of thesystem100. In particular, these techniques provide the ability to rapidly “roll on” one or morecamera array assemblies112 onto thesystem100, to immediately begin collecting image data of a target area and to quickly produce high-quality images because the individual sensors have been initially calibrated in the rigid mount unit(s) affixed to the rigid mount plate, as discussed above. In particular, the camera sensors are co-registered to calibrate the physical mount angle offset of each sensor relative to each other and/or to the nadir camera. In an embodiment, multiple, i.e., at least two, rigid mount units are affixed to a rigid mount plate and are co-registered. This provides an initial, “close” calibration. These initial calibration parameters may be entered into anonboard computer system104 in thesystem100, and updated during flight using oversampling techniques, as discussed above.

In an embodiment, thesystem100 comprises a RealTime, self-calibrating system to update the calibration parameters. In particular, theonboard computer104 software comprises a RealTime software “daemon” (i.e., a background closed-loop monitoring software) to constantly monitor and update the calibration parameters using the co-mounted, co-registered oversampling and flight line oversampling techniques, as discussed above. In a preferred embodiment, the RealTime daemon combines subpixel calibration, self-locking flight path and stereographic techniques to improve the stereographic image resolution and overall geospatial horizontal and vertical accuracy. In particular, stereographic techniques are used to match known elevation data to the improved stereographic datasets. Accordingly, the combined subpixel calibration, self-locking flight path and stereographic techniques provide a greatly improved Digital Elevation Model, which results in superior image.

In an embodiment, thesystem100 comprises a RealTime GPS data system to provide GPS input data. Calibration accuracy is driven by input data from electronic devices such as a GPS and an IMU, and by calibration software which is augmented by industry standard GPS and IMU software systems. Accordingly, a key component of this RealTime, self-calibrating system is a RealTime GPS input data via a potentially low bandwidth communication channel such as satellite phone, cell phone, RF modem, or similar device. Potential sources for the RealTime GPS input data include project controlled ad-hoc stations, fixed broadcast GPS locations (or similar) or inertial navigation via an onboard IMU.

The modules, algorithms and processes described above can be implemented in a number of technologies and configurations. Embodiments of the present invention may comprise functional instances of software or hardware, or combinations thereof. Furthermore, the modules and processes of the present invention may be combined together in a single functional instance (e.g., one software program), or may comprise operatively associated separate functional devices (e.g., multiple networked processor/memory blocks). All such implementations are comprehended by the present invention.

The embodiments and examples set forth herein are presented to best explain the present invention and its practical application and to thereby enable those skilled in the art to make and utilize the invention. However, those skilled in the art will recognize that the foregoing description and examples have been presented for the purpose of illustration and example only. The description as set forth is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are possible in light of the above teaching without departing from the spirit and scope of the following claims.

Claims (77)

What is claimed is:

1. A system for generating a map of a target area, comprising:

a global positioning receiver;

an imaging sensor system having a view of the target area, comprising:

a rigid mount unit having at least two imaging sensors disposed within the mount unit,

wherein a first imaging sensor and a second imaging sensor each has a focal axis passing through an aperture in the mount unit,

wherein the first imaging sensor generates a first image area comprising a first data array of pixels and the second imaging sensor generates a second image area comprising a second data array of pixels,

wherein the first and second imaging sensors are offset to have a first image overlap area in the target area,

wherein the first sensors image data bisects the second sensors image data in the first image overlap area; and

a computer in communication with the a global positioning antenna, the first imaging sensor, and the second imaging sensor; correlating at least a portion of the image areas from the first imaging sensor and the second imaging sensor to a portion of the target area based on input from the global positioning antenna.

2. The system ofclaim 1 further comprising:

a third imaging sensor disposed within the mount unit, wherein the third imaging sensor has a focal axis passing through the aperture in the mount unit, wherein the third imaging sensor generates a third image area comprising a third data array of pixels.

3. The system ofclaim 2, further comprising:

a fourth imaging sensor disposed within the mount unit, wherein the fourth imaging sensor has a focal axis passing through the aperture in the mount unit, wherein the fourth imaging sensor generates a fourth image area comprising a fourth data array of pixels, wherein the third and fourth imaging sensors are offset to have a second image overlap area in the target area, wherein the third sensors image data bisects the fourth sensors image data in the second image overlap area.

4. The system ofclaim 3, wherein a first sensor array comprising the first and second image sensors and a second sensor array comprising the third and fourth image sensors are offset to have a third image overlap area in the target area, wherein the first sensor arrays image data bisects the second sensor arrays image data in the third overlap area.

5. The system ofclaim 3, wherein the first sensors arrays image data completely overlaps the second sensors arrays image data.

6. The system ofclaim 3, wherein third and fourth imaging sensors are selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

7. The system ofclaim 3, wherein the first and second imaging sensors are a digital camera and the third imaging sensor is a light detection and ranging (LIDAR).

8. The system ofclaim 2, wherein the third imaging sensor is selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

9. The system ofclaim 2, wherein the third imaging sensor is selected from the group consisting of a digital camera having a hyperspectral filter and a light detection and ranging (LIDAR).

10. The system ofclaim 2, wherein the first and second imaging sensors are a digital camera and the third imaging sensor is a light detection and ranging (LIDAR).

11. The system ofclaim 1, wherein the mount unit flexes less than 100th of a degree during operation.

12. The system ofclaim 11, wherein the mount unit flexes less than 1,000th of a degree during operation.

13. The system ofclaim 12, wherein the mount unit flexes less than 10,000th of a degree during operation.

14. The system ofclaim 1, wherein the first imaging sensor is calibrated relative to one or more attitude measuring devices selected from the group consisting of a gyroscope, an inertial measurement unit (IMU), and a global positioning system (GPS).

15. The system ofclaim 1, wherein the first and second imaging sensors are selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

16. An imaging sensor system comprising:

a mount unit in alignment with a target area, having at least two imaging sensors disposed within the mount unit, wherein a first imaging sensor and a second imaging sensor each has a focal axis passing through an aperture in the mount unit, wherein the first imaging sensor generates a first image area comprising a first data array of pixels and the second imaging sensor generates a second image area comprising a second data array of pixels, wherein the first and second imaging sensors are offset to have a first image overlap area in the target area, wherein the first sensors image data bisects the second sensors image data in the first image overlap area.

17. The system ofclaim 16 further comprising:

a third imaging sensor disposed within the mount unit, wherein the third imaging sensor has a focal axis passing through the aperture in the mount unit, wherein the third imaging sensor generates a third image area comprising a third data array of pixels.

18. The system ofclaim 17 further comprising:

a fourth imaging sensor disposed within the mount unit, wherein the fourth imaging sensor has a focal axis passing through the aperture in the mount unit, wherein the fourth imaging sensor generates a fourth image area comprising a fourth data array of pixels, wherein the third and fourth imaging sensors are offset to have a second image overlap area in the target area, wherein the third sensors image data bisects the fourth sensors image in the second image overlap area.

19. The system ofclaim 18, wherein a first sensors array comprising the first and the second image sensor and a second sensors array comprising the third and the fourth image sensor are offset to have a third image overlap area in the target area, wherein first sensor arrays image data bisects the second sensor arrays image data in the third image overlap area.

20. The system ofclaim 18, wherein the first sensors arrays image data completely overlaps the second sensors arrays image data.

21. The system ofclaim 18, wherein the third and fourth imaging sensors are selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

22. The system ofclaim 18, wherein the first and second imaging sensors are a digital camera and the third imaging sensor is a light detection and ranging (LIDAR).

23. The system ofclaim 17, wherein the third imaging sensor is selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

24. The system ofclaim 17, wherein the third imaging sensor is selected from the group consisting of a digital camera having a hyperspectral filter and a light detection and ranging (LIDAR).

25. The system ofclaim 17, wherein the first and second imaging sensors are a digital camera and the third imaging sensor is a light detection and ranging (LIDAR).

26. The system ofclaim 16, wherein the mount unit flexes less than 100th of a degree during operation.

27. The system ofclaim 26, wherein the mount unit flexes less than 1,000th of a degree during operation.

28. The system ofclaim 27, wherein the mount unit flexes less than 10,000th of a degree during operation.

29. The system ofclaim 16, wherein the first imaging sensor is calibrated relative to one or more attitude measuring devices selected from the group consisting of a gyroscope, an inertial measurement unit (IMU), and a global positioning receiver (GPS).

30. The system ofclaim 16, wherein the first and second imaging sensors are selected from the group consisting of digital cameras, light detection and ranging (LIDAR), infrared, heat-sensing and gravitometers.

31. A method of calibrating imaging sensors comprising the steps of:

performing an initial calibration of the imaging sensors comprising:

determining the position of an attitude measurement unit (AMU) selected from the group consisting of a gyroscope, an inertial measurement unit (IMU), and a global positioning system (GPS);

determining the position of a first imaging sensor within a rigid mount unit relative to the AMU;

determining the position of a second imaging sensor within the rigid mount unit relative to the AMU;

calibrating the first imaging sensor against a target area and determining a boresight angle of the first imaging sensor; and

calculating the position of one or more subsequent imaging sensors within the rigid mount unit relative to the first imaging sensor; and

calibrating the one or more subsequent imaging sensors using the boresight angle of the first imaging sensor; and

using oversampling techniques to update at least one initial calibration parameter of the first imaging sensor against a target area and the boresight angle of the first imaging sensor;

using oversampling techniques to update the position of one or more subsequent imaging sensors within the rigid mount unit relative to the first imaging sensor; and

updating at least one calibration parameter of one or more subsequent imaging sensors within the rigid mount using the updated boresight angle of the first imaging sensor.

32. The method ofclaim 31, wherein the initial calibration step further comprises the step of:

calibrating the second imaging sensor using the updated boresight angle of the first imaging sensor.

33. The method ofclaim 32, further comprising the step of:

using oversampling techniques to update the position of the second imaging sensor within the rigid mount unit relative to the first imaging sensor.

34. The method ofclaim 31, further comprising the steps of:

using flight line oversampling techniques to update the calibration of the first imaging sensor against a target area and the boresight angle of the first imaging sensor; and

using flight line oversampling techniques to update the position of one or more subsequent imaging sensors within the rigid mount unit relative to the first imaging sensor.

35. The method ofclaim 34, further comprising the steps of:

using flight line oversampling techniques to update the position of the second imaging sensor within the rigid mount unit relative to the first imaging sensor;

using flight line oversampling techniques to update the position of one or more subsequent imaging sensors within the rigid mount unit relative to the first imaging sensor; and

updating at least one calibration parameter of one or more subsequent imaging sensors within the rigid mount using the updated boresight angle of the first imaging sensor.

36. A system for generating a map of a surface, comprising:

a global position receiver;

a global positioning antenna;

an imaging array, having a view of the surface, comprising:

a mount unit;

an aperture, formed in the mount unit;

a first imaging sensor, coupled to the mount unit, having a first focal axis passing through the aperture, wherein the first image sensor generates a first image area of the surface comprising a first data array of pixels, wherein the first data array of pixels is at least two dimensional; and

a second imaging sensor, coupled to the mount unit and offset from the first imaging sensor, having a second focal axis passing through the aperture and intersecting the first focal axis, wherein the second imaging sensor generates a second image area of the surface comprising a second data array of pixels, wherein the second data array of pixels is at least two dimensional; and

a computer, connected to the global positioning antenna, and first and second imaging sensors; correlating at least a portion of the image area from the first and second imaging sensors to a portion of the surface based on input from the global positioning antenna.

37. The system ofclaim 36, further comprising a third imaging sensor, coupled to the mount unit and offset from the first imaging sensor, having a third focal axis passing through the aperture and intersecting the first focal axis within an intersection area.

38. The system ofclaim 37, wherein the focal axes of the third imaging sensor lies in a common plane with the focal axes of the first and second imaging sensors.

39. The system ofclaim 37, wherein the focal axes of the first and second imaging sensors lie in a first common plane and the focal axis of the third imaging sensor lies in a plane orthogonal to the first common plane.

40. A system for generating a map of a surface, comprising:

a global position receiver;

a global positioning antenna;

a first imaging sensor, having a view of the surface, having a focal axis disposed in the direction of the surface, wherein the first imaging sensor generates an image area comprising a first data array of pixels, wherein the first data array of pixels is at least two dimensional; and

a computer, connected to the global positioning antenna, and the first imaging sensor; generating a calculated longitude and calculated latitude value for a coordinate corresponding to at least one pixel in the array based on input from the global positioning antenna.

41. A system for generating a map of a target area, comprising:

a global position receiver;

a global positioning antenna;

an imaging sensor system, having a view of the target area, comprising:

a mount unit, having a first and second imaging sensor disposed within the mount unit, wherein the first and second imaging sensors each have a focal axis passing through an aperture in the mount unit, wherein the first imaging sensor generates a first image area comprising a first data array of pixels and second imaging sensor generates a second image area comprising a second data array of pixels, wherein the first and second data array of pixels is at least two dimensional; and

a computer in communication with the global positioning antenna, the first imaging sensor, and the second imaging sensor; correlating at least a portion of the image area from the first imaging sensor and the second imaging sensor to a portion of the target area based on input from the global positioning antenna.

42. The system ofclaim 41, further comprising a third imaging sensor disposed within the mount unit, wherein the third imaging sensor has a focal axis passing through an aperture in the mount unit, wherein the third imaging sensor generates a third image area comprising a third data array of pixels.

43. An imaging sensor system comprising:

a mount unit, having a first and second imaging sensors disposed within the mount unit, wherein the first imaging and second imaging sensors each have a focal axis passing through an aperture in the mount unit, wherein the first imaging sensor generates a first image area comprising a first data array of pixels and the second imaging sensor generates a second image area comprising a second data array of pixels, wherein the first and second data array of pixels is at least two dimensional.

44. The system ofclaim 43, further comprising a third imaging sensor disposed within the mount unit, wherein the third imaging sensor has a focal axis passing through an aperture in the mount unit, wherein the third imaging sensor generates a third image area comprising a third data array of pixels.

45. A system for generating a map of a target area, comprising:

a global positioning receiver;

an imaging sensor system having a view of the target area, comprising:

a rigid mount unit having a first imaging sensor and a second imaging sensor,

wherein the first imaging sensor generates a first image area comprising a first data array of pixels and the second imaging sensor generates a second image area comprising a second data array of pixels,

wherein the first and second imaging sensors are offset to have a first image overlap area in the target area,

wherein the first data array of pixels bisects the second data array of pixels in the first image overlap area; and

a computer in communication with the global positioning receiver, the first imaging sensor, and the second imaging sensor, wherein at least a portion of the first image area from the first imaging sensor is correlated to a portion of the target area based on input from the global positioning receiver or other geographical positioning technique.

46. The system of claim 45, wherein the first image overlap area comprises at least one oversampling pattern.

47. The system of claim 46, wherein the at least one oversampling pattern comprises a plurality of pixels that have been positioned using an oversampling technique to a precision that is less than one pixel in magnitude.

48. The system of claim 47, wherein the mount unit flexes less than 100th of a degree during operation.

49. The system of claim 47, wherein the mount unit flexes less than 1,000th of a degree during operation.

50. The system of claim 47, wherein the mount unit flexes less than 10,000th of a degree during operation.

51. An imaging sensor system comprising:

a rigid mount unit in alignment with a target area;

a first imaging sensor rigidly connected to the mount unit, wherein the first imaging sensor generates a first image area comprising a first data array of pixels;

a second imaging sensor rigidly connected to the mount unit, wherein the second imaging sensor generates a second image area comprising a second data array of pixels;

wherein the first and second imaging sensors are offset to have a first image overlap area in the target area,

wherein the first data array of pixels bisects the second data array of pixels in the first image overlap area.

52. The system of claim 51, wherein the first image overlap area comprises at least one oversampling pattern.

53. The system of claim 52, wherein the at least one oversampling pattern comprises a plurality of pixels that have been positioned using an oversampling technique to a precision that is less than one pixel in magnitude.

54. A method of calibrating imaging sensors comprising the steps of:

performing an initial calibration of the imaging sensors comprising:

determining the position of an attitude measurement unit (AMU) selected from the group consisting of a gyroscope, an inertial measurement unit (IMU), and a global positioning system (GPS);

determining the position of a first imaging sensor relative to the AMU;

calibrating the first imaging sensor against a target area and determining a boresight angle of the first imaging sensor; and

determining the position of the second imaging sensor relative to the first imaging sensor; and

calibrating the second imaging sensor using the boresight angle of the first imaging sensor; and

using oversampling techniques to update at least one initial calibration parameter of the first imaging sensor against a target area and the boresight angle of the first imaging sensor;

using oversampling techniques to update the position of the second imaging sensor relative to the first imaging sensor; and

updating at least one calibration parameter of the second imaging sensor using the updated boresight angle of the first imaging sensor.

55. The method of claim 54, wherein the oversampling techniques comprise flight line oversampling.

56. The method of claim 54, wherein the oversampling techniques comprise lateral oversampling.

57. The method of claim 54, wherein the oversampling techniques comprise generating at least one pixel that has been positioned using the oversampling techniques to a precision that is less than one pixel in magnitude.

58. The method of claim 54, wherein the oversampling techniques comprise generating a plurality of pixels that have been positioned using the oversampling techniques to a precision that is less than one pixel in magnitude.

59. A system for generating an image of a surface, comprising:

a global position receiver;

an imaging array, having a view of the surface, comprising:

a mount unit;

a first imaging sensor, coupled to the mount unit, having a first focal axis passing through an aperture in the mount unit,

wherein the first image sensor generates a first image area of the surface comprising a first data array of pixels,

wherein the first data array of pixels is at least two dimensional; and

a second imaging sensor, coupled to the mount unit and offset from the first imaging sensor, having a second focal axis passing through an aperture in the mount unit,

wherein the second imaging sensor generates a second image area of the surface comprising a second data array of pixels,

wherein the second data array of pixels is at least two dimensional; and

a computer, connected to the global position receiver, the first imaging sensor, and the second imaging sensor, wherein at least a portion of the first image area from the first imaging sensor is correlated to a portion of the surface based on input from the global position receiver.

60. The system of claim 59, wherein the mount unit flexes less than 100th of a degree during operation.

61. The system of claim 59, wherein the mount unit flexes less than 1,000th of a degree during operation.

62. The system of claim 59, wherein the mount unit flexes less than 10,000th of a degree during operation.

63. A system for generating an image of a surface, comprising:

a global position receiver;

a first imaging sensor adapted to view a surface and disposed at least partially in a mount unit, wherein the mount unit flexes less than 100th of a degree during operation,

wherein the first imaging sensor generates a first image area comprising a first data array of pixels,

wherein the first data array of pixels is at least two dimensional; and

a computer connected to the global position receiver and the first imaging sensor, wherein a calculated longitude value and a calculated latitude value are generated for at least one pixel in the first data array of pixels based on input from the global position receiver.

64. The system of claim 63, wherein the mount unit flexes less than 1,000th of a degree during operation.

65. The system of claim 63, wherein the mount unit flexes less than 10,000th of a degree during operation.

66. A system for generating an image, comprising:

a rigid mount unit;

a first imaging sensor disposed at least partially within the rigid mount unit, wherein the first imaging sensor generates a first image area comprising a first two-dimensional data array of pixels;

a second imaging sensor disposed at least partially within the rigid mount unit, wherein the second imaging sensor generates a second image area comprising a second two-dimensional data array of pixels;

wherein the first imaging sensor and the second imaging sensor are offset such that the first image area overlaps with the second image area to form a first image overlap area;

a computer in communication with the first imaging sensor and the second imaging sensor;

a mosaicking module associated with the computer for balancing the color of the second two-dimensional data array of pixels based on the average intensity of green-dominant pixels in the first two-dimensional data array of pixels.

67. The system of claim 66, wherein the first image overlap area comprises at least one oversampling pattern.

68. The system of claim 67, wherein the at least one oversampling pattern comprises a plurality of pixels that have been positioned using an oversampling technique to a precision that is less than one pixel in magnitude.

69. The system of claim 66, wherein the mount unit flexes less than 100th of a degree during operation.

70. The system of claim 66, wherein the mount unit flexes less than 1,000th of a degree during operation.

71. The system of claim 66, wherein the mount unit flexes less than 10,000th of a degree during operation.

72. An imaging sensor system comprising:

a rigid mount unit;

a first imaging sensor disposed at least partially within the rigid mount unit, wherein the first imaging sensor generates a first image area comprising a first two-dimensional data array of pixels;

a second imaging sensor disposed at least partially within the rigid mount unit, wherein the second imaging sensor generates a second image area comprising a second two-dimensional data array of pixels;

wherein the first imaging sensor and the second imaging sensor are offset such that the first image area overlaps with the second image area to form a first image overlap area;

a computer in communication with the first imaging sensor and the second imaging sensor; and

an intensity balancing module associated with the computer for balancing the intensity of the first two-dimensional data array of pixels and the second two-dimensional data array of pixels using a balancing correlation matrix.

73. The system of claim 72, wherein the first image overlap area comprises at least one oversampling pattern.

74. The system of claim 73, wherein the at least one oversampling pattern comprises a plurality of pixels that have been positioned using an oversampling technique to a precision that is less than one pixel in magnitude.

75. The system of claim 72, wherein the mount unit flexes less than 100th of a degree during operation.

76. The system of claim 72, wherein the mount unit flexes less than 1,000th of a degree during operation.

77. The system of claim 72, wherein the mount unit flexes less than 10,000th of a degree during operation.

Images