US20040012573A1 - Passive touch system and method of detecting user input - Google Patents

Abstract

A passive touch system includes a passive touch surface and at least two cameras associated with the touch surface. The at least two cameras acquire images of the touch surface from different locations and have overlapping fields of view. A processor receives and processes images acquired by the at least two cameras to detect the existence of a pointer therein and to determine the location of the pointer relative to the touch surface. Actual pointer contact with the touch surface and pointer hover above the touch surface can be determined.

Description

FIELD OF THE INVENTION
• The present invention relates generally to input devices and in particular to a passive touch system and method of detecting user input.[0001]
BACKGROUND OF THE INVENTION
• Touch systems are well known in the art and typically include a touch screen having a touch surface on which contacts are made using a pointer in order to generate user input. Pointer contacts with the touch surface are detected and are used to generate corresponding output depending on areas of the contact surface where the contacts are made. There are basically two general types of touch systems available and they can be broadly classified as “active” touch systems and “passive” touch systems.[0002]
• Active touch systems allow a user to generate user input by contacting the touch surface with a special pointer that usually requires some form of on-board power source, typically batteries. The special pointer emits signals such as infrared light, visible light, ultrasonic frequencies, electromagnetic frequencies, etc. that activate the touch surface.[0003]
• Passive touch systems allow a user to generate user input by contacting the touch surface with a passive pointer and do not require the use of a special pointer in order to activate the touch surface. A passive pointer can be a finger, a cylinder of some material, or any suitable object that can be used to contact some predetermined area of interest on the touch surface.[0004]
• Passive touch systems provide advantages over active touch systems in that any suitable pointing device, including a user's finger, can be used as a pointer to contact the touch surface. As a result, user input can easily be generated. Also, since special active pointers are not necessary in passive touch systems, battery power levels and/or pointer damage, theft, or misplacement are of no concern to users.[0005]
• Passive touch systems have a number of applications relating to computer operation and video display. For example, in one interactive application, as is disclosed in U.S. Pat. No. 5,448,263 to Martin, assigned to the assignee of the present invention, the passive touch system is coupled to a computer and the computer display is projected onto the touch surface of the touch screen. The coordinates representing specific locations on the touch surface are mapped to the computer display. When a user contacts the touch surface, the coordinates of the contact are fed back to the computer and mapped to the computer display thereby allowing the user to operate the computer in a manner similar to using a computer mouse simply by contacting the touch surface. Furthermore, the coordinates fed back to the computer can be recorded in an application and redisplayed at a later time. Recording contact coordinates is typically done when it is desired to record information written or drawn on the touch surface by the user.[0006]
• The resolution of a passive touch screen determines if the touch system is suitable for recording information written or drawn on the touch screen or only useful for selecting areas on the touch screen mapped to large regions on the computer or video display in order to manipulate the computer or video display. Resolution is typically measured in dots per inch (DPI). The DPI is related to the size of the touch screen and the sampling ability of the touch system hardware and software used to detect contacts on the touch surface.[0007]
• Low-resolution passive touch screens only have enough DPI to detect contacts on the touch surface within a large group of pixels displayed by the computer or video system. Therefore, these low-resolution passive touch screens are useful only for manipulating the computer or video display.[0008]
• On the other hand, high-resolution passive touch screens have sufficient DPI to detect contacts that are proportional to a small number of pixels or sub-pixels of the computer or video display. However, a requirement for high-resolution touch screens is the ability to detect when the pointer is in contact with the touch surface. This is necessary for writing, drawing, mouse-click operations, etc. Without the ability to detect pointer contact with the touch screen, writing and drawing would be one continuos operation, and mouse clicks would not be possible thereby making computer display manipulation impossible. A secondary requirement is the ability to detect when the pointer is “hovering” above the touch surface. Although not required for writing or drawing, today's computer operating systems are increasingly using hover information to manipulate computer or video displays or pop-up information boxes.[0009]
• Passive touch screens are typically either of the analog resistive type, Surface Acoustic Wave (SAW) type or capacitive type. Unfortunately, these touch screens suffer from a number of problems or shortcomings as will be described.[0010]
• Analog resistive touch screens typically have a high-resolution. Depending on the complexity of the touch system, the resolution of the touch screen can produce 4096×4096 DPI or higher. Analog resistive touch screens are constructed using two flexible sheets that are coated with a resistive material and arranged as a sandwich. The sheets do not come into contact with each other until a contact has been made. The sheets are typically kept separated by insulating microdots or by an insulating air space. The sheets are constructed from ITO, which is mostly transparent. Thus, the touch screen introduces some image distortion but very little parallax.[0011]
• During operation of an analog resistive passive touch screen, a uniform voltage gradient is applied in one direction along a first of the sheets. The second sheet measures the voltage along the first sheet when the two sheets contact one another as a result of a contact made on the touch surface. Since the voltage gradient of the first sheet can be translated to the distance along the first sheet, the measured voltage is proportional to the position of the contact on the touch surface. When a contact coordinate on the first sheet is acquired, the uniform voltage gradient is then applied to the second sheet and the first sheet measures the voltage along the second sheet. The voltage gradient of the second sheet is proportional to the distance along the second sheet. These two contact coordinates represent the X-Y position of the contact on the touch surface in a Cartesian coordinate system.[0012]
• Since mechanical pressure is required to bring both sheets into contact, analog resistive touch screens can only detect contact when there is sufficient pressure to bring the two sheets together. Analog resistive passive touch screens cannot sense when a pointer is hovering over the touch surface. Therefore, contact events and positions can only be detected when actual contacts are made with the touch surface.[0013]
• Surface Acoustic Wave (SAW) touch screens typically provide for medium resolution and are not suitable for recording good quality writing. SAW touch screens employ transducers on the borders of a glass surface to vibrate the glass and produce acoustic waves that ripple over the glass surface. When a contact is made on the glass surface, the waves reflect back and the contact position is determined from the signature of the reflected waves.[0014]
• Unfortunately, SAW touch screens exhibit noticeable parallax due to the thickness of the vibrating glass which is placed over the surface of the video or computer display. Also, contact events and positions can only be detected when actual contacts are made with the glass surface. Furthermore, SAW touch screens do not scale beyond a few feet diagonal.[0015]
• Capacitive touch screens provide for low resolution because contacts can only be determined in large areas (approximately ½″×½″). As a result, capacitive touch screens cannot be used for recording writing or drawing and are suitable for selecting areas on the touch screen corresponding to computer generated buttons displayed on the video or computer display. These touch screens also suffer disadvantages in that they are sensitive to temperature and humidity. Similar to analog resistive touch screens and SAW touch screens, capacitive touch screens can also only detect contact events and positions when actual contacts are made with the touch surface.[0016]
• Scalability of passive touch screens is important since the demand for larger electronic digitizers is increasing. Where digitizers were once small desktop appliances, today they have found there way onto electronic whiteboarding applications. The need to build a passive touch sensitive “wall” has become a requirement for new touch screen applications. Existing passive touch screens of the types discussed above are all limited in the maximum size where they are still functional.[0017]
• As will be appreciated, improvements to passive touch systems are desired. It is therefore an object of the present invention to provide a novel passive touch system and method of detecting user input.[0018]
SUMMARY OF THE INVENTION
• According to one aspect of the present invention there is provided a passive touch system comprising:[0019]
• a passive touch surface;[0020]
• at least two cameras associated with said touch surface, said at least two cameras acquiring images of said touch surface from different locations and having overlapping fields of view; and[0021]
• a processor receiving and processing images acquired by said at least two cameras to detect the existence of a pointer therein and to determine the location of said pointer relative to said touch surface.[0022]
• In a preferred embodiment, the at least two cameras are two-dimensional image sensor and lens assemblies having fields of view looking along the plane of the touch surface. The processor determines the location of the pointer relative to the touch screen using triangulation. The processor also determines when the pointer is in contact with the touch surface and when the pointer is hovering above the touch surface.[0023]
• In one embodiment, the processor selects pixel subsets of images acquired by the image sensor and lens assemblies and processes the pixel subsets to determine the existence of the pointer. The processor includes a digital signal processor associated with each image sensor and lens assembly and a master digital signal processor in communication with the digital signal processors. The digital signal processors associated with each image sensor and lens assembly select the pixel subsets and process the pixel subsets to determine the existence of the pointer. The master digital signal processor receives pixel characteristic data from the digital signal processors and triangulates the pixel characteristic data to determine the location of the pointer relative to the touch surface.[0024]
• According to another aspect of the present invention there is provided a method of detecting the position of a pointer relative to a passive touch surface comprising the steps of:[0025]
• acquiring images of said touch surface from different locations using cameras having overlapping fields of view; and[0026]
• processing said images to detect the existence of a pointer therein and to determine the location of said pointer relative to said touch surface.[0027]
• The present invention provides advantages in that the passive touch system is of high resolution and allows actual pointer contacts with the touch surface as well as pointer hovers above the touch surface to be detected and corresponding output generated. Also, the present passive touch system provides advantages in that it does not suffer from parallax, image distortion, pointer position restrictions, image projection and scalability problems that are associated with prior art passive touch systems.[0028]
BRIEF DESCRIPTION OF THE DRAWINGS
• An embodiment of the present invention will now be described more fully with reference to the accompanying drawings in which:[0029]
• FIG. 1 is a schematic diagram of a passive touch system in accordance with the present invention;[0030]
• FIG. 2 is an isometric view of a touch screen forming part of the passive touch system of FIG. 1;[0031]
• FIG. 3 is an isometric view of a corner portion of the touch screen of FIG. 2;[0032]
• FIG. 4 is a schematic diagram of a camera assembly forming part of the touch screen of FIG. 2;[0033]
• FIG. 5 is a front elevational view of the touch screen of FIG. 2 showing the fields of view of two camera assemblies;[0034]
• FIG. 6 is a schematic diagram of a master controller forming part of the passive touch system of FIG. 1;[0035]
• FIG. 7 is case diagram showing functions executed by the camera assemblies;[0036]
• FIG. 8 is a flowchart showing the steps performed during execution of afindPointerMotion( ) function;[0037]
• FIG. 9 is a flowchart showing the steps performed during execution of an autoSelectThres( ) function;[0038]
• FIG. 10 is a flowchart showing the steps performed during execution of an extractPointer function;[0039]
• FIG. 11 is a flowchart showing the steps performed during execution of a centerOfMass( ) function;[0040]
• FIG. 12 is a flowchart showing the steps performed during execution of a processROI( ) function;[0041]
• FIG. 13 is a flowchart showing the steps performed during execution of a getHighestRegion( ) function;[0042]
• FIG. 14 shows an acquired image and a pixel subset of the image that is processed;[0043]
• FIG. 15 shows a region of interest within the pixel subset;[0044]
• FIG. 16 shows the triangulation geometry used to calculate a pointer contact position on the touch surface of the touch screen;[0045]
• FIG. 17 shows an image acquired by an image sensor and lens assembly including the pointer and its median line; and[0046]
• FIG. 18 shows pointer contact and pointer hover for different orientations of the pointer.[0047]
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
• Turning now to FIG. 1, a passive touch system in accordance with the present invention is shown and is generally indicated to by[0048]reference numeral50. As can be seen,passive touch system50 includes atouch screen52 coupled to amaster controller54.Master controller54 is also coupled to acomputer56.Computer56 executes one or more application programs and provides display output that is projected onto thetouch screen52 via aprojector58. Thetouch screen52,master controller54,computer56 andprojector58 form a closed-loop so that user contacts with thetouch screen52 can be recorded as writing or drawing or used to control execution of application programs executed by thecomputer56.
• FIGS.[0049]2 to4 better illustrate thetouch screen52.Touch screen52 includes atouch surface60 bordered by aframe62.Touch surface60 is passive and is in the form of a rectangular planar sheet of material. Camera subsystems are associated with each corner of thetouch screen52. Each camera subsystem includes acamera assembly63 mounted adjacent a different corner of thetouch screen52 by aframe assembly64. Eachframe assembly64 includes anangled support plate66 on which thecamera assembly63 is mounted. Supportingframe elements70 and72 are mounted on theplate66 by posts74 and secure theplate66 to theframe62.
• Each[0050]camera assembly63, in this embodiment, includes a camera in the form of a two-dimensional CMOS camera image sensor and associatedlens assembly80, a first-in-first-out (FIFO)buffer82 coupled to the image sensor andlens assembly80 by a data bus and a digital signal processor (DSP)84 coupled to theFIFO82 by a data bus and to the image sensor andlens assembly80 by a control bus. Aboot EPROM86 and apower supply subsystem88 are also included.
• In the present embodiment, the CMOS camera image sensor is a Photobit PB300 image sensor configured for a 20×640 pixel subarray that can be operated to capture image frames at rates in excess of 200 frames per second. The[0051]FIFO buffer82 is manufactured by Cypress under part number CY7C421 IV and theDSP84 is manufactured by Analog Devices under part number ADSP2185M.
• The[0052]DSP84 provides control information to the image sensor andlens assembly80 via the control bus. The control information allows theDSP84 to control parameters of the image sensor andlens assembly80 such as exposure, gain, array configuration, reset and initialization. TheDSP84 also provides clock signals to the image sensor andlens assembly80 to control the frame rate of the image sensor andlens assembly80.
• As shown in FIG. 5, each image sensor and[0053]lens assembly80 has a 55° field of view. The angle of theplate66 is selected so that the field of view of each image andlens assembly80 includes at least the majority of a different peripheral edge of thetouch surface60. In this way, theentire touch surface60 is within the fields of view of the image sensor andlens assemblies80.
• [0054]Master controller54 is best illustrated in FIG. 6 and includes aDSP90, aboot EPROM92, aserial line driver94 and apower supply subsystem95. TheDSP90 communicates with theDSPs84 of thecamera assemblies63 over a data bus through aserial port96 and communicates with thecomputer56 over a data bus through aserial port98 and theserial line driver94. In this embodiment, theDSP90 is also manufactured by Analog Devices under part number ADSP2185M. Theserial line driver94 is manufactured by Analog Devices under part number ADM222.
• The[0055]master controller54 and eachcamera assembly63 follow a communication protocol that enables bi-directional communications via a common serial cable similar to a universal serial bus (USB). The transmission bandwidth is divided into thirty-two (32) 16-bit channels. Of the thirty-two channels, five (5) channels are assigned to each of theDSPs84 in thecamera assemblies63 and to theDSP90 in themaster controller54 and the remaining seven (7) channels are unused. Themaster controller54 monitors the twenty (20) channels assigned to thecamera assembly DSPs84 while theDSPs84 in thecamera subsystems63 monitor the five (5) channels assigned to themaster controller DSP90. Communications between themaster controller54 and thecamera assemblies63 are performed as background processes in response to interrupts.
• The general operation of the[0056]passive touch system50 will now be described. Eachcamera assembly63 acquires images of thetouch surface60 within the field of view of its image sensor andlens assembly80 at the frame rate established by the DSP clock signals and processes the images to determine if a pointer is in the acquired images. If a pointer is in the acquired images, the images are further processed to determine characteristics of the pointer contacting or hovering above thetouch surface60. Pointer characteristics are then converted into pointer information packets (PIPs) and the PIPs are queued for transmission to themaster controller54. Thecamera assemblies63 also receive and respond to diagnostic PIPs generated by themaster controller54.
• The[0057]master controller54 polls thecamera assemblies63 at a set frequency (in thisembodiment70 times per second) for PIPs and triangulates pointer characteristics in the PIPs to determine pointer position data. Themaster controller54 in turn transmits pointer position data and/or status information to thepersonal computer56. In this manner, the pointer position data transmitted to thepersonal computer56 can be recorded as writing or drawing or can be used to control execution of application programs executed by thecomputer56. Thecomputer56 also updates the display output conveyed to theprojector58 so that information projected onto thetouch surface60 reflects the pointer activity.
• The[0058]master controller54 also receives commands from thepersonal computer56 and responds accordingly as well as generates and conveys diagnostic PIPs to thecamera assemblies63.
• Specifics concerning the processing of acquired images and the triangulation of pointer characteristics in PIPs will now be described with particular reference to FIGS.[0059]7 to13.
• Initially, an alignment routine is performed to align the image sensor and[0060]lens assemblies80. During the alignment routine, a pointer is held in the approximate center of thetouch surface60. Subsets of the pixels of the image sensor andlens assemblies80 are then selected until a subset of pixels for each image sensor andlens assembly80 is found that captures the pointer and the pointer tip on thetouch surface60. This alignment routine allows for a relaxation in mechanical mounting of the image sensor and lens assemblies on theframe assemblies64. The identification of the pointer tip on thetouch surface60 also gives a calibration that determines the row of pixels of each image sensor andlens assembly80 that detects actual contacts made with the touch surface. Knowing these pixel rows allows the difference between pointer hover and pointer contact to be determined.
• In this embodiment, since a computer display is projected onto the[0061]touch surface60, during the alignment routine several known coordinate locations are also displayed and the user is required to touch these coordinate locations in sequence using the pointer so that the subset of pixels for each of image sensor andlens assembly80 includes all of the coordinate locations as well. Calibration data is then stored for reference so that pointer contacts on thetouch surface60 can be mapped to corresponding areas on the computer display.
• As mentioned above, each[0062]camera assembly63 acquires images of thetouch surface60 within its field of view. The images are acquired by the image andlens assembly80 at intervals in response to the clock signals received from theDSP84. Each image acquired by the image andlens assembly80 is sent to theFIFO buffer82.
• The[0063]DSP84 in turn reads each image from theFIFO buffer82 and processes the image to determine if a pointer is located in the image and if so, to extract the pointer and related pointer statistical information. To avoid processing significant numbers of pixels containing no useful information, only the subset of the pixels in the image determined during the alignment routine are actually processed as is shown in FIG. 14.
• In order to determine if a pointer is located in the image and extract pointer and related pointer statistical information, the[0064]DSP84 executes a main findPointerMotion( ) function120 that calls a number of other functions, namely an autoSelectThres( )function122, anextractPointer function124, a centerOfMass( )function126, and a processROI( ) function128 (see FIG. 7). The extractPointer( )function128 also calls a getHighestRegion( )function130.
• ThefindPointerMotion( )[0065]function120 is used to extract the pointer from the image. Turning now to FIG. 8, the steps performed during execution of the findPointerMotion( )function120 is shown. When thefindPointerMotion( ) function is called, a check is made to determine if a previous image iPrev including a pointer exists (step150). If no previous image iPrev exists, center of mass parameters Cx and Cz are assigned zero values (step152). The current image iCurr being processed is then designated as the previous image iPrev (step154) to complete the findPointerMotion( ) function.
• At[0066]step150, if a previous image iPrev exists, the current image iCurr is subtracted from the previous image iPrev and the absolute value of the difference image iDiff is taken (step156). By forming the difference image iDiff, background features and noise are removed. The autoSelecthres( )function122 is then called to select a threshold value tValue for the difference image iDiff (step158) based on the histogram of the difference image iDiff. The threshold iThres of the difference image iDiff is then taken (step160) to highlight further the pointer within the current image iCurr. During thresholding a grayscale image is mapped to the binary difference image iDiff. Pixels in the difference image with values equal to or less than the threshold value tValue are made black while all other pixels are made white. The result is a binary image containing the pointer and some noise both designated by white pixels.
• Once the difference image has been thresholded, the[0067]extractPointer function124 is called (step162) to extract the pointer ptr from the difference image iDiff and ignore the noise. The size of the pointer ptr is then examined to determine if it is greater than a threshold value MIN_PTR_SIZE (step164).
• If the size of the pointer is greater than the threshold value MIN_PTR_SIZE, the centerOfMass( )[0068]function126 is called (step166) to determine the center of the pointer. Following this, the processROI( )function128 is called (step168) to select a region of interest ROI within the difference image iDiff and extract the pointer from the region of interest.
• Once the pointer has been extracted from the region of interest ROI, the size of the extracted pointer is determined (step[0069]170). If the pointer size is greater than zero, a medianLine function is called (step172). During execution of the medianLine function, the median line of the pointer (i.e. the pointer location within the region of interest) is calculated using linear least squares. The current image iCurr is then designated as the previous image iPrev to complete thefindPointerMotion( ) function.
• At[0070]step164, if the pointer size is equal to zero, the center of mass parameters Cx and Cz are examined (step174). If both of the center of mass parameters Cx and Cz have values greater zero, the processROI( )function128 is called (step168) to select a region of interest ROI within the difference image iDiff and extract the pointer from the region of interest. Atstep174, if one or both of the center of mass parameters Cx and Cz have values equal to zero or atstep170, if the size of the pointer is less than zero, a no pointer found condition is determined (step176). At this stage, the current image iCurr is designated as a calibration image glRef. ThefindPointerMotion( ) function then proceeds to step152 where the center of mass parameters Cx and Cz are assigned zero values.
• As mentioned above, at[0071]step158, when thefindPointerMotion( ) function120 calls the autoSelectThres( )function122, a threshold value for the difference image iDiff is selected based on the histogram of the difference image so that when the difference image iDiff is thresholded, the pointer is further highlighted from background features and noise. Selection of the threshold value in this manner is more robust than hardcoding the threshold value.
• Turning now to FIG. 9, the steps performed during execution of the autoSelectThres( )[0072]function122 are illustrated. As can be seen, in order to select the threshold level, a histogram of the difference image iDiff is taken and the number of pixels in each bin of the histogram are counted (step200). The number of pixels in the bin having the highest count is used as a peak parameter and the threshold value is initially assigned a value of one (step202). The number of bins having non-zero counts is then examined to determine if more than eight (8) bins have non-zero counts (step204). If less than eight (8) bins have non-zero counts, the threshold value remains at its initially assigned value and the autoSelectThres( ) function is completed.
• At[0073]step204, if more than eight (8) bins have non-zero counts, the number of non-zero bins is checked again to determine if an entire difference image is being processed (i.e. the autoSelectThres( ) function was called by the findPointerMotion( ) function120) or if a region of interest ROI within the difference image is being processed (i.e. the autoSelectThres( ) function was called by the processROI( ) function128) (step206). If the entire difference image iDiff is being processed, a threshold minimum parameter (tMin) is set to a value of twelve (12) and a Peak Div parameter is set to a value of eight (8) (step208). A minimum count parameter minCount is then calculated by dividing the peak parameter determined atstep202 by the Peak_Div parameter (step210). If a region of interest is being processed, the threshold minimum parameter (tMin) is set to a value of forty (40) and the Peak_Div parameter is set to a value of thirty-two (32) (step212) before proceeding to step210.
• Once minCount has been determined, the peak level is checked to determine if it is greater than the threshold minimum tMin (step[0074]214). Peak level is the grayscale level that contains the most pixels. In the case of a tie, the grayscale level with the highest numerical value (i.e. the closest to255) is chosen. If the peak level is greater than the threshold minimum tMin, a startLevel parameter is assigned a value equal to the peak level+1 (step216). Atstep214, if the peak level is less than the threshold minimum tMin, the startLevel parameter is assigned a value equal to the threshold minimum tMin (step218).
• Following[0075]step216 or218, a loop is entered. During the loop, the levCount for each bin having a bin number between the value of the startLevel parameter and two hundred and fifty-five (255) is examined to determine if it is greater than zero and if it is less than the minCount parameter determined at step210 (step220). If the condition is met during the loop, the loop is exited and the threshold value is assigned a value equal to the bin number having the levCount that resulted in the loop being exited+1 (step222). If the condition is not met, the loop is exited after the levCount forbin number255 has been checked.
• Once the loop has been exited, the threshold value is checked to determine if it is less than the minimum threshold value tMin (step[0076]224). If not, a check is again made to determine if an entire difference image is being processed or whether a region of interest ROI is being processed (step226). If the threshold value is less than the minimum threshold value tMin, the threshold value is set equal to the minimum threshold value tMin (step228) before the check is made to determine if an entire difference image is being processed or whether a region of interest is being processed (step226).
• At[0077]step226, if a difference image iDiff is being processed, the autoSelectThres( ) function is completed. However, if a region of interest is being processed, a parameter p is assigned a value corresponding to the first grayscale level at which 90% or more of the pixels will go black (step230). The parameter p is then compared to the threshold level (step232). If the parameter p is less than the threshold level, the autoSelectThres( ) function is completed. If the parameter p is greater than the threshold level, the threshold value is set to the value of parameter p (step234) and the autoSelectThres( ) function is completed.
• As mentioned above, at[0078]step162 thefindPointerMotion( ) function120 calls the extractPointer( )function124 to extract the pointer from the binary image and ignore the noise. This is done by selecting the “white region” in the binary image that is greater than or equal to a certain minimum size and is the highest region (i.e. the largest in the y-axis (20 pixel axis)). Specifically, when the extractPointer( )function124 is called, the extractPointer( ) function calls the getHighestRegion( ) function130 (step250). The getHighestRegion( ) function130 uses the threshold value and tol parameters to select the appropriate white region szRegion in the thresholded difference image. The tol parameter is used to avoid situations where board surface noise is mistaken as a pointer. FIG. 13 shows the steps performed during this function.
• Once the white region szRegion has been selected, the white region szRegion is checked to see if it is greater than zero (step[0079]252). If not, a no pointer condition is determined (step254). If the white region szRegion is greater than zero, morphological operator of erosion and dilation are used on the white region to reduce further noise (steps256 to264) and the extractPointer( ) function is completed.
• As mentioned above, at[0080]step166 thefindPointerMotion( ) function120 calls the centerOfMass( )function126 to determine the center of the pointer. During this function, the black pixels in the binary image are treated as having a mass of zero (0) and the white pixel are treated as having a mass of one (1). The physics formulae for center-of-mass are used. The equation below gives the center of mass in the x-direction:
• C_x=sum(X_i)/M
• where:[0081]
• X[0082]_iare the x-coordinates of the white pixels in the binary image; and
• M is the number of white pixels in the binary image.[0083]
• Initially, once the centerOfMass( ) function is executed, the center of mass parameters massX, massZ and a mass parameter are assigned zero values (see[0084]step300 in FIG. 11). A loop is then entered to calculate the center of mass parameters massX and massZ using the above equation and to calculate the mass parameter (step302).
• Upon exiting the loop, the mass parameter is checked to determine if its value is greater than zero (step[0085]304). If the value of the mass parameter is equal to zero, the center of mass parameters Cx and Cz are assigned values of zero (step306) and the centerOfMass( )function126 is completed. Atstep304, if the value of the mass parameter is greater than zero, the center of mass coordinates Cx and Cz are calculated (step308) using the equations:
• Cx=massX/mass; and
• Cz=massZ/mass.
• Once the center of mass coordinates have been calculated, the centerOfMass( )[0086]function126 is completed.
• As mentioned above, at[0087]step168 thefindPointerMotion( ) function120 calls the processROI( )function128 to process the region-of-interest in a manner similar to the findPointerMotion( ) function120 except, here the image size is 100×2 pixels and a calibration image including only background (i.e. no pointer) is used in place of the previous image. Upon execution of the processROI( )function128, xLeft and xRight parameters are calculated by subtracting and adding fifty (50) to the center of mass parameter Cx (step350). The value of parameter xLeft is then checked to determine if it is less than one (1) (step352). If the parameter xLeft has a value less than one (1), the parameter xRight is recalculated and the parameter xLeft is assigned a value of one (1) (step354) to define boundaries of the region of interest as shown in FIG. 15. A difference image iDiff of the region of interest is then calculated by subtracting the region of interest of the current image from the region of interest of the calibration image glRef determined atstep176 of thefindPointerMotion( )function120 and taking the absolute value of the difference (step356).
• At[0088]step352, if the parameter xLeft has a value greater than one (1), the parameter xRight is checked to determine if it has a value greater than 640 (step358). If the parameter xRight has a value greater than 640, the parameter xLeft is recalculated and the parameter xRight is assigned a value of one (1) (step360) to define boundaries of the region of interest. The processROI( ) function128 then proceeds to step356 to calculate the difference image iDiff of the region of interest. Atstep358, if the parameter xRight has a value less than 640, the processROI( ) function128 proceeds directly to step356 to calculate the difference image iDiff of the region of interest.
• Once the difference image iDiff of the region of interest has been calculated, the autoSelectThres( )[0089]function122 is called to select a threshold value for the difference image iDiff of the region of interest (step362) in the manner described above with reference to FIG. 9. The difference image iDiff of the region of interest is then thresholded (step364). Following this, the extractPointer( )function124 is called to extract the pointer from the difference image iDiff of the region of interest (step366) in the manner described above with reference to FIG. 10.
• Once the acquired image has been processed in the above manner, a PIP for the acquired image is created by the[0090]DSP84. The PIP is a five (5) word packet and has a layout including camera identification, an LRC checksum to ensure data integrity and a valid tag to ensure zero packets are not valid. The valid tag indicates whether the PIP relates to a pointer characteristic packet (10), a diagnostic packet for a specific camera assembly63 (01) or a diagnostic packet for all camera assemblies63 (11). Table 1 below shows the PIP layout.

  TABLE 1
  Word	15	14	13	12	11	10	9	8	7	6	5	4	3	2	1	0

  12/34 Pointer Characteristics packet (generated by Camera)

  0	Valid	Camer	X intercept (at Y0)
  	tag	a #

  1

  Frame rate

  intensity/color

  2	Packet#	pointer area
  3	Unused	X intercept (at Y19)

  4

  Unused

  Zposition

  LRC checksum

  12/34 Diagnostic Packet (generated by Camera or Master)

  0	Valid	Camer
  	tag	a #
  1

  2	Packet #
  3

  4	LRC checksun

• As mentioned above, each[0091]camera assembly63 acquires and processes an image in the manner described above in response to each clock signal generated by itsDSP84. The PIPs created by theDSPs84 are only sent to themaster controller54 when thecamera assemblies63 are polled by the master controller. TheDSPs84 create PIPs faster than themaster controller54 polls thecamera assemblies63. PIPs that are not sent to themaster controller54 are overwritten.
• When the[0092]master controller54 polls thecamera assemblies63, frame sync pulses are sent to thecamera assemblies63 to initiate transmission of the PIPs created by theDSPs84. Upon receipt of a frame sync pulse, eachDSP84 transmits the PIP to themaster controller54 over the data bus. The PIPs transmitted to themaster controller54 are received via theserial port96 and auto-buffered into theDSP90.
• After the[0093]DSP90 has polled thecamera assemblies63 and has received PIPs from each of the camera assemblies, theDSP90 processes the PIPs using triangulation to determine the location of the pointer relative to thetouch surface60 in (x,y) coordinates. Specifically, the PIPs from pairs ofcamera assemblies63 are processed using triangulation.
• FIG. 16 shows that two angles φ[0094]₁and φ₂are needed to triangulate the position (x₀,y₀) of the pointer relative to thetouch screen60. The PIPs generated by eachcamera assembly63 include a number θ∈[0,sensorResolution−1] (see FIG. 17) identifying the median line of the pointer. The sensorResolution, in the case of the Photobit PB300 image sensor, is 640. The equations below relate the angle φ to the position θ taking into account the field-of-view of the image sensor and lens assembly80:$\begin{array}{cc}\phi =\frac{\theta }{\mathrm{sensorResolution}}\times F_{\mathrm{ov}}-\delta & \left(1\right)\\ \phi =\frac{\mathrm{SensorResolution}-\theta }{\mathrm{sensoreResolution}}\times F_{\mathrm{ov}}-\delta & \left(2\right)\end{array}$

  
• The above equations subtract away an angle δ that allows the image sensor and[0095]lens assembly80 to have some overlap with theframe62. The overlap with theframe62 is desired due to mechanical tolerance issues in the frame assemblies64 (i.e. the angle of theplate66 can have an error of 1° to 2°). The angle δ is allowed to be negative, meaning that there is no overlap with theframe62, in fact part of thetouch surface60 along theframe62 is missed.Equation 1 or 2 is used to determine φ, depending on the mounting and/or optical properties of the image sensor andlens assembly80. If the image acquired by thecamera assembly63 is rotated as a result of the mounting and/or optical properties of the image sensor andlens assembly80, thenequation 2 is used.Equation 1 is used otherwise. In the present embodiment,equation 1 is used with thecamera assemblies63 positioned at the top left and bottom right corners of thetouch screen52 andequation 2 is used with thecamera assemblies63 positioned at the bottom left and top right corners of thetouch screen52.
• As discussed above,[0096]equations 1 and 2 allow the pointer median line data included in the PIPs to be converted by theDSP90 into an angle φ with respect to the x-axis. When two such angles are available, the intersection of the median lines extending at these angles from theirrespective camera assemblies63 yields the location of the pointer relative to thetouch surface60.
• In this embodiment, since the[0097]touch screen52 includes fourcamera assemblies63, six pairs of camera assemblies can be used for triangulation. The following discussion describes how a pointer position is determined by triangulation for each pair of thecamera assemblies63.
• In order to determine a pointer position using the PIPs received from the[0098]camera assemblies63 along the left side of thetouch screen52, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₀and φ₁for the upper and lower camera assemblies:$\begin{array}{cc}{x}_0=\frac{h}{w}\times \frac{1}{\tan \left({\phi }_0\right)+\tan \left({\phi }_1\right)}& \left(3\right)\\ {\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=\frac{\tan \left({\phi }_0\right)}{\tan \left({\phi }_0\right)+\tan \left({\phi }_1\right)}& \left(4\right)\end{array}$

  
• where:[0099]
• h is the height of the[0100]touch screen52 i.e. the vertical distance from camera assembly focal point-to-focal point;
• w is the width of the[0101]touch screen52 i.e. the horizontal distance from camera assembly focal point-to-focal point; and
• φ[0102]_iis the angle with respect to the horizontal, measured using camera assembly i andequation 1 or 2.
• For the[0103]camera assemblies63 along on the right side of thetouch screen52, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₂and φ₃for the upper and lower camera assemblies:$\begin{array}{cc}{x}_0=1-\frac{h}{w}\times \frac{1}{\tan \left({\phi }_2\right)+\tan \left({\phi }_3\right)}& \left(5\right)\\ {\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=1-\frac{\tan \left({\phi }_2\right)}{\tan \left({\phi }_2\right)+\tan \left({\phi }_3\right)}& \left(6\right)\end{array}$

  
• The similarity between[0104]equations 3 and 5, i.e. equation 5=1−equation 3 once φ₂and φ₃have been substituted intoequation 3 for φ₁and φ₂respectively should be apparent. Equations 4 and 6 are related in a similar manner.
• In order to determine a pointer position using the[0105]camera assemblies63 along the bottom of thetouch screen52, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₀and φ₃for bottom left and bottom right camera assemblies:$\begin{array}{cc}{x}_0=\frac{\tan \left({\phi }_3\right)}{\tan \left({\phi }_0\right)+\tan \left({\phi }_3\right)}& \left(7\right)\\ \begin{array}{c}{\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=\frac{w}{h}\times \frac{\tan \left({\phi }_3\right)}{\tan \left({\phi }_0\right)+\tan \left({\phi }_3\right)}\times \tan \left({\phi }_0\right)\\ =\frac{w}{h}\times x_0\times \tan \left({\phi }_0\right)\end{array}& \left(8\right)\end{array}$

  
• In order to determine a pointer position using the[0106]camera assemblies63 along the top of thetouch screen52, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₁and φ₂for the top left and top right camera assemblies:$\begin{array}{cc}{x}_0=\frac{\tan \left({\phi }_2\right)}{\tan \left({\phi }_1\right)+\tan \left({\phi }_2\right)}& \left(9\right)\\ \begin{array}{c}{\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=1-\frac{w}{h}\times \frac{\tan \left({\phi }_2\right)}{\tan \left({\phi }_1\right)+\tan \left({\phi }_2\right)}\times \tan \left({\phi }_1\right)\\ =1-\frac{w}{h}\times x_0\times \tan \left({\phi }_1\right)\end{array}& \left(10\right)\end{array}$

  
• The similarity between equations 7 and 9, i.e. equation 9=equation 7 once φ[0107]₁and φ₂have been substituted into equation 7 for φ₀and φ₃should be apparent.Equations 8 and 10 have the following relationship:equation 10=1−equation 8 once φ₁and φ₂have been substituted intoequation 8 for φ₀and φ₃respectively.
• In order to determine a pointer position using the[0108]camera assemblies63 across the bottom left to top right corner diagonal, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₀and φ₂for bottom left and top right camera assemblies:$\begin{array}{cc}{x}_0=\frac{\frac{h}{w}-\tan \left({\phi }_2\right)}{\tan \left({\phi }_0\right)-\tan \left({\phi }_2\right)}& \left(11\right)\\ {\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=\frac{1-\frac{w}{h}-\tan \left({\phi }_2\right)}{\tan \left({\phi }_0\right)-\tan \left({\phi }_2\right)}\times \tan \left({\phi }_0\right)& \left(12\right)\end{array}$

  
• In order to determine a pointer position using the[0109]camera assemblies63 across the bottom right to top left diagonal, the following equations are used to determine the (x₀,y₀) coordinates of the pointer position given the angles φ₁and φ₃for the bottom right and top left camera assemblies:$\begin{array}{cc}{x}_0=\frac{\frac{h}{w}-\tan \left({\phi }_3\right)}{\tan \left({\phi }_1\right)-\tan \left({\phi }_3\right)}& \left(13\right)\\ {\mathrm{<figure-callout\; label="y"\; filenames="US20040012573A1-20040122-D00005.png,US20040012573A1-20040122-D00006.png"\; state="\{\{state\}\}">y</figure-callout>}}_0=1-\frac{1-\frac{w}{h}-\tan \left({\phi }_3\right)}{\tan \left({\phi }_1\right)-\tan \left({\phi }_3\right)}\times \tan \left({\phi }_1\right)& \left(14\right)\end{array}$

  
• The similarity between equations 11 and 13, i.e. equation 13=equation 11 once °1 and °3 have been substituted into equation 11 for φ[0110]₀and φ₂should be apparent.Equations 12 and 14 have the following relationship: equation 14=1−equation 12 once φ₁and φ₃have been substituted intoequation 12 for φ₀and φ₂respectively.
• As will be appreciated, the above equations generate the coordinates x[0111]₀and y₀on a scale of [0, 1]. Therefore, any appropriate coordinate scale can be reported by multiplying x₀and y₀by the maximum X and maximum Y values respectively.
• In the present embodiment, the[0112]DSP90 calculates the pointer position using triangulation for each camera pair excluding the diagonal pairs. The resulting pointer positions are then averaged and the resulting pointer position coordinates are queued for transmission to thepersonal computer56 via theserial port98 and theserial line driver94. Since the rows of pixels of the image sensor andlens assemblies80 that correspond to actual contacts with thetouch surface60 are known, any Z-position in a PIP that does not correspond with one of these rows is by definition a pointer hover event.
• If desired, pointer velocity and angle can be calculated by the[0113]DSP90 as shown in FIG. 18. The velocity of the pointer is calculated by examining the changes in the Z-position (or X-intercept) of the pointer in successive PIPs and knowing the camera frame rate. For example, if the camera frame rate is 200 frames per second and the Z-position changes by 1 pixel per frame, the pointer velocity is 200 pixels per second.
• The angle of the pointer can be determined due to the fact that the PIP includes the X-intercept at[0114]pixel rows0 and19 of the median line. Since the X distance (the difference between X-intercepts) and the Y distance (the number of pixel rows) are known, all of the information necessary to calculate the pointer angle is available.
• The present invention provides advantages in that the[0115]passive touch system50 does not suffer parallax and/or image distortion problems due to the fact that a glass or other transparent overlay over a computer or video display is not required. In addition, the presentpassive touch system50 allows both pointer contact and pointer hover over thetouch surface60 to be detected by using two-dimensional image sensor andlens assemblies80 in the plane of thetouch surface60. Pointer contact with thetouch surface60 is defined only when the pointer is in very close proximity of the touch surface. The present invention also provides advantages in that the pointer position with respect to the touch surface is not restricted since the image sensor andlens assemblies80 look along the plane of thetouch surface60.
• With respect to resolution, the resolution of the passive touch system is a function of the distances of the pointer with respect to the image sensor and[0116]lens assemblies80, the number of pixel elements in the image sensor and lens assemblies and the fields of view of the image sensor and lens assemblies. Since image sensor and lens assemblies are available with pixel elements that range in number from tens of thousand to many millions and since the number of pixel elements in image sensors and lens assemblies of this nature is only expected to increase, the resolution of the presentpassive touch system50 is high.
• The[0117]passive touch system50 also provides advantages in that alignment is automatically corrected since only pixel subsets of images that include the touch surface and the pointer are processed. In addition, the present passive touch system allows for very fast acquisition of image data since the image sensor and lens assemblies can be triggered to capture images at rates exceeding two hundred frames per second.
• The present[0118]passive touch system50 is scaleable and can include atouch surface60 of arbitrary size. When used in conjunction with a projected computer image, the number of pixels of the image sensor and lens assemblies should be proportional to the number of pixels being displayed on thetouch surface60. For example, if a projected computer image is 1024×768 pixels, the size of the projected image is not be of concern provided the image sensor andlens assemblies80 are able to resolve the (x,y) coordinates with sufficient accuracy with respect to the displayed pixels.
• Although the[0119]passive touch system50 is shown includingcamera assemblies63 associated with each corner of thetouch screen52, those of skill in the art will appreciate that only two camera assemblies are required. In this case, the fields of view of the image sensor and lens assemblies are preferably selected so that theentire touch surface60 is encompassed since the locations of pointer contacts are determined only when they occur within the overlapping fields of view of thecamera assemblies63.
• Also, although the[0120]passive touch system50 is described as including a projector to display the computer display output onto thetouch surface60, this is not required. No information need be displayed on the touch surface.
• Although a preferred embodiment of the present invention has been described, those of skill in the art will appreciate that variations and modifications may be made without departing from the spirit and scope thereof as defined by the appended claims.[0121]

Claims (18)

What is claimed is:

1. A passive touch system comprising:

a passive touch surface;

at least two cameras associated with said touch surface, said at least two cameras acquiring images of said touch surface from different locations and having overlapping fields of view; and

a processor receiving and processing images acquired by said at least two cameras to detect the existence of a pointer therein and to determine the location of said pointer relative to said touch surface.

2. A passive touch system as defined inclaim 1 wherein said at least two cameras are two-dimensional image sensor and lens assemblies having fields of view looking along the plane of said touch surface.

3. A passive touch system as defined inclaim 2 wherein said processor determines the location of said pointer relative to said touch screen using triangulation.

4. A passive touch system as defined inclaim 3 wherein said processor determines when the pointer is in contact with said touch surface and when said pointer is hovering above said touch surface.

5. A passive touch system as defined inclaim 2 wherein said processor selects pixel subsets of images acquired by said image sensor and lens assemblies and processes said pixel subsets to determine the existence of said pointer.

6. A passive touch system as defined inclaim 5 wherein said pixel subsets are determined during an alignment routine.

7. A passive touch system as defined inclaim 2 wherein said processor includes a digital signal processor associated with each image sensor and lens assembly and a master digital signal processor in communication with the digital signal processors associated with each image sensor and lens assembly, the digital signal processors associated with each image sensor and lens assembly selecting the pixel subsets and processing the pixel subsets to determine the existence of said pointer, the master digital signal processor receiving pixel characteristic data from the digital signal processors associated with each image sensor and lens assembly and triangulating the pixel characteristic data to determine the location of said pointer relative to said touch surface.

8. A passive touch system as defined inclaim 7 wherein said pixel characteristic data includes a median line of the pointer.

9. A passive touch system as defined inclaim 8 wherein said pixel characteristic data is packaged by the digital signal processors associated with each image sensor and lens assembly into pointer information packets (PIPs).

10. A passive touch system as defined inclaim 9 wherein said master digital signal processor polls the digital signal processors associated with each image sensor and lens assembly for the PIPs at a rate less than the rate at which said image sensor and lens assemblies acquire images.

11. A passive touch system as defined inclaim 7 including an image sensor and lens assembly adjacent each corner of said touch screen.

12. A passive touch system as defined inclaim 11 further including a computer coupled to said master digital signal processor, said computer receiving pointer location data from said master digital signal processor and processing the same.

13. A method of detecting the position of a pointer relative to a passive touch surface comprising the steps of:

acquiring images of said touch surface from different locations using cameras having overlapping fields of view; and

processing said images to detect the existence of a pointer therein and to determine the location of said pointer relative to said touch surface.

14. The method ofclaim 13 wherein during said acquiring step, said images are acquired using two-dimension image sensor and lens assemblies having fields of view looking along the plane of said touch surface.

15. The method ofclaim 14 wherein during the processing step, the location of said pointer relative to said touch screen is determined using triangulation.

16. The method ofclaim 15 wherein during said processing step, the images are processed to determine when said pointer is in contact with said touch surface and when said pointer is hovering above said touch surface.

17. The method ofclaim 16 further comprising the step of selecting pixel subsets of images acquired by said image sensor and lens assemblies prior to processing said images.

18. The method ofclaim 17 wherein during said processing step the existence of said pointer is determined by calculating median lines of the pointer and wherein the location of said pointer is determined by calculating the intersection point of median lines and using triangulation to determine the coordinates of said intersection point.

Images