in which DCM_cam2bodyFor a conversion matrix of pixel orientation into a star system, K, taking into account the internal and external orientation elements of the camera_{pixel_i}Coordinates of pixel points corresponding to the ith image control point, L_pixelIs the pixel size and f is the focal length;

step five: two nonparallel target vectors obtained in the third step are selected at will and are respectively R_{sat2tar_sen_i}And R_{sat2tar_sen_j}Establishing the following three orthogonal unit vectors of the star sensor measurement system:

two unparallel target vectors obtained in the fourth step are randomly selected and are respectively R_{sat2tar_body_i}And R_{sat2tar_body_j}Establishing three orthogonal unit vectors under a camera body coordinate system:

step six: calculating to obtain a direction cosine array from the star sensitive measurement system to the camera body coordinate system according to three orthogonal unit vectors under the star sensitive measurement system and the camera body coordinate system in the fifth step, namely a star sensor installation matrix:

DCM_sen2body＝Mat_{body_ij}^T*Mat_{sen_ij} (6)

step seven: when N is 2, directly solving the star sensor installation matrix to obtain an on-orbit calibration result of the star sensor installation matrix;

and when N is larger than 2, optimally solving the star sensor mounting matrix by using a least square method to obtain the optimal estimation of the star sensor mounting matrix.

Compared with the prior art, the invention has the following beneficial effects:

the method for installing and calibrating the remote sensing satellite star sensor based on the image control points can realize the on-orbit calibration of the satellite attitude determination system, continuously monitor the on-orbit change of the installation condition of the satellite attitude determination system and effectively improve the accuracy and the reliability of the attitude determination system.

Drawings

FIG. 1 is a flow chart of a remote sensing satellite star sensor installation calibration method based on image control points.

Detailed Description

The technical solution of the present invention will be described in detail with reference to the accompanying drawings and preferred embodiments.

The invention aims to provide an on-orbit calibration method of an attitude determination sensor installation matrix based on optical load imaging, which is suitable for a planar/linear array push-broom remote sensing satellite, wherein the method comprises the following basic principles of identifying and extracting the geographic positions of some control points in an original image, and calculating the optimal estimation of a star sensor installation matrix by combining the time, the navigation and the output value of the star sensor during the satellite imaging:

according to the double-vector attitude determination principle, the relative relationship between the two coordinate systems can be calculated by only knowing the representation of the two non-collinear vectors under the two coordinate systems. For an area array imaging satellite, only two or more control points are needed in one frame of image to determine the instantaneous satellite attitude and calculate the installation matrix by combining the output of the time sensor. However, the linear array imaging satellite only acquires one row of images at one imaging moment, and because the control points are sparse, a plurality of control points are difficult to be arranged in one row, the difficulty in determining the instantaneous attitude is high. Because the sensor installation matrix is essentially the relative relation from the measurement system to the satellite system, the relative relation is determined by the satellite structure, the single rigid body satellite imaged by the whole maneuvering can be regarded as constant, and is irrelevant to the time and the whole satellite attitude, so the vector can be considered under the measurement system and the satellite system, the vector under the inertial system from the current satellite to the control point can be obtained by identifying the latitude and longitude height of the control point and the navigation data of the current satellite, and the current satellite sensitive output is essentially the rotation matrix from the inertial system to the measurement system, so the target direction vector under the satellite sensitive measurement system can be obtained; in addition, the vector of the pixel visual axis corresponding to the control point under the star system can be obtained according to the calibration results of the inner part and the outer part of the camera. Two vectors (essentially, the same vector is represented under two coordinate systems) can be obtained every time one control point is obtained, so that the relative relation between the system and the measurement system can be calculated only by obtaining two control points. If the number of points is more than two, the optimal estimation of the installation matrix can be obtained according to a least square method.

The invention realizes the method for calibrating the star sensor for the remote sensing satellite installed on the orbit by the following technical scheme, and the method is suitable for the star sensor of the area array push-broom remote sensing satellite or the linear array push-broom remote sensing satellite of the whole star motor-driven imaging, and specifically comprises the following steps:

the method comprises the following steps: selecting a row with image control points with higher precision from an original image which is dense, cloud-free, good in imaging quality and not subjected to splicing and embedding in a remote sensing satellite ground imaging task, preferably, the image control points are preferably positioned at two sides of the image to improve an observation base line, then extracting the positions and row numbers of pixel points corresponding to the image control points, and identifying and obtaining the latitude and longitude height [ Lat ] corresponding to the ground position according to the image control points_i Lon_i H_i]Where i ═ 1,2, …, N]And N is more than or equal to 2.

Extracting image line number corresponding time T and satellite current position coordinate R from auxiliary data_sat84Star sensitive output Q_sen(if the imaging line transfer center time is not consistent with the navigation time and the star sensor sampling time, the position/star sensor output can be interpolated, and all data are aligned to the imaging time).

Step two: the height of the latitude and longitude [ Lat ] of the corresponding ground position of each pixel point_i Lon_i H_i]Transformed into WGS84 system coordinate R_{tar84_i}And coordinate R is determined_{tar84_i}And the coordinate R of the current position of the remote sensing satellite_{sat_84}Subtracting to obtain the target vector under WGS 84:

R_{sat2tar84_i}＝R_{tar84_i}-R_{sat_84} (1)

step three: combining the time T corresponding to the line number to make WGS84 the target vector R_{sat2tar84_i}Transferring to J2000 series to obtain target vector R under J2000 series_{sat2tar2000_i}At this time, the star sensor output is represented by a direction cosine array as DCM_senThen, the target vector under the star sensitivity measurement system is:

R_{sat2tar_sen_i}＝DCM_sen*R_{sat2tar2000_i}, (2)

step four: calculating vector R of pixel point under camera body coordinate system by inner and outer orientation elements of camera_{sat2tar_body_i}The calculation formula is as follows:

in which DCM_cam2bodyConversion matrix of pixel orientation to star system, K, for taking into account internal and external orientation elements such as camera mounting and focal plane shift axes_{pixel_i}The coordinates of the pixel point corresponding to the ith control point (i.e. the position number of the pixel point counted from the main point), L_pixelIs the pixel size and f is the focal length.

Then, according to the third step and the fourth step, two unparallel target vectors (namely two non-coincident control points) are arbitrarily taken and respectively expressed in a camera body coordinate system and a satellite sensitive measurement system (the larger the included angle between the vectors is, the better), and the current attitude of the satellite can be determined by double-vector attitude determination after unitization.

in this step, the two target vectors in the arbitrarily selected step three or step four may be vectors at the same time, or two vectors at different times, that is, the original image of the remote sensing satellite in the present invention does not require all pixel points to be shot at the same time, and the vectors do not require time alignment for use.

Step six: calculating to obtain a direction cosine array from the star sensor measurement system to the camera body coordinate system according to the three orthogonal unit vectors under the star sensor measurement system and the three orthogonal unit vectors under the camera body coordinate system in the fifth step:

DCM_sen2body＝Mat_{body_ij}^T*Mat_{sen_ij} (6)

the direction cosine matrix is the star sensor installation matrix.

and when N is larger than 2, the number of the target vectors under the star sensor measurement system and the number of the target vectors under the camera body coordinate system respectively exceed two, and at the moment, the least square method is utilized to optimally solve the star sensor installation matrix to obtain the optimal estimation of the star sensor installation matrix so as to reduce the random calibration error.

The star sensor mounting matrix can convert the vectors of the batch star sensitive measurement system into the camera body coordinate system:

obviously, only two non-parallel vectors are required, the matrix Mat_{sen_ij}Full rank, the equation above can solve for DCM_sen2body。

If the number of the vectors is more than two (namely N is more than 2), the vectors are an over-constraint equation system, and the known vectors of the installation matrix to be solved in the formula (7) are expressed as DCM_sen2bodyX, the vector under the star-sensitive measurement system is a coefficient matrix [ Mat_{sen_ij} Mat_{sen_kw} ...]^TThe vector in the camera body coordinate system is expressed as [ Mat [ ] A_{body_ij} Mat_{body_kw} ...]^TEquation (7) can be expressed as a simple system of linear equations:

A*x＝b (8)

solving the optimal solution of the system of linear equations, i.e. the order

The minimum, namely:

namely:

x＝(A^TA)^-1A^Tb (10)

therefore, when obtaining a plurality of groups of control point data, the optimal estimation of the star sensor installation matrix is as follows:

because the same attitude-fixing quaternion is used for calibration, the relative installation between the two star sensors can be kept consistent, the situation that the attitude-fixing quaternion is changed suddenly after the star sensors are switched does not exist, and the relative installation calibration is not needed.

In order to further verify the effectiveness of the method provided by the invention, the technical effect of the method is further explained by combining an actual star sensor on-orbit installation calibration experiment. In the experiment, a Jilin I constellation high score 03D series satellite is used as a remote sensing satellite.

A Jilin I star seat height division 03D series satellite is successfully launched into orbit in a Taiyuan launching center in 2021, 7 months and 3 days, is a sub-meter resolution optical remote sensing satellite, adopts a whole-star maneuvering linear array push-broom imaging, is compact in structure, good in rigidity, visible as a single rigid body, is provided with two star sensors, and is suitable for in-orbit installation calibration experiments.

Selecting 'Jilin I' high-score 03D01 satellite 7 to form image data, identifying 10 control points in each group, and respectively calculating the optimal estimation of the installation matrix of the two star sensors in the above way, wherein the result is as follows:

TABLE 1 Star sensor 1 installation calibration results

TABLE 2 Star sensor 2 installation calibration results

As can be seen from tables 1 and 2, the satellite sensitive actual mounting matrix can be effectively obtained by adopting the linear array push-broom image control point calibration strategy, the calibration result is stable, the calibration error 3 sigma value is less than 0.2 degrees, the error in the actual calibration process is mainly concentrated in the rotation direction around the visual axis of the camera, the influence on the image positioning precision is small, the mounting calibration precision in the direction of the load visual axis is less than 0.05 degrees, and the method is obviously improved compared with the method without calibration.

The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present invention, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

Translated fromChinese

1.一种基于图像控制点的遥感卫星星敏感器安装标定方法，其特征在于，包括以下步骤：1. a remote sensing satellite star sensor installation calibration method based on image control point, is characterized in that, comprises the following steps:

步骤一：在遥感卫星原始图像中选取存在精度较高图像控制点的行，提取各个图像控制点对应的像素点的位置、行号，并根据各个图像控制点识别获得对应地面位置纬经高[Lat_i Lon_i H_i]，其中i＝[1,2,…,N]且N≥2；Step 1: In the original image of the remote sensing satellite, select the row with high-precision image control points, extract the position and row number of the pixel point corresponding to each image control point, and obtain the corresponding ground position latitude, longitude and height according to the identification of each image control point [ Lat_i Lon_i H_i ], where i=[1,2,...,N] and N≥2;

步骤二：将各个所述像素点的对应地面位置纬经高[Lat_i Lon_i H_i]转为WGS84系下坐标R_{tar84_i}，并将坐标R_{tar84_i}与遥感卫星当时位置坐标R_{sat_84}相减，得到WGS84系下目标矢量：Step 2: Convert the corresponding ground position latitude, longitude and height [Lat_i Lon_i H_i ] of each pixel point to the coordinate R_{tar84_i} under the WGS84 system, and subtract the coordinate R_{tar84_i} from the current position coordinate R_{sat_84} of the remote sensing satellite to obtain Target vector under WGS84 system:

R_{sat2tar84_i}＝R_{tar84_i}-R_{sat_84} (1)R_{sat2tar84_i} =R_{tar84_i} -R_{sat_84} (1)

步骤三：结合所述行号对应的时刻将目标矢量R_{sat2tar84_i}转到J2000系，得到J2000系下目标矢量R_{sat2tar2000_i},，此时星敏输出用方向余弦阵表示为DCM_sen，则在星敏测量系下目标矢量为：Step 3: Transfer the target vector R_{sat2tar84_i} to the J2000 system according to the time corresponding to the row number, and obtain the target vector R_{sat2tar2000_i} under the J2000 system. At this time, the star-sensing output is expressed as DCM_sen by the direction cosine matrix. The lower target vector is:

R_{sat2tar_sen_i}＝DCM_sen*R_{sat2tar2000_i}, (2)R_{sat2tar_sen_i} =DCM_sen *R_{sat2tar2000_i} , (2)

步骤四：由相机内外方位元素计算所述像素点在相机本体坐标系下目标矢量R_{sat2tar_body_i}，计算公式如下：Step 4: Calculate the target vector R_{sat2tar_body_i} of the pixel point in the camera body coordinate system from the inside and outside orientation elements of the camera, and the calculation formula is as follows:

其中DCM_cam2body为考虑相机内外方位元素在内的像素指向到星体系转换矩阵，K_{pixel_i}为第i个图像控制点对应的像素点坐标，L_pixel为像元尺寸，f为焦距；Among them, DCM_cam2body is the pixel-to-star system transformation matrix considering the internal and external orientation elements of the camera, K_{pixel_i} is the pixel coordinate corresponding to the i-th image control point, L_pixel is the pixel size, and f is the focal length;

步骤五：任意选取步骤三中得到的两个不平行的目标矢量，分别为R_{sat2tar_sen_i}和R_{sat2tar_sen_j}，建立星敏测量系下三个正交单位向量：Step 5: Arbitrarily select the two non-parallel target vectors obtained in Step 3, which are R_{sat2tar_sen_i} and R_{sat2tar_sen_j respectively} , and establish three orthogonal unit vectors under the star-sensing measurement system:

任意选取步骤四中得到的两个不平行的目标矢量，分别为R_{sat2tar_body_i}和R_{sat2tar_body_j}，建立相机本体坐标系下三个正交单位向量：Arbitrarily select two non-parallel target vectors obtained in step 4, R_{sat2tar_body_i} and R_{sat2tar_body_j respectively} , and establish three orthogonal unit vectors in the camera body coordinate system:

步骤六：根据步骤五中星敏测量系下和相机本体坐标系下三个正交单位向量，计算得到星敏测量系到相机本体坐标系方向余弦阵，即为星敏感器安装矩阵：Step 6: According to the three orthogonal unit vectors under the star-sensing measurement system and under the camera body coordinate system in step 5, the cosine array in the direction from the star-sensing measurement system to the camera body coordinate system is calculated, which is the star sensor installation matrix:

DCM_sen2body＝Mat_{body_ij}^T*Mat_{sen_ij} (6)DCM_sen2body = Mat_{body_ij}^T *Mat_{sen_ij} (6)

步骤七：当N＝2时，对所述星敏感器安装矩阵直接求解，得到星敏感器安装矩阵的在轨标定结果；Step 7: when N=2, directly solve the star sensor installation matrix to obtain the on-orbit calibration result of the star sensor installation matrix;

当N＞2时，利用最小二乘法对所述星敏感器安装矩阵进行最优求解，得到星敏感器安装矩阵的最优估计。When N>2, use the least squares method to optimally solve the installation matrix of the star sensor, and obtain the optimal estimate of the installation matrix of the star sensor.

2.根据权利要求1所述的一种基于图像控制点的遥感卫星星敏感器安装标定方法，其特征在于，所述图像控制点位于所述遥感卫星原始图像的两侧。2 . The method for installing and calibrating a remote sensing satellite star sensor based on an image control point according to claim 1 , wherein the image control point is located on both sides of the original image of the remote sensing satellite. 3 .

3.根据权利要求1所述的一种基于图像控制点的遥感卫星星敏感器安装标定方法，其特征在于，所述遥感卫星原始图像采用遥感卫星在对地成像任务中图像控制点密集、无云、成像质量好且未经过拼接镶嵌的图像。3. a kind of remote sensing satellite star sensor installation calibration method based on image control point according to claim 1, is characterized in that, described remote sensing satellite original image adopts remote sensing satellite in the ground imaging task, image control point is dense, has no Clouds, images with good image quality and without mosaicking.

4.根据权利要求1所述的一种基于图像控制点的遥感卫星星敏感器安装标定方法，其特征在于，所述遥感卫星为整星机动成像的面阵推扫遥感卫星或者线阵推扫遥感卫星。4. a kind of remote sensing satellite star sensor installation calibration method based on image control point according to claim 1, is characterized in that, described remote sensing satellite is the area array push-broom remote-sensing satellite or linear array push-broom of whole star motorized imaging Remote Sensing Satellites.

5.根据权利要求1所述的一种基于图像控制点的遥感卫星星敏感器安装标定方法，其特征在于，所述遥感卫星原始图像不需要所有像素点拍摄于同一时间，矢量对使用无需时间对齐。5. a kind of remote sensing satellite star sensor installation and calibration method based on image control point according to claim 1, is characterized in that, described remote sensing satellite original image does not need all pixel points to be photographed at the same time, and vector pair use does not need time Align.