Camera and radar photoelectric axis on-orbit calibration method based on weight model

Abstract

A camera and radar photoelectric axis on-orbit calibration method based on a weight model comprises the following steps: controlling the attitude of a whole satellite to enable the optical axis of a camera to point to a target, and recording the position of the target under a camera system; determining a radar scanning range and a scanning step length, establishing a radar scanning coordinate system, and determining radar scanning points; traversal scanning is carried out in a determined scanning range according to a predetermined scanning step length in sequence, and a radar sends a radar beam to each scanning point; virtual scatter points are divided in a radar scanning area, the weight of each virtual scatter point in the scanning range is determined according to the echo intensity and the positions of the virtual scatter points, and the weight of each virtual scatter point in the radar scanning range is continuously updated according to the radar echo condition of each time of scanning; calculating coordinates of the target in the radar scanning coordinate system according to each virtual scatter weight; and after scanning is completed, calculating the position of the target under the radar scanning system, determining an included angle between a camera optical axis and a radar electric axis by combining the position of the target under the camera system, and completing calibration of the camera and the radar.

Description

Technical Field

[0001] This invention relates to an on-orbit calibration method for the photoelectric axes of cameras and radars based on a weighted model, and more particularly to an on-orbit calibration method for spaceborne radars and optical cameras without angle measurement functions based on a weighted model, belonging to the field of optical imaging device calibration. Background Technology

[0002] Because radar and optical cameras possess high ranging and angle measurement accuracy respectively, they are often used together in space-based detection to achieve target tracking, detection, and relative navigation. The prerequisite for accurately describing target motion using radar and camera measurement data is a unified measurement benchmark, i.e., calibrating the relative relationship between the radar's electrical axis and the optical camera's optical axis.

[0003] In-orbit thermal conditions can cause a certain deviation in the photoelectric axis of the camera / radar relative to ground calibration. If the radar has angle measurement capability, in the optical camera self-tracking state (the satellite completes the tracking and pointing of the target based on the camera's angle measurement results, and the azimuth and elevation error voltages are all near zero), the target angle measurement results output by the radar can be directly read, and the photoelectric deviation can be calculated by statistical averaging. This yields the target's angle measurement values ​​in the camera coordinate system and the radar coordinate system, thereby achieving in-orbit calibration of the photoelectric axis.

[0004] However, when the radar lacks angle measurement capabilities, it's impossible to perform on-orbit photoelectric axis calibration of the camera and radar using the conventional methods described above. Without on-orbit calibration of the radar's electrical axis and the camera's optical axis, not only will the accuracy of composite detection be reduced, but when the radar beam angle is narrow, the misalignment between the radar's electrical axis and the camera's optical axis may prevent the radar beam from accurately targeting the target satellite when the camera's optical axis is pointing towards it, or even prevent the reception of the target satellite's echo. The on-orbit calibration technology for the photoelectric bias of spaceborne radar and cameras based on a weighted model solves the problem of on-orbit calibration of radars and cameras without angle measurement capabilities.

[0005] Current on-orbit calibration of radar and cameras relies on their angle measurement capabilities, and there is still a gap in high-precision calibration methods for radar and cameras without angle measurement capabilities. Summary of the Invention

[0006] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide an on-orbit calibration method for camera and radar photoelectric axes based on a weighted model, which solves the problem of on-orbit calibration of radar and camera without angle measurement function.

[0007] The technical solution of this invention is: an on-orbit calibration method for camera and radar photoelectric axes based on a weighted model, comprising: Select the target for calibration, control the overall satellite attitude to make the camera optical axis point to the target, and record the target's position under the camera system; Determine the radar scanning range and scanning step size, establish the radar scanning coordinate system, and determine the radar scanning points; The radar sequentially scans the defined scanning range with a predetermined scanning step size. At each scanning point, the radar sends a radar beam and divides the radar scanning area into several virtual scattered points. The weight of each virtual scattered point in the current scanning range is determined based on the echo intensity and the position of the virtual scattered points. The weight of each virtual scattered point in the radar scanning range is continuously updated based on the radar echo situation of each scan. Calculate the target's coordinates in the radar scanning coordinate system based on the weights of each virtual scatter point; After scanning is completed, the target's position in the radar scanning system is calculated, and the angle between the camera's optical axis and the radar's electrical axis is determined by combining the target's position in the camera system, thus completing the calibration of the camera and radar.

[0008] Preferably, the selection of the target needs to simultaneously satisfy: The target is within the radar's operating range; The target's magnitude is superior to the detection sensitivity of an optical camera.

[0009] Preferably, when determining the radar scanning range: Maximum photoelectric axis deviation caused by on-orbit thermal and mechanical conditions θ max It is estimated that the radar scanning range is no less than 2. θ max ; The scanning step size is less than the radar half-beamwidth U.

[0010] Preferably, the camera and radar are linked, and when establishing the radar scanning coordinate system: Point the camera's optical axis toward the target, establish a radar scanning coordinate system with the radar's current position, and take the direction of the radar's current electrical axis as the starting point for scanning.

[0011] Preferably, virtual scattered points are divided within the radar scanning area, specifically as follows: Let the coordinates be in the radar scanning coordinate system, at the... i During the next scan, the theoretical scan point of the radar is located at... S i , Indicates the scan count index value. And it is a positive integer. Indicates the total number of scans; Virtual scatter points are obtained by dividing the radar scanning area into steps with a step size de. P j , express S i The index values ​​of virtual scatter points within a certain surrounding area. And it is a positive integer. This indicates the total number of virtual scatter points within the radar scanning area.

[0012] Preferably, when determining the weights of each virtual scatter point at the current scanning position based on the echo intensity: If the radar echo signal-to-noise ratio is greater than the target detection signal-to-noise ratio, then the... i The weights of each virtual scatter point during the second scan W i ( j )for: W i ( j )= w f j +(1- w ) g j If no radar echo is received or the echo signal-to-noise ratio is lower than the target detection signal-to-noise ratio, then the... i The weights of each virtual scatter point during the next scan are: W i ( j )= u f j +(1- u ) g j w Indicates the probability of radar false alarms. u Indicates the probability of target detection; f j 、g j These represent the first and second basic weights of each virtual scatter point, determined based on the echo intensity.

[0013] Preferably, the first basic weight is determined. f j At that time, by scanning point S i Set 2n+1 regions around the center, where n = floor( e max / de floor () () indicates rounding down; When virtual scatter points P j The position satisfies:

[0014] Then the first basic weight f j for:

[0015] When virtual scatter points P j The position satisfies:

[0016] in, k =2~2n and are integers, then the first basic weight f j for:

[0017] When scatter P j Located outside the scanning range region 1 to 2n, the first basic weight f j for: f j = ε Where ε represents less than 10 -4 The minimum value; F ( () indicates satellite pointing error e The cumulative distribution function; e max Indicates satellite pointing error e 3σ or maximum value; Dis ( S i , P j )represent S i Connection with satellite, P j The angle between the line connecting it to the satellite.

[0018] Preferably, the second basic weight is determined. g j At that time, by scanning point S i Set 2n+1 regions around the center, where n = floor( e max / de floor () () indicates rounding down; When virtual scatter points P j The position satisfies:

[0019] Then the second basic weight g j for:

[0020] When virtual scatter points P j The position satisfies:

[0021] in, k =2~2n and are integers, then the second basic weight g j for:

[0022] When scatter P j Located outside the scanning range region 1 to 2n, the second basic weight g j for: g j = ε Where ε represents less than 10 -4 The minimum value; F ( () indicates satellite pointing error e The cumulative distribution function; e max This indicates a satellite pointing error of 3σ or its maximum value; Dis ( S i , P j ) represents the scan point S i Connections to satellites, virtual scatter plots P j The angle between the line connecting it to the satellite.

[0023] Preferably, the weights of each scatter point within the radar scanning range are updated by combining the results of multiple scans, specifically as follows:

[0024] W i ( j ) indicates through the first i The first scan obtained the j The weights of the virtual scatter points W I ( j ) indicates a synthesis of the 1st, 2nd, ... I The first scan result after the second scan j The weights of each virtual scatter point.

[0025] Preferably, the positions and weights of each virtual scatter point are weighted to estimate the target's position in the radar scanning system. T l :

[0026] W I ( j ) indicates a synthesis of the 1st, 2nd, ... I The first scan result after the second scan j Individual point weights.

[0027] Compared with the prior art, the present invention has the following advantages: This invention proposes to divide the radar scanning range into several scattered points and assign corresponding weights to each scattered point according to the radar scanning results, thereby estimating the position of the target in the radar scanning coordinate system and realizing the on-orbit calibration of the photoelectric axis of the camera / radar. This solves the problem of high-precision on-orbit calibration of radar and camera without angle measurement function. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of the on-orbit calibration of the radar and camera according to the present invention; Figure 2 This is a flowchart of the on-orbit calibration process for the radar and camera of the present invention; Figure 3 The results of 100 simulations of target position estimation based on the algorithm presented in this invention are as follows; Figure 4 This is a diagram showing the division of different regions within the scanning range when the radar echo is strong, as described in an embodiment of the present invention. Figure 5 This is a diagram showing the division of different regions within the scanning range when the radar echo is weak, as described in an embodiment of the present invention. Detailed Implementation

[0029] Calibration requires obtaining the target satellite's position in the radar scanning system. However, microwave radars lacking differential channels do not have angle measurement capabilities. Therefore, accurately determining the target's position in the radar scanning coordinate system is crucial for calibration. This invention addresses the on-orbit calibration problem of radar and camera systems lacking angle measurement capabilities by proposing a weighted model-based on-orbit calibration method for camera / radar photoelectric axis deviation. By establishing a weighted model of the radar echo situation, false alarm rate, satellite pointing error, and possible distributions of the target at different positions in the radar scanning system during the scanning process, the target's position in the radar scanning system is estimated. In other words, radar scanning based on the weighted model enables the radar to measure the target's angle, thereby achieving on-orbit calibration of the radar and camera photoelectric axes.

[0030] The objective of this invention is achieved through the following technical solutions. Figure 1This demonstrates the basic principles of on-orbit calibration. Figure 2 The on-orbit calibration process was demonstrated.

[0031] Step 1: Select the target satellite for calibration, control the overall satellite attitude to point the camera's optical axis at the target satellite, and record the target's position under the camera system. T c ; First, select an on-orbit target satellite for calibration. The selection of the target satellite needs to consider two points: 1) the calibration satellite is within the radar's working range; 2) the magnitude of the calibration satellite is better than the detection sensitivity of the optical camera. Then, the entire satellite is adjusted in attitude so that the calibration satellite is located at the center of the optical camera's field of view. Step 2: Determine the radar scanning range and scanning step size, divide the radar's full field of view into several virtual points, and establish a radar scanning coordinate system; First, the scanning range is determined based on the possible relative positional changes of the antenna electrical axis and camera optical axis after they enter orbit, and the maximum deviation of the photoelectric axis that may be caused by on-orbit mechanical and thermal conditions is then considered. θ max It is estimated that the radar scanning range is no less than 2. θ max First, ensure that the radar beam emitted at a certain scanning point covers the target. Then, divide the scanning range into several scattered points. For example, if the scanning range is [-0.5°, 0.5°] × [-0.5°, 0.5°], the scattered points can be set to [-0.5°: 0.01°: 0.5°] × [-0.5°: 0.01°: 0.5°]. The denser the scattered points, the more accurate the calculation results will be, but the computational load will increase. Next, determine the scanning step size. The step size should be less than the radar half-beamwidth U. The smaller the step size, the higher the calibration accuracy, but the longer the calibration time. Finally, establish the radar scanning coordinate system. The scanning start point is when the camera optical axis points towards the target. Define the radar scanning coordinate system as follows: Origin O t It is the scan start point, X t The axis is along the X direction of the scanning plane, and the Y direction is... t The axis is along the Y direction of the scanning plane (X and Y are orthogonal), Z t The axis is along the optical axis of the camera and satisfies the right-hand rule, such as... Figure 1 As shown.

[0032] Step 3: Using satellite attitude control, the radar traverses the scanning range in a predetermined step size according to a specific method (helical scan, grid scan, etc.). The radar transmits a signal at each scanning point, and updates the weights of each point within the radar scanning range based on the echo intensity and weight model. Let the target position be T in the radar scanning coordinate system, and let the target position be T at the nth time. i During the next scan, the radar scanning position is Si The actual location of the radar beam center R i Due to satellite pointing error e The presence of radar scanning location S i and actual beam center R i They do not overlap.

[0033] set up P j (j=1,2,…, J If ) represents several discrete points (virtual scattered points) in the scanning area, then the weights of each scattered point can be updated according to the radar echo situation of each scan.

[0034] Specifically: Let the coordinates be in the radar scanning coordinate system, at the... i During the next scan, the theoretical scan point of the radar is located at... S i , Indicates the scan count index value. And it is a positive integer. Indicates the total number of scans; Virtual scatter points within the radar scanning area, divided by step size de. P j , express S i The index values ​​of virtual scatter points within a certain surrounding area. And it is a positive integer. This indicates the total number of scatter points within the radar's full field of view scanning area; Virtual scatter position P j These are virtual scattered points set up to assist in the calculation. There are a total of several scattered points, namely... J The exact value depends on the calibration accuracy desired by the designer. For example: When the scanning range is [-0.2°:0.2°]×[-0.2°:0.2°] and the scanning step size is 0.02°, the scanning points are approximately (0.4° / 0.02°). (0.4° / 0.02°) = 400, which means there will be 400 scans (400 scan points). S i The radar needs to sequentially target 400 scanning points and emit 400 beams. If we take several scattered points within the scanning range of [-0.2°:0.01:0.2°] × [-0.2°:0.01:0.2°], with a spacing of 0.01° between each point, then the number of virtual scattered points is approximately (0.4° / 0.01°). (0.4° / 0.01°) = 1600. The weights of these 1600 virtual points are updated based on the scan results during each scan.

[0035] When determining the weights of virtual scatter points at the current scan position based on echo intensity: If the radar echo signal-to-noise ratio is greater than the target detection signal-to-noise ratio, then the scanning point... S i Overall weight W i ( j )for: W i ( j )= w f j +(1- w ) g j If no radar echo is received or the echo signal-to-noise ratio is lower than the target detection signal-to-noise ratio, the scanning point S i The overall weight is: W i ( j )= u f j +(1- u ) g j w Indicates the probability of radar false alarms. u Indicates the probability of target detection; f j 、g j These represent the first and second basic weights of each virtual scatter point, determined based on the echo intensity.

[0036] 3.1 Determine the first basic weight f j At that time, by scanning point S i Set the center of the circle to 2n+1 regions (n=floor( e max / de (de represents the step size for scatter plotting), and the scatter plots within region 1 to region 2n+1. P j Corresponding to different basic weights fi floor ( () indicates rounding down; specifically: (1) Region 1 Scattered points in area 1 P j : ; Weights of each scatter point in region 1 f i : .

[0037] (2) Region k ( k =2~2n) area k Scattered points P j : ; area k Weights of each scatter point f i : .

[0038] (3) Region 2n+ 1 Scattered points outside the region 1 to 2n within the scan range P j Its weight f j =ε.

[0039] Where ε represents less than 10 -4 The minimum value; F ( . () indicates the attitude pointing error e The cumulative distribution function; e max This indicates a pointing error of 3σ or the maximum value; Dis (a,b) represents the angle between "a - our satellite - b": Dis ( S i , P j )represent S i Connection with satellite, P j The angle between the line connecting the point and the satellite; point P j The higher the weight, the greater the target distance. P j The closer, the greater the target distance. P j The farther away.

[0040] 3.2 Determine the basic weights g j At that time, by scanning point S i Set the center of the circle to 2n+1 regions (n ​​= floor( e max / de ), de represents the step size for scatter plotting), floor ( () indicates rounding down; scattered points within region 1 to region 2n+1 P j Corresponding to different basic weights g i Specifically: (1) Region 1 Scattered points in area 1 P j : ; Weights of each scatter point in region 1 g i : .

[0041] (2) Region k ( k =2~2n) area k Scattered points P j : ; area k Weights of each scatter point g i : .

[0042] (3) Region 2n+ 1 Scattered points outside the region 1 to 2n within the scan range P j Its weight g j =ε.

[0043] Where ε represents less than 10 -4 The minimum value; F ( () indicates satellite pointing error e The cumulative distribution function; e max This indicates a pointing error of 3σ or the maximum value; Dis (a,b) represents the angle between "a - our satellite - b". Dis ( S i ,P j )represent S i Connection with satellite, P j The angle between the line connecting to the satellite; scatter points P j The higher the weight, the greater the target distance. P j The closer, the greater the distance to the target. P j The farther away.

[0044] Step 4: Based on the weights of each scatter point W i (j) Calculate the target's coordinates in the radar scanning coordinate system.

[0045] First, based on the weights determined after each scan echo, the weights of each scatter point can be updated. Scatter points farther from the target will have their weights decrease rapidly after multiple scans, while the weights of scatter points closer to the target will increase. The weights of each scatter point are determined on the [number]th scan. I The following updates were made after the second scan:

[0046] Then, the positions and weights of each scatter point are weighted to estimate the target's position in the radar scanning system. T l :

[0047] Step 5: After scanning is complete, calculate the target's position in the radar scanning system. T l Combined with camera measurement results T c This allows for the calibration of the camera's optical axis and the radar's electronic axis.

[0048] Example: The invention will now be illustrated through specific examples.

[0049] A schematic diagram of on-orbit calibration of spaceborne radar / camera is shown below. Figure 1 As shown, the calibration flowchart is available here. Figure 2 Assume the radar half-beamwidth is U = 0.15°, and the satellite pointing error is e(e...). max =0.02°), the calibration steps are as follows: Step 1: The satellite issues a command, and the camera's optical axis is adjusted to point towards the satellite target via attitude control; Step 2: Determine the radar scanning range and scanning step size, and divide the scanning range into several scattered points; 1) Determine the scanning range as [-0.2°:0.2°] × [-0.2°:0.2°], with a scanning step size of 0.02°; 2) Within the scanning range, take several scattered points in the range of [-0.2°:0.01:0.2°]×[-0.2°:0.01:0.2°], with a spacing of 0.01° between each scattered point; Step 3: Traverse the scanning range in a certain way (spiral scan, raster scan, etc.) with a predetermined scanning step size; 1) At each radar scan position S i The radar is controlled to send signals to the target, and the weights of each scatter point are updated based on the echo intensity. a) Calculate the first basic weight of each scatter point within the scan range. f j If the scanning area division step size de is set to 0.01, the scanning range can be divided into region I, region II, region III, region IV, and region V, see [reference needed]. Figure 4 Different weights are assigned to each scatter point within each region based on the distribution of random errors. f j As shown in Table 1.

[0050] Table 1. Weight assignment results for each scatter point within the scan range

[0051] in, F ( ) is the cumulative distribution function of the satellite pointing error e; ε is less than 10. -4 The smaller value can be assigned after estimating the probability that the target may be located in region VI based on the radar echo intensity. Dis (a,b) represents the angle between "a - our satellite - b", scatter plot. P j The higher the weight, the greater the target distance. P j The closer, the greater the distance to the target. P j The farther away.

[0052] b) Calculate the second basic weight of each scatter point within the scan range. g j With a scan area division step size of 0.01, the scan range can be divided into region I, region II, region III, region IV, and region V. (See...) Figure 5 .

[0053] Different weights are assigned to the scattered points within each region. g jAs shown in the table below.

[0054] Table 2. Results of assigning the second basic weights to each scatter point within the scan range.

[0055] c) If the radar detects a target, the overall weight of each scatter point... W i ( j )for: W i ( j )= w f j +(1- w ) g j Among them, radar false alarm probability w (False alarm rate), detection probability is u .

[0056] d) If the radar does not detect the target, the overall weight of each scatter point. W i ( j )for: W i ( j )= u f j +(1- u ) g j 2) Repeat the above process, updating the weights of each scatter point after each scan. The weights of each scatter point are determined on the th scan. I The updates after the second scan are as follows:

[0057] Indicates the scan count index value. Indicates the sequence number of each scatter point; Indicates the total number of scans; Indicates the total number of scatter points; Step 4: Weight the positions and weights of each scatter point, and estimate the target's position in the radar scanning system using the following formula. T l The estimation error under 100 simulations is shown in the figure. Figure 3 .

[0058]

[0059] Figure 3 The results are from 100 Monte Carlo simulations performed on the example described in the specific implementation method. The innovation of this invention lies in establishing a weighted model to estimate the target in a radar scanning system, thus solving the on-orbit calibration problem for radars and cameras without angle measurement capabilities. The target estimation error based on the weighted model is as follows: Figure 5 As shown, the radar beamwidth is 0.3°, while the target estimation error is 0.0135°, and the maximum error in 100 Monte Carlo simulations does not exceed 0.0025°. Simulation results demonstrate that although the radar itself does not have an angle measurement function, the technical solution of this invention can achieve high-precision target position estimation, thereby solving the problem of on-orbit calibration of the photoelectric axis of camera / radar without angle measurement function.

[0060] To address the on-orbit calibration problem of combined microwave radar and camera detection without angle measurement capabilities, this invention proposes an on-orbit calibration process based on a weighted model, solving the challenge of estimating the target's position in the radar scanning coordinate system. This invention divides the radar scanning range into several scattered points and assigns corresponding weights to each point based on the radar's scanning results, thereby estimating the target's position in the radar scanning coordinate system and achieving on-orbit calibration of the camera / radar's photoelectric axes.

[0061] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A method for on-orbit calibration of camera and radar photoelectric axes based on a weighted model, characterized in that... include: Select the target for calibration, control the overall satellite attitude to make the camera optical axis point to the target, and record the target's position under the camera system; Determine the radar scanning range and scanning step size, establish the radar scanning coordinate system, and determine the radar scanning points; The radar sequentially scans the defined scanning range with a predetermined scanning step size. At each scanning point, the radar sends a radar beam and divides the radar scanning area into several virtual scattered points. The weight of each virtual scattered point in the current scanning range is determined based on the echo intensity and the position of the virtual scattered points. The weight of each virtual scattered point in the radar scanning range is continuously updated based on the radar echo situation of each scan. Calculate the target's coordinates in the radar scanning coordinate system based on the weights of each virtual scatter point; After scanning is completed, the target's position in the radar scanning system is calculated, and the angle between the camera's optical axis and the radar's electrical axis is determined by combining the target's position in the camera system, thus completing the calibration of the camera and radar.

2. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 1, characterized in that: The selection of an objective must simultaneously satisfy the following: The target is within the radar's operating range; The target's magnitude is superior to the detection sensitivity of an optical camera.

3. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 1, characterized in that: When determining the radar scanning range: Maximum photoelectric axis deviation caused by on-orbit thermal and mechanical conditions θ max It is estimated that the radar scanning range is no less than 2. θ max ; The scanning step size is less than the radar half-beamwidth U.

4. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 1, characterized in that: When the camera and radar are linked, and the radar scanning coordinate system is established: Point the camera's optical axis toward the target, establish a radar scanning coordinate system with the radar's current position, and take the direction of the radar's current electrical axis as the starting point for scanning.

5. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 1, characterized in that: The radar scanning area is divided into virtual scattered points, specifically: Let the coordinates be in the radar scanning coordinate system, at the... i During the next scan, the theoretical scan point of the radar is located at... S i , Indicates the scan count index value. And it is a positive integer. Indicates the total number of scans; Within the radar scanning area, in steps de Virtual scatter points are obtained by partitioning P j , express S i The index values ​​of virtual scatter points within a certain surrounding area. And it is a positive integer. This indicates the total number of virtual scatter points within the radar scanning area.

6. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 5, characterized in that: When determining the weights of virtual scatter points at the current scan position based on echo intensity: If the radar echo signal-to-noise ratio is greater than the target detection signal-to-noise ratio, then the... i The weights of each virtual scatter point during the second scan W i ( j )for: W i ( j )= w f j +(1- w ) g j If no radar echo is received or the echo signal-to-noise ratio is lower than the target detection signal-to-noise ratio, then the... i The weights of each virtual scatter point during the next scan are: W i ( j )= u f j +(1- u ) g j w Indicates the probability of radar false alarms. u Indicates the probability of target detection; f j 、g j These represent the first and second basic weights of each virtual scatter point, determined based on the echo intensity.

7. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 6, characterized in that: Determine the first basic weight f j At that time, by scanning point S i Set 2n+1 regions around the center, where n = floor ( e max / de floor () () indicates rounding down; When virtual scatter points P j The position satisfies: Then the first basic weight f j for: When virtual scatter points P j The position satisfies: in, k =2~2n and are integers, then the first basic weight f j for: When scatter P j Located outside the scanning range region 1 to 2n, the first basic weight f j for: f j = e Where ε represents less than 10 -4 The minimum value; F ( () indicates satellite pointing error e The cumulative distribution function; e max Indicates satellite pointing error e 3σ or maximum value; Dis ( S i , P j )represent S i Connection with satellite, P j The angle between the line connecting it to the satellite.

8. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 6, characterized in that: Determine the second basic weight g j At that time, by scanning point S i Set 2n+1 regions around the center, where n = floor ( e max / de floor () () indicates rounding down; When virtual scatter points P j The position satisfies: Then the second basic weight g j for: When virtual scatter points P j The position satisfies: in, k =2~2n and are integers, then the second basic weight g j for: When scatter P j Located outside the scanning range region 1 to 2n, the second basic weight g j for: g j = e Where ε represents less than 10 -4 The minimum value; F ( () indicates the attitude pointing error e The cumulative distribution function; e max This indicates a radar pointing error of 3σ or its maximum value; Dis ( S i , P j ) represents the scan point S i Connections to satellites, virtual scatter plots P j The angle between the line connecting it to the satellite.

9. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 6, characterized in that: Based on the combined results of multiple scans, the weights of each scatter point within the radar scan range are updated as follows: W i ( j ) indicates through the first i The first scan obtained the j The weights of the virtual scatter points W I ( j ) indicates a synthesis of the 1st, 2nd, ... I The first scan result after the second scan j The weights of each virtual scatter point.

10. The on-orbit calibration method for camera and radar photoelectric axes based on a weighted model according to claim 9, characterized in that: By weighting the positions and weights of each virtual scatter point, the position of the target in the radar scanning system is estimated. T l : W I ( j ) indicates a synthesis of the 1st, 2nd, ... I The first scan result after the second scan j Individual point weights.

Images