A method for on-orbit dark field correction of star sensors

Abstract

A method for on-orbit dark-field correction of a star sensor involves setting the star sensor to operate in window tracking mode, recording the pixel coordinates and grayscale of each frame of the tracking window image during on-orbit operation, and determining when a pixel has accumulated a certain number of frames. Then, the grayscale values ​​of that pixel are processed, and the average value is calculated after removing the highest grayscale values. This average value is used as the on-orbit dark-field correction coefficient for that pixel. Once the on-orbit dark-field correction coefficients for all pixels within the star sensor's detector imaging area have been calculated, on-orbit dark-field correction can be performed on the star sensor, thereby improving its measurement accuracy.

Description

Technical Field

[0001] This invention belongs to the field of star sensor detector calibration technology, and relates to an on-orbit calibration method for a star sensor photoelectric detector, including dark field coefficient calculation and dark field correction. Background Technology

[0002] Star sensors are key components in spacecraft attitude determination systems, widely used in aviation, aerospace, navigation, and weaponry. As space missions demand increasingly higher positioning accuracy from spacecraft, the overall spacecraft system places higher demands on the attitude measurement accuracy of space pointing instruments, which in turn sets higher requirements for the calibration accuracy of these instruments. Achieving high-precision attitude measurement with star sensors requires rigorous design and precise measurement of each component. The photodetector is the core component of the star sensor. During measurement, the photodetector receives photons converged by the optical system, and through the photoelectric effect, ultimately outputs an image of star points. The star sensor performs centering calculations on the imaged star points on the image plane and finally solves for the direction of the star sensor's optical axis. The quality of the star points imaged by the photodetector determines the accuracy of star point centering and ultimately affects the attitude measurement accuracy of the star sensor.

[0003] The quality of star-shaped spots in photodetector imaging depends primarily on the detector's systematic and random errors. Systematic errors mainly arise from the detector's response inhomogeneity and dark-field inhomogeneity, while random errors primarily stem from readout noise and quantum noise, which are Gaussian and Poisson noise, respectively. Improving the accuracy of photodetector imaging from the perspective of photodetectors mainly involves correcting or reducing both systematic and random errors. Random errors can be reduced through filtering or superposition, while systematic errors can be corrected through prior calibration in the laboratory.

[0004] However, as the star sensor is exposed to the space environment during its long-term operation in orbit, the detector's non-uniformity and noise characteristics will change. The ground calibration coefficients for the detector's system errors will no longer be applicable. Summary of the Invention

[0005] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a method for on-orbit calibration of the dark field correction coefficient of a star sensor photodetector. By using the dark background in orbit, the dark field coefficient of the star sensor detector is calibrated and corrected in orbit, thereby achieving accurate dark field correction of the star sensor in orbit and improving the on-orbit measurement accuracy of the star sensor.

[0006] The technical solution of this invention is: an on-orbit dark field correction method for a star sensor, comprising the following steps:

[0007] (1) When the star sensor is working in window tracking mode, determine whether the running angular velocity of the star sensor meets the requirements for dark field coefficient calculation. Proceed to the next step if and only if the requirements for dark field coefficient calculation are met.

[0008] (2) Establish an m×N queue to store the grayscale data of each pixel, where m is the number of pixels in the star point imaging range on the star sensor detector and N is the number of frames to be stored.

[0009] (3) Collect and record the column coordinate x, row coordinate y and gray value DN(x,y) of each pixel in each frame of the image in window tracking mode. Based on the column coordinate and row coordinate values ​​of the pixel, put the gray value of the pixel into an m×N queue. At the same time, calculate the cumulative frame number n of each pixel until the cumulative frame number n of each pixel is N.

[0010] (4) Calculate the on-orbit dark field coefficient D(x,y) for each pixel;

[0011] (5) Perform dark field correction on each pixel of each window in window tracking mode, specifically:

[0012] E'(x,y)=E(x,y)-D(x,y)+E0

[0013] Where E'(x,y) is the pixel grayscale value after dark field correction, E(x,y) is the original pixel grayscale value, and E0 is a constant.

[0014] Furthermore, the determination of whether the operating angular velocity of the star sensor meets the requirements for calculating the dark field coefficient is as follows: if the combined angular velocity v of the two directions perpendicular to the optical axis during the operation of the star sensor is greater than a*IFOV*T, then the operating angular velocity of the star sensor meets the requirements for calculating the dark field coefficient; where a is the Gaussian radius of the star spot pixel, IFOV is the instantaneous field of view of the star sensor, and T is the period of the star sensor's output attitude in window tracking mode.

[0015] Preferably, N > 20 frames.

[0016] Furthermore, the method for calculating the cumulative frame count n for each pixel is as follows: whenever a grayscale value added to this pixel queue increases by one, the cumulative frame count n+1, with n starting from 0.

[0017] Furthermore, the calculation of the on-orbit dark field coefficient D(x,y) for each pixel is specifically as follows:

[0018] 51) Calculate the sequence grayscale values ​​(E1, E2, E3, ..., E...) in this pixel queue. N The mean M and standard deviation σ of ( );

[0019] 52) Sequentially determine the grayscale values ​​of the sequence; if Ei If the value is greater than M+3×σ, then the gray value will be removed from the queue, i=1,2,3…N;

[0020] 53) Repeat steps 51) and 52) K times, where K is a positive integer;

[0021] 54) Calculate the mean value Me of the queue grayscale values ​​after the last removal operation, and use it as the on-orbit dark field coefficient D(x,y) of the pixel.

[0022] Preferably, K is 3, 4 or 5.

[0023] Furthermore, the value of E0 is the mean of D(x,y).

[0024] The advantages of this invention compared to existing technologies are as follows: The method of this invention calculates the on-orbit dark field coefficient using a window image obtained from a star sensor operating in window tracking mode. When the star sensor has an angular velocity, after a period of accumulation, the window image can traverse the effective imaging pixels of the detector. After acquiring multiple frames of grayscale data for each pixel, bright spots with high grayscale values ​​when a star passes through that pixel are removed, and the average value is calculated as the dark field coefficient for that pixel. Finally, this coefficient is used to correct for dark field non-uniformity errors. After the star sensor operates in orbit, its dark field noise will gradually change with the amount of irradiance it receives during on-orbit operation. On-orbit dark field calibration can calibrate the continuously changing dark field coefficient in real time, thereby performing corresponding corrections. Attached Figure Description

[0025] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0026] The present invention will now be described in further detail with reference to the accompanying drawings.

[0027] like Figure 1 The diagram shows a flowchart of the on-orbit dark field correction method for star sensors proposed in this invention. The method includes the following steps:

[0028] 1) When the star sensor is working in window tracking mode, determine whether the star sensor's angular velocity meets the requirements for calculating the dark field coefficient.

[0029] When the star sensor is in window tracking mode, it acquires multiple window images centered on the star positions on the detector's image plane by predicting the star positions. The number of windows is determined by the number of star points in the downward field of view of that direction in the star catalog.

[0030] This method improves computational efficiency by reusing stellar window data for on-orbit dark field calibration in window tracking mode.

[0031] The condition for determining whether the operating angular velocity of the star sensor meets the dark field coefficient calculation is that the combined angular velocity v in the two directions perpendicular to the optical axis during the operation of the star sensor is greater than a*IFOV*T, where a is the Gaussian radius of the star spot pixel, IFOV is the instantaneous field of view of the star sensor, and T is the period of the star sensor's output attitude in window tracking mode.

[0032] IFOV*T represents the displacement of the same star point on the image plane between two consecutive frames. The angular velocity is greater than a*IFOV*T to ensure that when the star point moves on the image plane, most of the grayscale data collected from the same pixel in multiple frames are dark background grayscale.

[0033] 2) When the operating angular velocity of the star sensor meets the requirements, the star sensor processing software establishes an m×N queue to store the grayscale data of each pixel.

[0034] Here, m is the number of pixels within the star-point imaging range on the star sensor detector, and N is the number of frames to be stored, where N > 20 frames. The larger N is, the larger the required storage space.

[0035] 3) The star sensor processing software collects and records the column coordinate x, row coordinate y, and gray value DN(x,y) of each pixel in each frame of the image in window tracking mode. Based on the column coordinate and row coordinate values ​​of the pixel, the gray value of the pixel is put into the queue, and the cumulative frame number n(x,y) of each pixel is calculated.

[0036] The cumulative frame count n(x,y) for each pixel is calculated as follows: whenever a grayscale value is added to the pixel queue, the cumulative frame count n+1 is added, and n starts from 0.

[0037] 4) If the cumulative frame count n(x,y) of a certain pixel reaches the maximum value N of the queue, then if there is more grayscale data for that pixel, it will no longer be stored.

[0038] 5) Once all pixel data within the effective imaging range has reached the maximum value N, begin calculating the on-orbit dark field coefficient for each pixel.

[0039] The steps for calculating the on-orbit dark field coefficient of a certain pixel are as follows:

[0040] 51) Calculate the sequence grayscale values ​​(E1, E2, E3, ..., E...) in this pixel queue. N The mean M and standard deviation σ of ( );

[0041] 52) Sequentially determine the grayscale values ​​of the sequence; if E i If the value is greater than M+3×σ, then the gray value will be removed from the queue, thus removing the brighter gray values, i=1,2,3…N;

[0042] 53) Repeat steps 51) and 52) 3 to 5 times to calculate the mean and standard deviation and remove brighter grayscale values.

[0043] 54) Calculate the mean Me of the queue gray values ​​that have been removed from the brighter gray values, and use it as the on-orbit dark field coefficient D(x,y) of the pixel, where x and y are the column coordinates and row coordinates of the pixel, respectively.

[0044] 6) After completing the on-orbit dark field coefficient calculation, dark field correction will be performed on each pixel of each window in the window tracking mode, and the corrected grayscale value will be used to complete the subsequent centroid extraction and attitude calculation.

[0045] The correction formula for dark field correction of each pixel in each window in window tracking mode is:

[0046] E'(x,y)=E(x,y)-D(x,y)+E0

[0047] Where: x and y are the column coordinates and row coordinates of the pixel, respectively; E' is the pixel grayscale value after dark field correction; E is the original pixel grayscale value; and E0 is a constant, which takes the mean value of D(x,y) to make the corrected pixel grayscale value positive.

[0048] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A method for on-orbit dark field correction of a star sensor, characterized in that: Includes the following steps: (1) When the star sensor is working in window tracking mode, determine whether the running angular velocity of the star sensor meets the requirements for dark field coefficient calculation. Specifically, if the combined angular velocity v of the two directions perpendicular to the optical axis during the operation of the star sensor is greater than a*IFOV*T, then the running angular velocity of the star sensor meets the requirements for dark field coefficient calculation. Where a is the Gaussian radius of the star spot pixel, IFOV is the instantaneous field of view of the star sensor, and T is the period time of the star sensor's output attitude in window tracking mode. Proceed to the next step if and only if the requirements for dark field coefficient calculation are met. (2) Establish an m×N queue to store the grayscale data of each pixel, where m is the number of pixels in the star point imaging range on the star sensor detector and N is the number of frames to be stored. (3) Collect and record the column coordinate x, row coordinate y and gray value DN(x,y) of each pixel in each frame of the image in window tracking mode. Based on the column coordinate and row coordinate values ​​of the pixel, put the gray value of the pixel into an m×N queue. At the same time, calculate the cumulative frame number n of each pixel until the cumulative frame number n of each pixel is N. (4) Calculate the on-orbit dark field coefficient D(x,y) for each pixel; specifically: 41) Calculate the sequence grayscale values ​​(E1, E2, E3, ..., E...) in this pixel queue. N The mean M and standard deviation σ of ( ); 42) Sequentially determine the grayscale values ​​of the sequence; if E i If the value is greater than M+3×σ, then the gray value will be removed from the queue, i=1,2,3…N; 43) Repeat steps 41) and 42) K times, where K is a positive integer; 44) Calculate the mean value Me of the grayscale values ​​in the queue after the last removal operation, and use it as the on-orbit dark field coefficient D(x,y) of the pixel; (5) Perform dark field correction on each pixel of each window in window tracking mode, specifically: E'(x,y)=E(x,y)-D(x,y)+E0 Where E'(x,y) is the pixel grayscale value after dark field correction, E(x,y) is the original pixel grayscale value, and E0 is a constant.

2. The method for on-orbit dark field correction of a star sensor according to claim 1, characterized in that: The N>20 frames.

3. The method for on-orbit dark field correction of a star sensor according to claim 1, characterized in that: The method for calculating the cumulative frame count n for each pixel is as follows: whenever a grayscale value added to this pixel queue increases by one, the cumulative frame count n+1, with n starting from 0.

4. The method for on-orbit dark field correction of a star sensor according to claim 1, characterized in that: The value of K is 3, 4, or 5.

5. The on-orbit dark field correction method for a star sensor according to claim 4, characterized in that: The value of E0 is the mean of D(x,y).

Images