A method for pre-treating a star sensor damaged by nuclear radiation neutrons

Abstract

The present application belongs to the technical field of shielding radiation of photoelectric sensor, and specifically relates to a method for preprocessing a star sensor damaged by neutron radiation in nuclear radiation, which comprises the following steps: step 1, analyzing a photoelectric image damaged by neutron radiation to obtain defect distribution response characteristics; step 2, according to the relationship between the defect response characteristics and the gray scale, preliminarily processing the defects in the star sensor image; and step 3, reprocessing the defects with large deviation after compensation and correction, wherein the correctness of the defects after correction is judged through correlation due to the high correlation of the star sensor image. The method provided by the present application improves the neutron radiation resistance of equipment, improves the detection image, maintains the measurement accuracy, prolongs the service life of the equipment, increases the service life of the photoelectric measuring equipment by more than 20 times, and maintains the high accuracy of the equipment.

Description

Technical Field

[0001] This invention relates to the field of radiation shielding technology for photoelectric sensors, and in particular to a pretreatment method for a star sensor damaged by nuclear radiation neutrons. Background Technology

[0002] Nuclear radiation decays primarily consist of alpha radiation (helium nuclei), gamma radiation (photons, electromagnetic radiation), and neutron radiation. Among these, neutron radiation has a relatively low dose but high energy. It contains no electric charge, so it can directly enter the atomic nucleus and undergo nuclear reactions upon collision, producing secondary radiation. Very few substances can effectively block neutron radiation.

[0003] In existing technologies, large nuclear facilities can be shielded using large concrete domes, large amounts of water, and thick boron-containing polyethylene, but this significantly increases weight and space requirements. However, shielding is practically impossible in special applications such as missiles. This is because neutron radiation damages all photoelectric sensors, and protection can only be achieved through external shielding; the damage cannot be mitigated by strengthening the sensor chips themselves or optimizing circuit design.

[0004] Nuclear radiation can damage the silicon lattice of photoelectric sensors (CCD, CMOS), leading to increased device defects, manifested as decreased transfer efficiency, increased dark current, and more blemishes. Under high doses of radiation or prolonged exposure, photoelectric sensors exhibit significant performance degradation.

[0005] Photoelectric sensors damaged by neutron radiation exhibit numerous defects in their images, as shown in the image. Figure 1 All defects are caused by flaws in the sensor's photosensitive material, resulting in current leakage into the pixel potential well. The leakage current equals the current magnitude multiplied by the leakage time (exposure and data conversion time). Neutron radiation-damaged photoelectric sensor images are densely packed with defects. Processing defects using the general photoelectric sensor defect handling method (taking the average grayscale value of four neighboring pixels to represent the defect) presents several problems:

[0006] 1. Defect assessment is difficult to determine because neutrons are emitted from the radiation source in all directions, causing varying degrees of damage to each pixel on the photoelectric sensor. Some pixels will have higher leakage current and be brighter, while others will have lower leakage current and be darker. Therefore, there is no unified standard for determining which pixels are defects, as it requires replacing the grayscale values ​​of surrounding pixels.

[0007] 2. Even if a defect judgment threshold is selected, the number of defects is too large. Basically, several defects will be concentrated in a target area. The gray value of the defect is replaced by the gray value of the surrounding pixels, resulting in a large loss of target information and a significant decrease in the accuracy of target centroid extraction.

[0008] 3. The pixels used to replace defects are also damaged, and their grayscale cannot reflect the true grayscale of light energy conversion. The image obtained by the grayscale replacement method cannot truly reflect the received light energy image. Using such an image to perform high-precision centroid calculation of small targets will result in a large error.

[0009] The long-term operation of photoelectric imaging measurement equipment under neutron radiation is a completely new field in China, lacking research and relevant data internationally. According to general damage assessment requirements for photoelectric sensors, the service life of such equipment under neutron radiation is generally no more than one year. Summary of the Invention

[0010] This invention addresses the impact of long-term neutron radiation. In cases where photoelectric sensors are already damaged, a special software image processing method is used to mitigate the effects of neutron radiation damage on the function and performance of star sensors. This achieves the goal of maintaining or minimizing the reduction in measurement sensitivity and accuracy, effectively improving the lifespan, performance, and reliability of star sensors.

[0011] Specifically, the present invention provides a pretreatment method for a star sensor damaged by nuclear radiation neutrons, the pretreatment method comprising the following steps:

[0012] Step 1: Analyze the photoelectric images after neutron radiation damage to obtain the defect distribution response characteristics;

[0013] Step 2: Based on the relationship between defect response characteristics and grayscale, perform preliminary processing on defects in the star sensor image;

[0014] Step 3: If some defects have large deviations after compensation and correction, they need to be reprocessed. Since the star sensor images have high correlation, the correctness of the defect correction is judged by the correlation.

[0015] Step 2 also includes the following steps:

[0016] Step 21: During the power-on self-test of the star sensor, a dark image is acquired as the background image;

[0017] Step 22: The star sensor acquires a 12-bit image with a maximum grayscale of 4095 and performs high-pass filtering.

[0018] Step 23: For defects in the image with a gray value less than 1024, the current gray value is subtracted from the gray value of the background image and treated as a normal pixel.

[0019] Step 24: For defects with grayscale values ​​of 1024 to 2047 in the processed image, compensate the response grayscale according to a fixed ratio;

[0020] Step 25: For defects in the image with a gray value greater than or equal to 2048, the neighborhood interpolation method or median substitution method shall be used for processing.

[0021] Furthermore, step 1 also includes the following steps:

[0022] Step 11: Analyze the photoelectric images after neutron radiation damage. Acquire dark images under dark conditions with 200ms exposure at room temperature and 200ms exposure at 58℃. Extract defects with a grayscale threshold of 64.

[0023] Step 12: Capture one frame of low-frequency illuminated image with the lens facing a white background, and subtract the gray values ​​of each pixel in the illuminated image from the gray values ​​of each pixel in the dark image.

[0024] Step 13: Calculate the average gray value of the 5*5 pixels around the defect in the subtracted image. Divide the gray value of the pixel at the defect in the subtracted image by the average gray value to obtain the response efficiency of this defect.

[0025] Furthermore, in step 24, the fixed ratio is:

[0026]

[0027] Where x represents the gray level of the defect under no-light conditions.

[0028] Furthermore, in step 25, if there are no defects in the neighboring pixels, the grayscale of the defects is replaced by the neighboring interpolation method; if there are still defects in the neighboring pixels, the median is used for replacement.

[0029] Furthermore, in step 3, the correlation determination method is as follows:

[0030] Centered on a pixel, take a 3×3 pixel block and calculate its grayscale mean and variance:

[0031]

[0032] To obtain a larger threshold, Indicates the grayscale mean. Represents variance. This represents the threshold.

[0033] The beneficial effects achieved by this invention are:

[0034] For a long time, in the nuclear neutron radiation environment, only physical shielding measures can be used for protection, such as using thick cement walls or large water pools for protection; otherwise, photoelectric detection equipment will not be able to work at all.

[0035] This invention addresses photoelectric detection and measurement equipment that must operate long-term in nuclear radiation environments but lacks effective shielding. It proposes, for the first time, a software algorithm-based method to enhance the equipment's resistance to neutron radiation, improve detection images, maintain measurement accuracy, and extend its service life. This method increases the lifespan of photoelectric measurement equipment by more than 20 times while maintaining high precision. Attached Figure Description

[0036] Figure 1 This is a preprocessing method for a nuclear radiation neutron-damaged star sensor, which involves exposing the sensor to a dark image at a high temperature (58°C) for 200ms.

[0037] Figure 2 This is a schematic diagram of the star sensor image processing and target detection process in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0038] Figure 3 This is a schematic diagram of the defect distribution of a star sensor at room temperature with an exposure time of 200ms, in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0039] Figure 4 This is a schematic diagram illustrating the statistical defects of a sensor exposed at room temperature for 200ms during a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0040] Figure 5 This is a schematic diagram of the grayscale response characteristics of defects in a sensor with a 200ms exposure time at room temperature, in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0041] Figure 6 This is a schematic diagram of the defect distribution during a high-temperature (58℃) 200ms exposure time in a pretreatment method for a nuclear radiation neutron-damaged star sensor.

[0042] Figure 7 This is a schematic diagram illustrating the statistical defects of a star sensor with a high temperature (58℃) exposure time of 200ms in a pretreatment method for a nuclear radiation neutron-damaged star sensor.

[0043] Figure 8 This is a schematic diagram of the grayscale response characteristics of defects during a high-temperature (58℃) exposure time of 200ms in a pretreatment method for a nuclear radiation neutron-damaged star sensor.

[0044] Figure 9 This is a 12-bit bright background image displayed in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0045] Figure 10 This is a schematic diagram of four-neighbor interpolation or median interpolation in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0046] Figure 11This is an embodiment of a star sensor image quality measurement method for a star sensor preprocessing method for nuclear radiation neutron-damaged star sensors.

[0047] Figure 12 This is a 12-bit display of the defect distribution near a target star in a preprocessing method for a nuclear radiation neutron-damaged star sensor.

[0048] Figure 13 This is a schematic diagram illustrating the detection of a large number of false targets in current defect detection methods. Detailed Implementation

[0049] The technical solution of the present invention will be described in more detail below with reference to the accompanying drawings. The present invention includes, but is not limited to, the following embodiments.

[0050] like Figure 2 As shown, the present invention provides a pretreatment method for a nuclear radiation neutron-damaged star sensor, which includes the following steps:

[0051] Step 1: Analyze the photoelectric images after neutron radiation damage to obtain the defect distribution response characteristics;

[0052] Furthermore, step 1 also includes the following steps:

[0053] Step 11: Analyze the photoelectric images after neutron radiation damage. Acquire dark images under dark conditions with 200ms exposure at room temperature and 200ms exposure at 58℃. Extract defects with a grayscale threshold of 64.

[0054] Step 12: Capture one frame of low-frequency illuminated image with the lens facing a white background, and subtract the gray values ​​of each pixel in the illuminated image from the gray values ​​of each pixel in the dark image.

[0055] Step 13: In the image after subtraction, calculate the average gray value of the 5*5 pixels around the defect (excluding the defect). Divide the gray value of the pixel at the defect in the subtracted image by the average gray value to obtain the response efficiency of this defect.

[0056] As attached Figure 3-8 As shown, at room temperature, defects are mostly isolated points, evenly distributed, with a small number of adjacent defects. As the temperature increases, the number of defects in the sensor image increases significantly, from 161 to 928, and the maximum grayscale value of the defects also increases from 488 to 3248. Defect statistics are performed using clustering methods to count the number of spots in the image. (See attached image.) Figure 7 The "total number" refers to the number of spots. Two spots connected together are counted as one spot, so the total number of spots is 921 = 914 + 14 / 2.

[0057] As can be seen from the defect grayscale-response efficiency chart, the response efficiency is highly correlated with its original grayscale value. As the defect grayscale value increases, the response efficiency decreases linearly. For defect grayscale values ​​below 1000, the response efficiency is generally >85%, while for defect grayscale values ​​above 1200, the response efficiency is below 80%. All defects respond to illumination.

[0058] At the same time, such as Figure 9 As shown, when a frame of bright background image is captured and black bad pixels are detected, there are no black bad pixels that cannot be detected, which is consistent with the leakage characteristics of the defect potential well, that is, the defect can only be a bright spot.

[0059] Step 2: Based on the relationship between defect response characteristics and grayscale, perform preliminary processing on defects in the star sensor image;

[0060] Step 21: During the power-on self-test of the star sensor, a dark image is acquired as the background image;

[0061] Step 22: The star sensor acquires a frame of image, which is represented in 12 bits with a maximum grayscale of 4095, and performs high-pass filtering to obtain the processed image.

[0062] Step 23: For defects in the image with a gray value less than 1024, based on the defect distribution response characteristics analysis, the photoelectric conversion efficiency of the defect is greater than 85%. The current gray value is subtracted from the gray value of the background image, and it is still treated as a normal pixel.

[0063] Step 24: For defects in the image with grayscale values ​​between 1024 and 2047, compensate for the response grayscale according to a linear equation. The linear equation is:

[0064]

[0065] Where x is the gray level of the defect under no light conditions; A is the slope, B is the intercept, and B is a constant.

[0066] For different detector models, A and B differ. For the same detector model, A and B are fixed values ​​in the linear equation. A and B are obtained through defect grayscale-response efficiency statistics, as in this product. .

[0067] Defect response characteristic statistical method: First, cut out defects in the dark (black) image using a fixed threshold (e.g., grayscale 64 as the threshold); then capture a low-frequency (gradually changing) illuminated image of a white paper with the lens; subtract the dark image from the illuminated image; calculate the average grayscale value of 5*5 pixels at each defect in the subtracted image (excluding the defect itself); divide the grayscale value of the pixels at the defect in the subtracted image by the average grayscale value to obtain the response efficiency of this defect.

[0068] After the defects were corrected, most of the defects were restored to the correct grayscale. The remaining defects were relatively few and discretely distributed, which can serve as the basis for the next step of processing.

[0069] Step 25: For defects in the image with a gray value greater than or equal to 2048, compensation is not suitable due to the low photoelectric conversion response efficiency and large fluctuations.

[0070] If there are no defects in the neighboring pixels, the defect grayscale is replaced using the 4-neighbor interpolation method:

[0071]

[0072] If neighboring pixels also have defects, median substitution is used, as follows: Figure 10 As shown, if both points (0,0) and (-1,0) in the figure are defects, then the median gray value of all pixels except (0,0) is used to replace pixel (0,0).

[0073] Using this defect handling method, defects with a grayscale value <2048 are processed first to ensure that the grayscale value of most defects has been restored to normal. On this basis, defects with a grayscale value ≥2048 and some incorrectly processed defects are processed using a neighborhood interpolation method.

[0074] Step 3: If some defects have large deviations after compensation and correction, they need to be reprocessed. Since the star sensor images have high correlation, the correctness of the defect correction is judged by the correlation.

[0075] The determination method is as follows:

[0076] Centered on a pixel, take a 3×3 pixel block, and calculate the grayscale mean and variance (excluding the center pixel). For the threshold, To obtain a larger threshold, Indicates the grayscale mean. Represents variance. This indicates the threshold; if the value exceeds the threshold, the pixel is dynamically judged to need further correction.

[0077] The selection criteria for the variance coefficient 6 and the cutoff threshold 20 are as follows:

[0078] The star image is Gaussian, with a standard deviation between 1 and 1.2. Therefore, the ratio of pixel grayscale value to variance in the star image, excluding the center point, is at most approximately 2, so a coefficient of 3 is used. The star center is calculated at the sub-pixel level, with a pixel grayscale value to variance ratio of 5.4. Therefore, a coefficient of 6 is used, which satisfies the purpose of removing defects without processing the actual star points.

[0079] The image preprocessing workflow includes 3×3 smoothing. The impact of isolated defects on grayscale is divided by 9. In the star target sensitivity calculation, the variance benchmark is extracted as 6. Using a 20-fold truncation threshold avoids calculating and comparing for each pixel, significantly reducing the computational load. On the other hand, after smoothing, defects with grayscale ≤20 have negligible impact on target position and sensitivity below the truncation threshold.

[0080] In one embodiment, 11 CCD (Computer-Controlled Discharge) photoelectric sensor samples from the neutron radiation acceleration experiment were selected, with an equivalent radiation period of 10 years. These were installed in a star sensor assembly, and the target surface was adjusted and calibrated. The assembly's self-test, single-star accuracy, single-point stability, false target detection rate, and target miss detection rate were tested at an operating temperature of 55°C. The single-point stability, false target detection rate, and target miss detection rate were statistically analyzed under the condition of a 5.5 magnitude star output from the single-star simulator (as required by the star measurement sensitivity). Figure 11 These are real defect images of the star sensor at its operating temperature of 55℃.

[0081] Single-star angle measurement accuracy test:

[0082] Place the star sensor on a high-precision turntable, with a single-star simulator in front of it. Rotate the turntable in two dimensions at 0.5° intervals, measure the angular error between the theoretical star position and the output position of the star sensor, and calculate the variance of the position deviation across the entire field of view, which is the single-star angular measurement accuracy.

[0083] Table 1 Single-satellite accuracy test data

[0084]

[0085]

[0086] 121.711625 88.711928 -53.524776 88.680673 1.117348 1.116646 2.527849 137.05165 88.391873 -39.482043 88.335142 1.413119 1.412998 0.43498 41.174708 87.576812 -137.461726 87.673414 2.501667 2.501984 -1.141431 50.31642 87.926727 -127.703264 88.015225 2.122826 2.122488 1.217362 62.694041 88.204178 -114.461216 88.277281 1.803981 1.803334 2.328945 78.531498 88.371692 -97.772491 88.41928 1.581837 1.581782 0.197487 96.304608 88.394482 -79.63272 88.410639 1.499993 1.499507 1.749242 112.970152 88.266829 -63.24952 88.252609 1.580397 1.579417 3.528273 126.40123 88.017585 -50.441679 87.979563 1.801454 1.800605 3.054283

[0087] From the “Deviation” column of Table 1, the variance of the single-star angle measurement accuracy is 1.38".

[0088] The centroid of the star point in the data set "Turntable azimuth 210.752754, Turntable zenith angle 89.209599, Star sensor output azimuth 29.376324, Star sensor zenith angle 89.105418, Turntable included angle 0.707615, Star sensor and zero position included angle 0.707756, Deviation -0.506943" is (608.996258, 664.021774). Taking (609, 664) as the center point, a 9×9 pixel block is taken, as follows... Figure 12 As shown, there are three relatively bright defects in this area, and their grayscale values ​​are shown in Table 2. The original image contained multiple defects near the star target. After defect processing, the star target could be extracted stably, and the sensitivity and accuracy met the requirements.

[0089] Table 2 Defect Location and Gray Scale

[0090]

[0091] Single-star stability test:

[0092] Place the star sensor on a high-precision turntable, with a single-star simulator in front of it, outputting a 5.5 magnitude star. Rotate the turntable so that the simulated star is imaged at the center and four corners of the star sensor's field of view. The star sensor outputs the star pixel coordinates (each pixel corresponds to an angle of 19"). Calculate the fluctuation (variance) of the star position, i.e., the single-point stability.

[0093] Table 3 Single-point stability test data

[0094] X coordinate Y coordinate X coordinate Y coordinate X coordinate Y coordinate X coordinate Y coordinate X coordinate Y coordinate 977.507 751.624 979.000 197.908 425.505 197.140 424.136 750.866 701.528 474.409 977.558 751.576 979.012 197.979 425.495 197.151 424.106 750.926 701.554 474.367 977.555 751.578 979.022 197.919 425.448 197.182 424.095 750.936 701.542 474.363 977.566 751.532 979.016 197.964 425.520 197.177 424.102 750.900 701.529 474.359 977.585 751.565 979.026 197.918 425.483 197.160 424.111 750.906 701.529 474.367 977.609 751.531 979.024 197.869 425.439 197.161 424.129 750.932 701.494 474.343 977.518 751.519 978.988 197.944 425.481 197.139 424.110 750.973 701.537 474.326 977.574 751.572 979.000 197.945 425.439 197.143 424.090 750.905 701.501 474.356 977.550 751.544 978.994 197.986 425.492 197.208 424.077 750.900 701.525 474.389 977.528 751.601 979.008 197.857 425.490 197.190 424.055 750.906 701.546 474.380 977.524 751.614 978.978 197.891 425.451 197.147 424.090 750.986 701.557 474.347 977.583 751.595 979.001 197.909 425.413 197.152 424.069 750.947 701.555 474.335 977.537 751.571 979.004 197.897 425.390 197.196 424.106 750.884 701.570 474.399 977.513 751.628 979.035 197.965 425.455 197.182 424.158 750.861 701.536 474.415 977.464 751.580 979.009 197.900 425.399 197.181 424.106 750.904 701.549 474.347 977.505 751.587 979.025 197.923 425.388 197.210 424.100 750.919 701.540 474.348 977.527 751.639 979.013 197.915 425.414 197.183 424.129 750.884 701.575 474.354 977.519 751.565 978.970 197.892 425.483 197.139 424.198 750.941 701.602 474.411 977.523 751.527 978.975 197.984 425.464 197.136 424.104 750.939 701.501 474.395 977.513 751.572 979.000 197.902 425.468 197.168 424.176 750.896 701.554 474.387 977.531 751.540 978.985 197.935 425.411 197.186 424.120 750.909 701.610 474.411 977.509 751.608 979.021 197.926 425.422 197.130 424.146 750.863 701.537 474.385 977.548 751.548 978.979 197.977 425.400 197.187 424.116 750.889 701.503 474.413 977.505 751.593 978.972 197.925 425.440 197.129 424.130 750.906 701.546 474.331 977.539 751.527 978.990 197.922 425.455 197.130 424.094 750.918 701.559 474.360 977.488 751.576 978.982 197.867 425.453 197.160 424.100 750.882 701.546 474.368 977.541 751.541 978.955 197.919 425.433 197.169 424.159 750.914 701.492 474.354 977.464 751.661 979.001 197.940 425.451 197.207 424.077 750.867 701.540 474.290 977.493 751.532 978.982 197.889 425.455 197.175 424.094 750.884 701.532 474.329 977.549 751.536 978.951 197.899 425.485 197.214 424.123 750.904 701.505 474.309

[0095] 977.552 751.589 979.021 197.925 425.430 197.223 424.123 750.887 701.470 474.333 977.566 751.557 979.037 197.909 425.435 197.170 424.084 750.872 701.527 474.369 977.524 751.541 979.078 197.888 425.499 197.149 424.094 750.914 701.486 474.331 977.481 751.533 978.962 197.928 425.424 197.174 424.076 750.896 701.511 474.404 977.562 751.589 978.973 197.927 425.463 197.182 424.111 750.924 701.549 474.389 977.480 751.614 979.038 197.921 425.474 197.109 424.149 750.901 701.567 474.328 977.463 751.562 978.961 197.940 425.439 197.149 424.090 750.947 701.576 474.344 977.472 751.484 978.978 197.909 425.452 197.160 424.060 750.904 701.592 474.328 977.570 751.479 978.992 197.904 425.456 197.178 424.108 750.878 701.544 474.308 977.580 751.466 978.994 197.893 425.453 197.158 424.052 750.901 701.503 474.303 977.496 751.558 979.012 197.858 425.474 197.226 424.081 750.922 701.549 474.328 977.580 751.571 978.961 197.895 425.442 197.223 424.086 750.921 701.589 474.359 977.549 751.654 978.965 197.900 425.429 197.158 424.069 750.927 701.523 474.352 977.588 751.631 978.934 197.904 425.486 197.184 424.078 750.875 701.527 474.367 977.546 751.520 978.969 197.965 425.452 197.188 424.095 750.953 701.524 474.300 977.498 751.528 979.060 197.888 425.382 197.136 424.125 750.930 701.554 474.368 977.548 751.584 979.020 197.941 425.369 197.231 424.142 750.918 701.509 474.330 977.537 751.514 978.934 197.941 425.468 197.139 424.067 750.910 701.499 474.354 977.529 751.546 978.999 197.932 425.439 197.217 424.125 750.913 701.540 474.370 977.529 751.489 979.042 197.940 425.427 197.135 424.091 750.918 701.536 474.387 977.501 751.535 978.983 197.905 425.486 197.195 424.063 750.867 701.486 474.363 977.476 751.524 978.939 197.871 425.439 197.183 424.110 750.888 701.523 474.318 977.497 751.494 978.986 197.980 425.433 197.174 424.130 750.893 701.524 474.399 977.468 751.572 978.943 197.892 425.439 197.204 424.130 750.922 701.519 474.343 977.459 751.635 979.025 197.927 425.425 197.160 424.078 750.922 701.498 474.365 977.414 751.618 978.947 197.896 425.407 197.188 424.137 750.914 701.523 474.357 977.422 751.532 979.027 197.906 425.461 197.200 424.111 750.874 701.514 474.286 977.468 751.496 978.978 197.918 425.455 197.203 424.110 750.867 701.555 474.343 977.488 751.531 979.033 197.878 425.412 197.210 424.117 750.892 701.531 474.360 977.484 751.562 978.995 197.888 425.368 197.221 424.065 750.921 701.559 474.342 977.479 751.541 978.992 197.883 425.387 197.178 424.106 750.861 701.503 474.380 977.479 751.557 979.021 197.931 425.429 197.196 424.115 750.941 701.564 474.367 977.550 751.622 978.938 197.848 425.448 197.187 424.094 750.935 701.520 474.344 977.533 751.565 978.999 197.912 425.405 197.162 424.111 750.893 701.522 474.404 977.550 751.566 979.009 197.887 425.447 197.188 424.136 750.939 701.497 474.373 977.490 751.536 979.046 197.930 425.447 197.161 424.045 750.965 701.503 474.365 977.510 751.530 979.025 197.895 425.433 197.199 424.076 750.927 701.541 474.348 977.558 751.530 978.995 197.895 425.416 197.242 424.103 750.919 701.548 474.291 977.524 751.523 978.996 197.935 425.460 197.212 424.093 750.924 701.503 474.380 977.579 751.540 978.991 197.902 425.438 197.190 424.041 750.940 701.564 474.367 977.544 751.543 979.048 197.932 425.403 197.171 424.104 750.965 701.520 474.344 977.553 751.536 978.978 197.944 425.450 197.188 424.108 750.913 701.567 474.339 977.529 751.583 978.977 197.895 425.508 197.239 424.065 750.876 701.561 474.277

[0096] 977.510 751.580 978.978 197.901 425.509 197.190 424.089 750.857 701.499 474.286 977.483 751.589 979.009 197.925 425.438 197.269 424.111 750.888 701.564 474.294 977.539 751.595 978.981 197.896 425.427 197.179 424.108 750.906 701.471 474.363 977.566 751.539 979.052 197.875 425.453 197.213 424.082 750.896 701.554 474.321 977.509 751.549 978.977 197.914 425.457 197.196 424.065 750.928 701.514 474.378 977.516 751.493 978.998 197.899 425.430 197.207 424.124 750.924 701.460 474.355 977.527 751.541 979.017 197.944 425.477 197.230 424.060 750.924 701.574 474.461 977.492 751.602 979.022 197.941 425.457 197.191 424.106 750.875 701.518 474.419 977.522 751.574 978.986 197.862 425.411 197.182 424.138 750.854 701.502 474.382 977.555 751.571 979.000 197.906 425.482 197.209 424.080 750.962 701.540 474.329 977.575 751.541 979.030 197.977 425.379 197.186 424.116 750.923 701.509 474.333 977.596 751.476 979.033 197.925 425.400 197.197 424.117 750.909 701.493 474.399 977.622 751.548 978.948 197.849 425.431 197.145 424.152 750.852 701.549 474.386 977.552 751.497 978.979 197.896 425.486 197.264 424.146 750.865 701.563 474.396

[0097] Single-point stability: x-axis 0.81", y-axis 0.83", 5.5 magnitude star stability ≤1".

[0098] False target rate and false negative rate:

[0099] The star sensor was placed on a high-precision turntable, with a single-star simulator in front of it, outputting a 5.5 magnitude star. Maintaining an operating temperature of 55°C, 1000 frames of star detection data were recorded. No false targets were detected, and no targets were missed. Without the processing algorithm of this invention, a large number of false targets would be detected. Figure 13 As shown.

[0100] The software algorithm of this invention solves the problem of a significant increase in sensor defects, increases the adaptability of star sensors to nuclear radiation environments, extends their working life, and ensures the sensitivity and measurement accuracy of star sensors.

[0101] The invention is not limited to the specific embodiments described above. Those skilled in the art can implement the invention using other specific embodiments based on the disclosed content of the embodiments and accompanying drawings. Therefore, any design that adopts the design structure and concept of the invention and makes some simple changes or modifications falls within the protection scope of the invention.

Claims

1. A method for preprocessing a nuclear radiation neutron-damaged star sensor, characterized in that, The pretreatment method for nuclear radiation neutron-damaged star sensors includes the following steps: Step 1: Analyze the photoelectric images after neutron radiation damage to obtain the defect distribution response characteristics; Step 2: Based on the relationship between defect response characteristics and grayscale, perform preliminary processing on defects in the star sensor image; Step 3: If some defects have large deviations after compensation and correction, they need to be reprocessed. Since the star sensor images have high correlation, the correctness of the defect correction is judged by the correlation. Step 1 also includes the following steps: Step 11: Analyze the photoelectric images after neutron radiation damage. Acquire dark images under dark conditions with 200ms exposure at room temperature and 200ms exposure at 58℃. Extract defects with a grayscale threshold of 64. Step 12: Capture one frame of low-frequency illuminated image with the lens facing a white background, and subtract the gray values ​​of each pixel in the illuminated image from the gray values ​​of each pixel in the dark image. Step 13: Calculate the average gray value of the 5*5 pixels around the defect in the subtracted image. Divide the gray value of the pixel at the defect in the subtracted image by the average gray value to obtain the response efficiency of this defect. Step 2 also includes the following steps: Step 21: During the power-on self-test of the star sensor, a dark image is acquired as the background image; Step 22: The star sensor acquires a 12-bit image with a maximum grayscale of 4095 and performs high-pass filtering. Step 23: For defects in the image with a gray value less than 1024, the current gray value is subtracted from the gray value of the background image and treated as a normal pixel. Step 24: For defects in the processed image with gray values ​​of 1024 to 2047, compensate the response gray value according to a linear equation; Step 25: For defects in the image with a gray value greater than or equal to 2048, use neighborhood interpolation or median substitution methods for processing.

2. The pretreatment method for nuclear radiation neutron-damaged star sensors according to claim 1, characterized in that, In step 24, the linear equation is: ； Where x represents the gray level of the defect under no-light conditions.

3. The pretreatment method for nuclear radiation neutron-damaged star sensors according to claim 1, characterized in that, In step 25, if there are no defects in the neighboring pixels, the grayscale of the defects is replaced by the neighboring interpolation method; if there are still defects in the neighboring pixels, the median is used for replacement.

4. The pretreatment method for nuclear radiation neutron-damaged star sensors according to claim 1, characterized in that, In step 3, the correlation determination method is as follows: Centered on a pixel, take a 3×3 pixel block and calculate its grayscale mean and variance: ； To obtain a larger threshold, Indicates the grayscale mean. Represents the variance of gray levels. Indicates the threshold; If the value exceeds the threshold, the pixel is judged to need further correction.

Images