A small celestial body centroid extraction method based on a joint pixel distribution frequency template

Abstract

The application discloses a small celestial body centroid extraction method based on a joint pixel distribution frequency template and belongs to the technical field of aerospace. The application realizes the method as follows: converting an RGB noise optical image of a small celestial body into a gray-scale image, segmenting the gray-scale image by taking the maximum gray-scale value of all pixels in the first k columns of the gray-scale image as a segmentation threshold to obtain a segmented image; marking and combining filled connected domains to determine a small celestial body connected image and a small celestial body connected domain; logically judging the gray-scale image by using the connected image to obtain a small celestial body connected gray-scale image reflecting brightness information; constructing a row-column direction pixel distribution frequency template; multiplying and smoothing the row-column distribution frequencies of corresponding pixel points to obtain a joint pixel distribution frequency template reflecting shape information; establishing a new representation fusing the brightness and shape information of the small celestial body; calculating the centroid coordinates of the small celestial body by weighted average to realize the centroid extraction of the small celestial body. The application has the advantages of high solution precision, strong solution stability and strong expansion capacity.

Description

TECHNICAL FIELD

[0001] The application relates to a small celestial body centroid extraction method based on a joint pixel distribution frequency template and belongs to the technical field of aerospace. BACKGROUND

[0002] With the continuous development of aerospace technology, deep space exploration missions for small celestial bodies and other targets have gradually become the focus of attention at home and abroad. Spacecraft navigation, guidance and control are the basis for the smooth implementation of related missions. At present, small celestial body relative orbit determination based on optical measurement is widely used in small celestial body exploration due to its strong autonomy and high precision. Target centroid extraction is a prerequisite and basis for implementing optical relative orbit determination, and the optical center or centroid is generally used to replace the centroid to provide the necessary relative line-of-sight information. Deep space small celestial body targets have different characteristics from general near-earth targets or spacecraft targets, mainly being dark and irregular, and being disturbed by much image noise. Therefore, it is very important to study the centroid extraction method for small celestial bodies.

[0003] Among the developed methods for replacing the centroid of the target celestial body with the optical center or the centroid, the prior art [1] (see: Jia H, Liu Q Y, Xie Y F, et al. High Precision Centroid Extraction Algorithm of Space Target for Autonomous Optical Navigation [C] / / 2022 41st Chinese Control Conference (CCC). IEEE, 2022: 6222-6227.) proposes an improved gray-scale centroid extraction algorithm to realize sub-pixel centroid extraction, adopts a regional centroid extraction method to realize preliminary centroid extraction, and then performs sub-pixel correction according to the gray-scale value and the image gradient to obtain a more accurate centroid position.

[0004] The prior art [2] (see: Cui P Y, Jia H, Zhu S Y. Small Celestial Body Irregularity and Optical Navigation Centroid Extraction and Application [J]. Journal of Spacecraft Technology, 2021, 42(01): 83-91.) proposes a centroid extraction method based on the minimum circumscribed figure for the case that the small celestial body has large irregularity. On the basis of edge detection of the projection image, the minimum circumscribed figure of the convex hull vertex is calculated to realize accurate extraction of the centroid of the irregular small celestial body. The method is less affected by the spin of the small celestial body, and can improve the extraction accuracy of the centroid of the small celestial body, and further improve the accuracy of autonomous optical navigation.

[0005] The two replacement forms of optical center and centroid have limitations in distance and characteristics of applicable scenarios, and the single evaluation of the brightness or shape of the target in the image is not comprehensive enough as the target centroid. Through a specially designed image preprocessing method, and by fusing the brightness and shape information of small celestial bodies in the noise optical image, the small celestial body centroid extraction precision can be effectively improved. SUMMARY

[0006] The application aims to provide a small celestial body centroid extraction method based on a joint pixel distribution frequency template. For the small celestial body centroid extraction problem, a segmentation threshold is designed to segment the small celestial body gray image, mark and combine the connected domain to determine the small celestial body connected image and the small celestial body connected domain. The connected image is used for logical judgment on the gray image to obtain a connected gray image reflecting the brightness information. Further, a row-column direction pixel distribution frequency template is constructed based on the connected domain. The joint pixel distribution frequency template reflecting the shape information is obtained by multiplying and smoothing the corresponding pixel row-column distribution frequency. A new representation fusing the small celestial body brightness and shape information is established to calculate the centroid coordinates of the small celestial body, and high-precision centroid extraction of the small celestial body is realized. The application has the advantages of high solution precision, strong solution stability and strong expansion capability.

[0007] The application is realized by the following technical scheme:

[0008] The small celestial body centroid extraction method based on the joint pixel distribution frequency template disclosed in the application is an improved gravity center method fusing the brightness and shape information. To solve the high-precision centroid extraction problem of the small celestial body target, the RGB noise optical image of the small celestial body is converted into a gray image. The maximum gray value of all pixels in the first k columns of the gray image is taken as a segmentation threshold T to segment the gray image. The target region greater than the segmentation threshold and the background region less than or equal to the segmentation threshold are distinguished, and a segmented image is obtained. For the segmented image, an 8-connected criterion is selected to mark the connected domain belonging to the same object in the segmented image. According to the number of pixels in the connected domain, a judgment threshold n is set. When the number of pixels in the connected domain is greater than the judgment threshold n, the connected domain is marked as a target region; otherwise, the connected domain is marked as a background region. The target region and the background region are marked in the segmented image. The target region is filled to obtain a connected image of the small celestial body. The connected image is used for logical judgment on the gray image to obtain a connected gray image reflecting the brightness information. Further, a row-column direction pixel distribution frequency template is constructed based on the connected domain. The joint pixel distribution frequency template reflecting the shape information is obtained by multiplying and smoothing the corresponding pixel row-column distribution frequency. A new representation fusing the small celestial body brightness and shape information is established to calculate the centroid coordinates of the small celestial body, and high-precision centroid extraction of the small celestial body is realized. The application has the advantages of high solution precision, strong solution stability and strong expansion capability. TFurther distinguish the background point region and one or more small celestial body illumination regions from the target region; combine all the small celestial body illumination regions to obtain a binary representation of a small celestial body combined connected image, further fill local 0 value holes in the small celestial body combined connected image, and determine a final small celestial body connected image and a small celestial body connected domain; use the small celestial body connected image to perform logical judgment on the gray scale image to obtain a small celestial body connected gray scale image reflecting small celestial body brightness information, which is used for subsequent small celestial body centroid extraction; on the basis of the small celestial body connected domain, construct a small celestial body row and column direction pixel distribution frequency template; comprehensively consider the row and column directions, multiply the corresponding pixel row and column distribution frequencies and smooth the pixel row and column distribution frequencies to obtain a joint pixel distribution frequency template reflecting small celestial body shape information; based on the gray scale value of the small celestial body connected gray scale image, establish a new representation fusing small celestial body brightness and shape information, and obtain the centroid coordinates (u c ,v c ) of the small celestial body through weighted average calculation. The application can be used for small celestial body detection, proximity navigation guidance control and other task scenarios.

[0009] The application discloses a small celestial body centroid extraction method based on a joint pixel distribution frequency template, and the method comprises the following steps:

[0010] Step one: convert the RGB noise optical image of the small celestial body into a gray scale image, use the maximum gray scale value of all the pixels in the first k columns of the gray scale image as a segmentation threshold T to segment the gray scale image, distinguish the target region greater than the segmentation threshold from the background region less than or equal to the segmentation threshold, and obtain a segmentation image, thereby realizing small celestial body threshold segmentation.

[0011] The maximum gray scale value of all the pixels in the first k columns of the small celestial body gray scale image is used as the segmentation threshold T:

[0012]

[0013] Wherein g(u,v) is the gray scale value of the gray scale image, and H is the image height.

[0014] The gray scale value s(u,v) of the segmentation image obtained by segmenting the gray scale image through the segmentation threshold T is:

[0015]

[0016] Wherein the region with a non-zero gray scale value is the target region, and the region with a zero gray scale value is the background region.

[0017] Step two: for the segmentation image obtained in step one, mark the connected domain belonging to the same object in the segmentation image by selecting an 8-connected criterion, and based on the pixel quantity in the connected domain, determine a judgment threshold n TFurther distinguish the background point region and one or more small celestial body illumination regions from the target region of step one; combine all the small celestial body illumination regions obtained in step 2.1 to obtain a small celestial body combined connected image represented in binary, further fill the local 0-value holes therein, determine the final small celestial body connected image and small celestial body connected domain C; use the small celestial body connected image to logically judge the gray-scale image of step one, obtain the gray-scale value I c (u,v) small celestial body connected gray-scale image of (u,v), for subsequent small celestial body centroid extraction, realize small celestial body connected domain labeling.

[0018] Step 2.1, for the segmented image obtained in step one, select 8-connected criterion to label the connected domain belonging to the same object in the segmented image, according to the number of pixels in the connected domain, set the judgment threshold n T Further distinguish the background point region and one or more small celestial body illumination regions from the target region of step one;

[0019] Obtain the connected image with label value l(u,v) through 8-connected criterion, the region with the same label value is the connected domain belonging to the same object; the connected domain with the largest number of pixels belongs to the background region obtained in step one, and the other connected domains belong to the target region; and the target region contains the background point region which may only occupy a few pixels, and one or more small celestial body illumination regions with moderate number of pixels;

[0020] Therefore, for different connected domains belonging to the target region, the number of pixels in each connected domain is n, and the judgment threshold n T Distinguish the background point region less than or equal to the judgment threshold and the small celestial body illumination region A greater than the judgment threshold from the target region.

[0021] Step 2.2, combine all the small celestial body illumination regions obtained in step 2.1 to obtain a small celestial body combined connected image represented in binary, further fill the local 0-value holes therein, determine the final small celestial body connected image and small celestial body connected domain C.

[0022] Obtain the small celestial body combined connected image represented in binary b(u,v) by combining all the small celestial body illumination regions:

[0023]

[0024] Further fill the zero to a few possible 0-value holes therein, i.e. reassign all 0 values in all holes to 1, determine the final small celestial body connected image represented in binary c(u,v), and the small celestial body connected domain C corresponding to the 1 value.

[0025] Step 2.3, using the small celestial body connected image of step 2.2 to logically judge the gray scale image of step one, to obtain a gray scale value I c reflecting the small celestial body brightness information.

[0026] Step three, on the basis of the small celestial body connected domain obtained in step two, construct a small celestial body row and column direction pixel distribution frequency template; by multiplying and smoothing the corresponding pixel point row and column distribution frequency, obtain a joint pixel distribution frequency template reflecting the small celestial body shape information.

[0027] Step 3.1, on the basis of the small celestial body connected domain obtained in step 2.2, construct a small celestial body row and column direction pixel distribution frequency template, respectively:

[0028]

[0029] Wherein, N is the number of all effective pixel points in the small celestial body connected domain, U and V are respectively the corresponding u row and v column of all pixel sets;

[0030] Step 3.2, by multiplying and smoothing the corresponding pixel point row and column distribution frequency, obtain a joint pixel distribution frequency template reflecting the small celestial body shape information:

[0031]

[0032] Wherein, the Hadamard product symbol ⊙ represents element multiplication;

[0033] Step four, based on the gray scale value of the small celestial body connected gray scale image obtained in step two, a new representation I c (u,v)·Φ(u,v) is established, which is calculated by weighted average to obtain the centroid coordinates (u c ,v c ) of the small celestial body, realizing the fusion representation centroid extraction.

[0034] According to formula (6), the centroid coordinates (u c ,v c ) of the small celestial body are calculated by weighted average.

[0035]

[0036] Also includes step five, using the small celestial body threshold segmentation of step one, the small celestial body connected domain marking of step two, the joint pixel distribution frequency template of step three, the fusion representation centroid extraction method of step four, realizing high precision small celestial body centroid extraction, providing basic information for subsequent small celestial body detection, approach navigation guidance control.

[0037] Advantages:

[0038] 1. The small celestial body centroid extraction method based on the joint pixel distribution frequency template disclosed in the application is characterized in that the segmentation threshold and the small celestial body noise optical image preprocessing mode such as connected domain combination and filling are designed in a targeted manner, and the row direction, the column direction and the joint pixel distribution frequency template are constructed, so that the small celestial body brightness and shape information are fused, the centroid extraction is realized through the weighted average of the fused representation, and the extraction precision of the small celestial body centroid is improved.

[0039] 2. The small celestial body centroid extraction method based on the joint pixel distribution frequency template disclosed in the application is characterized in that the determination of the small celestial body connected domain is more accurate through image preprocessing, the comprehensive information of the small celestial body brightness and shape is contained in the fused representation, and the stability of the solution is stronger.

[0040] 3. The small celestial body centroid extraction method based on the joint pixel distribution frequency template disclosed in the application is characterized in that the imaging characteristics of the deep space small celestial body target can also be extended to the case of observing other large celestial bodies at a long distance, is suitable for centroid extraction task scenes of various different targets, and has stronger expansion capability. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flowchart of the small celestial body centroid extraction method based on the joint pixel distribution frequency template of the application;

[0042] Figure 2 is a three-dimensional model of the small celestial body 433 Eros selected in the embodiment;

[0043] Figure 3 is a local enlarged image of the small celestial body noise optical image in the embodiment;

[0044] Fig. 4 is a local enlarged binary image of the segmented image and the connected image in the embodiment, Figure 4a ) is a local enlarged binary image of the segmented image in the embodiment, Figure 4b ) is a local enlarged binary image of the connected image in the embodiment;

[0045] Figure 5 is a local enlarged image of the small celestial body centroid extraction result in the embodiment, wherein the white color is the real centroid of the small celestial body, and the red color is the small celestial body centroid extracted by the method of the application;

[0046] Figure 6 is the small celestial body centroid extraction error under the condition of occupying different pixel numbers in the embodiment. DETAILED DESCRIPTION

[0047] To better illustrate the purpose and advantages of this invention, a simulation analysis of a method for extracting the centroid of small celestial bodies based on a joint pixel distribution frequency template is presented below to provide a detailed explanation of the invention.

[0048] This embodiment uses, as follows Figure 2 Taking the irregularly shaped small celestial body 433Eros as an example, we extract the coordinates of its centroid.

[0049] like Figure 1 As shown in the figure, this embodiment discloses a method for extracting the centroid of small celestial bodies based on a joint pixel distribution frequency template. The specific implementation steps are as follows:

[0050] Step 1: To eliminate the influence of noise in the low signal-to-noise ratio of the optical image with noise from small celestial bodies, a threshold segmentation method is used. The RGB noise optical image of the small celestial body is converted into a grayscale image. The maximum grayscale value of all pixels in the first k columns of the grayscale image is used as the segmentation threshold T to segment the grayscale image, distinguishing the target area with a value greater than the segmentation threshold from the background area with a value less than or equal to the segmentation threshold, and thus obtaining the segmented image.

[0051] Considering that small celestial bodies are usually located near the center of the image during tracking observations, and that the non-zero gray values ​​of the deep-space background are mainly Gaussian noise, the maximum gray value of all pixels in the first k columns of the small celestial body grayscale image is used as the segmentation threshold T:

[0052]

[0053] Where g(u,v) is the gray value of the grayscale image, and H is the image height; this segmentation threshold can effectively capture the maximum noise in the background, thereby segmenting away the vast majority of the background; as a preferred option, k = W / 5, that is, the first 1 / 5 of the image width W;

[0054] The grayscale image is segmented using a segmentation threshold T to obtain a segmented image. The grayscale values ​​s(u,v) of the segmented image are:

[0055]

[0056] The areas with non-zero grayscale values ​​are the target areas, and the areas with zero grayscale values ​​are the background areas.

[0057] Step 2: For the segmented image obtained in Step 1, select the 8-connectivity criterion to mark the connected components belonging to the same object in the segmented image. Based on the number of pixels in the connected component, and using the judgment threshold n... TFurther distinguish the background point region and one or more small celestial body illumination regions from the target region of step one; combine all the small celestial body illumination regions obtained in step 2.1 to obtain a small celestial body combined connected image represented by binary, further fill the local 0-value holes in the small celestial body combined connected image, determine the final small celestial body connected image and small celestial body connected domain C; use the small celestial body connected image to logically judge the gray-scale image of step one, obtain the gray-scale value I c (u,v) of the small celestial body connected gray-scale image, for subsequent small celestial body centroid extraction;

[0058] Step 2.1, for the segmented image obtained in step one, select an 8-connected criterion to mark the connected domain belonging to the same object in the segmented image, according to the number of pixels in the connected domain, set a judgment threshold n T Further distinguish the background point region and one or more small celestial body illumination regions from the target region of step one;

[0059] By considering the 8-connected criterion of edge contact and corner contact of the pixels at the same time, obtain the connected image with the marking value l(u,v), the region with the same marking value is the connected domain belonging to the same object; the connected domain with the largest number of pixels belongs to the background region obtained in step one, and the other connected domains belong to the target region; and the target region contains the background point region which may only occupy a few pixels, and one or more small celestial body illumination regions with moderate number of pixels;

[0060] Therefore, for different connected domains belonging to the target region, assuming that the number of pixels in each connected domain is n, set a judgment threshold n T , distinguish the background point region with the number of pixels less than or equal to the judgment threshold and the small celestial body illumination region A with the number of pixels greater than the judgment threshold from the target region; as preferred, take n T = 2;

[0061] Step 2.2, combine all the small celestial body illumination regions obtained in step 2.1 to obtain a small celestial body combined connected image represented by binary, further fill the local 0-value holes in the small celestial body combined connected image, determine the final small celestial body connected image and small celestial body connected domain C;

[0062] Considering the complex space lighting conditions and irregular small celestial body shapes, the large-area shadow on the surface of the small celestial body may separate multiple small celestial body illumination regions, therefore, combine all the small celestial body illumination regions to obtain a small celestial body combined connected image represented by binary b(u,v):

[0063]

[0064] In addition, the rugged terrain of small celestial body craters and other rough terrain may also make the small celestial body light area appear wrapped local shadow in the middle, and the current small celestial body genesis hypothesis does not support the emergence of holes, so further filling of zero to a plurality of possible 0-value holes, that is, reassigning all 0 values in all holes to 1, determines the final small celestial body connected image represented by c(u, v) binary, and the small celestial body connected domain C corresponding to the 1 value;

[0065] Step 2.3, using the small celestial body connected image of step 2.2 to logically judge the gray scale image of step one, to obtain the gray scale value I that can reflect the small celestial body brightness information c (u, v) small celestial body connected gray scale image, for subsequent small celestial body centroid extraction;

[0066] Step three, on the basis of the small celestial body connected domain obtained in step 2.2, construct the pixel distribution frequency template in the row and column directions of the small celestial body, wherein the pixel distribution frequency template in the row direction has the same value in the same row, reflecting the pixel proportion of the target shape in each row, and the column direction is the same;

[0067] Step 3.1, on the basis of the small celestial body connected domain obtained in step 2.2, construct the pixel distribution frequency template in the row and column directions of the small celestial body, respectively:

[0068]

[0069] Wherein, N is the number of all effective pixel points in the small celestial body connected domain, U and V are respectively the corresponding u-th row and v-th column of all pixel sets; the pixel distribution frequency template in the row direction has the same value in the same row, reflecting the pixel proportion of the target shape in each row, and the column direction is the same;

[0070] Step 3.2, by multiplying the row and column distribution frequencies of the corresponding pixel points and smoothing, a joint pixel distribution frequency template reflecting the shape information of the small celestial body is obtained:

[0071]

[0072] Wherein, the Hadamard product symbol ⊙ represents element multiplication;

[0073] Step four, based on the gray scale value of the small celestial body connected gray scale image obtained in step 2.3, a new representation I c (u, v)·Φ(u, v) is established, which calculates the centroid coordinates (u c ,v c ) of the small celestial body by weighted average.

[0074]

[0075] Step five, using the small celestial body threshold segmentation of step one, the small celestial body connected domain labeling of step two, the joint pixel distribution frequency template of step three, the fusion representation centroid extraction method of step four, realize high-precision small celestial body centroid extraction, provide basic information for subsequent small celestial body detection, approach navigation guidance control.

[0076] In order to verify the feasibility and effectiveness of the method, the simulation results are analyzed in this embodiment.

[0077] The local enlarged image of the small celestial body in the noise optical image of the small celestial body 433 Eros is shown in Fig. 3, and the small celestial body occupies about 170 pixels. The local enlarged binary image of the small celestial body segmentation image and the connected image is shown in Fig. 4, for the convenience of clear display, Figure 3 ) the segmented image with gray scale is binarized, wherein the white color is the target area and the black color is the background area. It can be seen that the target area contains a small celestial body illumination area with a large number of pixels, and a background point area occupying only a few pixels, and the two are not connected; Figure 4a ) the white color in the connected image is the small celestial body connected domain. It can be seen that the imaging part corresponding to the small celestial body is successfully determined. The extraction result of the small celestial body centroid coordinates is shown in Fig. 5, wherein the white cross is the real centroid of the small celestial body, and the red cross is the extracted small celestial body centroid. The two-dimensional centroid extraction error is only about 0.0292 pixel, realizing high-precision sub-pixel level centroid extraction of the small celestial body. Figure 4b Figure 5 ) the white color in the connected image is the small celestial body connected domain. It can be seen that the imaging part corresponding to the small celestial body is successfully determined. The extraction result of the small celestial body centroid coordinates is shown in Fig. 5, wherein the white cross is the real centroid of the small celestial body, and the red cross is the extracted small celestial body centroid. The two-dimensional centroid extraction error is only about 0.0292 pixel, realizing high-precision sub-pixel level centroid extraction of the small celestial body. Figure 6

[0078] The above specific description further describes the purpose, technical scheme and beneficial effects of the application. It should be understood that the above description is only a specific embodiment of the application, which is used to explain the application and does not limit the protection scope of the application. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of the application shall be included in the protection scope of the application.​​

Claims

1. A method for small body centroid extraction based on joint pixel distribution frequency template, characterized in that: The method comprises the following steps, Step one, the RGB noise optical image of the small celestial body is converted into a gray image, the maximum gray value of all pixels in the first k columns of the gray image is taken as a segmentation threshold T, the gray image is segmented, the target region greater than the segmentation threshold and the background region less than or equal to the segmentation threshold are distinguished, and a segmented image is obtained, and threshold segmentation of the small celestial body is realized; Step two, for the segmented image obtained in step one, select 8-connected criterion to mark the connected domain belonging to the same object in the segmented image, and according to the number of pixels in the connected domain, determine the threshold value n based on the judgment T Further distinguish the background point region and one or more small celestial body illumination regions from the target region in step one; combine all the obtained small celestial body illumination regions to obtain a binary representation of the small celestial body combined connected image, further fill the local 0 value holes therein, and determine the final small celestial body connected image and small celestial body connected domain C; use the small celestial body connected image to perform logical judgment on the gray scale image in step one to obtain a gray scale value I c (u,v) small celestial body connected gray scale image, which is used for subsequent small celestial body centroid extraction and small celestial body connected domain marking; Step three, on the basis of the small celestial body connected domain obtained in step two, a pixel distribution frequency template in the row and column directions of the small celestial body is constructed; by multiplying and smoothing the row and column distribution frequencies of the corresponding pixel points, a joint pixel distribution frequency template reflecting the shape information of the small celestial body is obtained; Step four, based on the gray value of the connected gray image of the small celestial body obtained in step two, a new representation I that fuses the brightness and shape information of the small celestial body is established c (u,v)·Φ(u,v), the centroid coordinates (u c ,v c ) of the small celestial body are calculated by weighted average, and the fusion representation centroid extraction is realized.

2. The method for small body centroid extraction based on joint pixel distribution frequency template according to claim 1, wherein: In step one, The maximum gray value of all pixels in the first k columns of the small celestial body gray image is taken as a segmentation threshold T: Wherein, g(u,v) is the gray value of the gray image, and H is the image height; The gray image is segmented by the segmentation threshold T to obtain a segmented image, and the gray value s(u,v) of the segmented image is: Wherein, the region with a non-zero gray value is a target region, and the region with a zero gray value is a background region.

3. The method for small body centroid extraction based on joint pixel distribution frequency template according to claim 2, characterized in that: The implementation method of step two is, Step 2.1, for the segmented image obtained in step one, select 8-connected criterion to mark the connected domain belonging to the same object in the segmented image, according to the number of pixels in the connected domain, based on the judgment threshold n T Further distinguish the background point region and one or more small celestial body illumination regions from the target region in step one; The connected image with a label value l(u,v) is obtained by the 8-connected criterion, and the regions with the same label value are connected domains belonging to the same object; It is known that the connected domain with the largest number of pixels belongs to the background region obtained in step one, and the other connected domains belong to the target region; and the target region includes a background point region occupying only a few pixels and one or more small celestial body illumination regions with a moderate number of pixels; Therefore, for different connected domains belonging to the target region, the number of pixels in each connected domain is n, and a judgment threshold n based on the number of pixels is set T A background point region less than or equal to the judgment threshold and a small celestial body illumination region A greater than the judgment threshold are distinguished from the target region. Step 2.2, all small celestial body illumination regions obtained in step 2.1 are combined to obtain a small celestial body combined connected image represented by a binary value, and local 0-value holes in the small celestial body combined connected image are further filled to determine a final small celestial body connected image and a small celestial body connected domain C; By combining all small celestial body illumination regions, a small celestial body combined connected image represented by a binary value b(u,v) is obtained: Further, the zero to a plurality of possible 0-value holes are filled, that is, all 0 values in the holes are revalued as 1, and a final small celestial body connected image represented by a binary value c(u,v) and a small celestial body connected domain C corresponding to the 1 value are determined; Step 2.

3. Logical judgment is made on the gray scale image of step 1 using the small celestial body connected image of step 2.2, and a gray scale value I reflecting the brightness information of the small celestial body is obtained c The small celestial body connected gray scale image of (u, v) is used for subsequent small celestial body centroid extraction.

4. The method for small body centroid extraction based on joint pixel distribution frequency template according to claim 3, characterized in that: The implementation method of step three is, Step 3.1, on the basis of the small celestial body connected domain obtained in step 2.2, a pixel distribution frequency template in the row and column directions of the small celestial body is constructed, respectively: Wherein, N is the number of all effective pixel points in the small celestial body connected domain, U and V are all pixel sets corresponding to the u-th row and the v-th column, respectively; Step 3.2, by multiplying and smoothing the row and column distribution frequencies of the corresponding pixel points, a joint pixel distribution frequency template reflecting the shape information of the small celestial body is obtained: Wherein, the Hadamard product symbol ⊙ represents element multiplication.

5. The method for small body centroid extraction based on joint pixel distribution frequency template according to claim 4, characterized in that: According to formula (6), the barycentric coordinates (u c ,v c ) of the small body are calculated by weighted average

Images