Star pattern recognition method based on radial features and inter-star angular distances

Abstract

The star map recognition method based on radial features and interstellar angular distance relates to the field of star map recognition in astronomical navigation, solves the problems that the existing star map recognition algorithm has less number of stars to be recognized at one time, and the recognition speed and robustness are difficult to be compatible, and the method comprises the following steps: a navigation star table is made, radial recognition feature library and interstellar position relationship library are constructed according to the position distribution features of each navigation star and surrounding near neighbor stars; the interstellar position relationship library is used for storing the star number and radial ring number of the surrounding near neighbor stars of each navigation star; then, initial recognition and accurate recognition are carried out. In the process of matching by using the radial feature library, an angular distance error tolerance is set, and a weight value is set for each ring number, so that the star position noise has strong robustness. After the initial recognition, the accurate recognition is carried out, the surrounding near neighbor stars of the main star are used, and the method has strong robustness to false star noise and missing stars. The method realizes the recognition of multiple stars at one time, and improves the initial attitude solution accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of star map recognition in celestial navigation, and particularly relates to a star map recognition method based on radial features and inter-star angular distances. BACKGROUND

[0002] In recent years, with the diversification and complication of space missions, a series of new spacecraft such as maneuverable satellites and space weapons have begun to have increasingly strong maneuverability, and the dynamic performance and precision requirements of the attitude measurement system have also been correspondingly improved. Star sensors are high-precision attitude measurement instruments and play an irreplaceable important role in various advanced space missions. Star map recognition is used to obtain initial attitude information of a star sensor. How to quickly and accurately perform star map recognition has always been a key factor in improving the performance of a star sensor.

[0003] Existing star map recognition methods can be mainly divided into subgraph isomorphism methods and pattern recognition methods. The subgraph isomorphism method such as the golden pyramid algorithm needs to construct a quadrilateral using four star points, match the angular distance between each two stars, and store a large amount of data for slow retrieval, but it is not easy to mis-match, and the recognition accuracy is relatively high. The pattern recognition method is to construct a pattern by the relative position relationship of the star points around a main star for recognition, and the storage data capacity is small and the retrieval is fast, but it is easy to be disturbed by external noise, such as the need to find accurate near neighbors in the grid algorithm, and the radial ring algorithm is prone to mis-match when the star position deviation is large. Therefore, the technical personnel in the field are committed to developing a fast star map recognition algorithm capable of recognizing multiple star points at a time, taking into account the accuracy of the subgraph isomorphism algorithm and the speed of the pattern recognition method, to provide a faster and more accurate initial attitude for a star sensor. SUMMARY

[0004] To solve the problems such as less star recognition at a time, and difficult to balance recognition speed and robustness of existing star map recognition algorithms, the present application provides a star map recognition method based on radial features and inter-star angular distances.

[0005] The star map recognition method based on radial features and inter-star angular distances is realized by the following steps:

[0006] Step one, select several stars as navigation stars in the original star catalog, make a navigation star catalog, and record the star number, right ascension, declination and spatial direction vector information of each navigation star;

[0007] Step two, according to the position distribution characteristics of each navigation star and its surrounding near neighbors, construct a radial recognition feature library and an inter-star position relationship library; the inter-star position relationship library is used to store the star number and radial ring number of the near neighbors around each navigation star;

[0008] Step three, select the star near the center of the field of view as the main star, and construct the radial identification mode according to the position of the companion star within the mode radius of the main star;

[0009] Step four, initial identification; the radial identification mode is used to search and match the radial identification feature library, the matching weight of each navigation star is calculated and sorted by size, and the initial identification is completed;

[0010] Step five, accurate identification;

[0011] The multi-star position relationship library of the navigation star is searched according to the weight size in turn, until more than three near stars can be matched and identified, and the accurate identification is completed.

[0012] The beneficial effects of the present application are:

[0013] 1) In the construction of the radial feature library, the present application adopts the method of equal-area ring division, which reduces the database storage capacity and improves the matching efficiency.

[0014] 2) In the process of matching the radial feature library, the present application sets an angular distance error tolerance and a weight value for each ring number, which has strong robustness to star position noise.

[0015] 3) After the initial identification, the present application performs accurate identification, which uses the near neighbor stars around the main star, and has strong robustness to false star noise and missing stars.

[0016] 4) The present application realizes the identification of multiple stars at a time, and improves the initial attitude solution accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Figure 1 The flowchart of the star identification method of the present application is shown.

[0019] Figure 2 The radial feature library of the navigation star is shown.

[0020] Figure 3 The multi-star position relationship library of the navigation star is shown.

[0021] Figure 4 The identification rate curve comparison diagram of different identification methods under different star position noise is shown.

[0022] Figure 5 This is a comparison of the recognition rate curves of different recognition methods under different false stars and missing star interference. Detailed Implementation

[0023] Combination Figures 1 to 5 This embodiment describes a star map identification method based on radial features and inter-star angular distances. The method is implemented through the following steps:

[0024] Step 1: Select several stars from the original star catalog as navigation stars, create a navigation star catalog, and record information such as the star number, right ascension, declination, and spatial direction vector of each navigation star;

[0025] In this embodiment, the principle for selecting several navigation stars is: after removing binary stars, select stars with a magnitude of 5.8 or higher as navigation stars.

[0026] Step 2: Based on the positional distribution characteristics of each navigation star and its neighboring stars, construct a radial feature library and a multi-star positional relationship library to store the star number and radial ring number of the neighboring stars around each navigation star.

[0027] In this embodiment, the specific steps for establishing the radial feature library and the inter-satellite position relationship library of the navigation satellite are as follows:

[0028] Step A1: Nearest neighbor stars refer to stars of magnitude 5.8 or higher within the radius of the navigation star model. Divide the area into equal-area distribution rings based on the angular distance between the navigation star and the nearest neighbor stars, ensuring that the probability of a nearest neighbor star falling into each ring is approximately equal.

[0029] Step A2: Calculate the inter-satellite angular distance between the navigation star and all its nearest neighbor stars, determine the ring number of the nearest neighbor stars, and store the navigation star's star number under the navigation star number stored in that ring number, thereby establishing a radial feature library of the navigation star.

[0030] like Figure 2 As shown, the radial feature library is searched by ring number, and each ring number stores the main star numbers of the neighboring stars of the corresponding ring number.

[0031] Step A3: Construct a multi-star position relationship database for each navigation star. Select an appropriate number of neighboring stars based on their magnitude and inter-star distribution, and store the neighboring star numbers and ring numbers of each navigation star as the multi-star position relationship database.

[0032] like Figure 3 As shown, the multi-star position relationship database consists of N (number of primary stars) matrices, each storing the nearest neighbor stars of the primary star and the ring number of the nearest neighbor stars.

[0033] Step 3: Select the star closest to the center of the field of view in the star chart as the primary star, and construct a radial recognition pattern based on the positions of companion stars within the pattern radius of the primary star;

[0034] Step 4: Initial identification. Use the radial identification mode to search and match the radial feature library, calculate the weight of each navigation star match and sort them by size to complete the initial identification.

[0035] In this embodiment, the principle of weight calculation is as follows: calculate the angular distance θ between the nearest star and the main star, set the angular distance tolerance error ±ε (the error is calculated based on the actual star sensitivity parameters, approximately 0.000147 (0.05°)), calculate the ring number range of the nearest star, take the navigation star number within the ring number and accumulate the weights of different ring numbers.

[0036] Step 5: Precise identification. Search the inter-satellite position relationship database of navigation stars sequentially according to their weights until three or more neighboring stars can be matched and identified, completing the precise identification. If the number of searches exceeds the search threshold, select other stars close to the center of the field of view as the primary stars and proceed to Step 6.

[0037] In this embodiment, the method for matching and identifying neighboring stars around the primary star is as follows:

[0038] Step B1: First, match the star number of the star point in the star map with the ring number of the nearest star in the multi-star position relationship database.

[0039] Step B2: Select two neighboring stars A and B. Based on the star designations of the neighboring stars and the primary star designation C, retrieve the direction vector from the navigation star table and calculate the angular distances between stars A and B, B and C, and A and C respectively. Then compare these angular distances with the actual star points calculated. If the error is within the range ω (the error is calculated based on the actual star sensitivity parameters and cannot be given a specific value, but is approximately 0.000147 (0.05°)), then the three star points are considered to have been successfully identified.

[0040] Step B3: Repeat step B2 until all neighboring stars matching the ring number have been identified. If the total number of identified stars is greater than 3, the matching is considered successful and accurate identification is completed.

[0041] Step 6: If the value of the primary star being replaced is greater than the threshold for replacing the primary star, the identification fails. Otherwise, a radial identification pattern is constructed based on the location of the primary star and the positions of the nearest neighboring stars within it, and step 4 is repeated.

[0042] like Figure 4 and Figure 5 As shown, Figure 4 This is a comparison of the recognition rate curves of different recognition methods under varying star point noise levels. The horizontal axis represents the standard deviation of star point noise, and the vertical axis represents the success rate of star map recognition. Star map parameters: resolution 1280*1024, pixel size 5.3um, focal length 25mm. Figure 5 This is a comparison of the recognition rate curves of different recognition methods under varying interference from false stars and missing stars. The horizontal axis represents the number of false stars and missing stars, and the vertical axis represents the success rate of star map recognition.

[0043] Each technical feature of the above-described embodiments can be combined with any other technical feature, and for the sake of brevity, not all possible combinations are described, but it is understood that the scope of the present disclosure encompasses all such possible combinations.

[0044] The above-described embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be noted that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of the present application. Therefore, the scope of the patent of the present application should be subject to the appended claims.

Claims

1. A star map recognition method based on radial features and inter-star angular distances, characterized by: This method is implemented by the following steps: Step 1: Select several stars from the original star catalog as navigation stars, create a navigation star catalog, and record the star number, right ascension, declination, and spatial direction vector information of each navigation star; Step 2: Based on the positional distribution characteristics of each navigation star and its neighboring stars, construct a radial identification feature library and a multi-star positional relationship library; the multi-star positional relationship library is used to store the star number and radial ring number of the neighboring stars around each navigation star; In step two, the specific steps for establishing the radial identification feature library and the inter-satellite position relationship library for navigation stars are as follows: Step A1: Divide the navigation star and the nearest neighboring star into equally distributed rings to ensure that the nearest neighboring star has an equal probability of falling into each ring; Step A2: Calculate the inter-satellite angular distance between the navigation star and all its nearest neighbor stars, determine the ring number of the nearest neighbor stars, store the star number of the navigation star under the navigation star number stored in the ring number, and establish a radial feature library of the navigation star. Step A3: Build a multi-satellite position relationship database for each navigation satellite; Select an appropriate number of neighboring stars based on their magnitude and inter-star distribution, and store the star numbers of the neighboring stars and the ring numbers of the neighboring stars as a multi-star position relationship database for each navigation star. Step 3: Select the star closest to the center of the field of view in the star chart as the primary star, and construct a radial recognition pattern based on the positions of companion stars within the pattern radius of the primary star; Step 4: Initial Identification; The radial recognition feature library is retrieved and matched using the radial recognition pattern. The weight of each navigation star is calculated and sorted by size to complete the initial recognition. In step four, the principle of weight calculation is as follows: calculate the angular distance θ between the nearest star and the main star, set the angular distance tolerance error ±ε, calculate the ring number range of the nearest star, take the star number of the navigation star within the ring number and accumulate the weights of different ring numbers. Step 5: Accurate Identification; Search the multi-star position relationship database of navigation stars sequentially according to their weight values ​​until three or more neighboring stars can be matched and identified to complete accurate identification; The specific process of accurate identification is as follows: Step B1: Match the star number of the star point in the star map with the ring number of the nearest star in the multi-star position relationship database; Step B2: Take two nearby stars A and B. Based on the star designation of the nearby stars and the star designation of the primary star C, take the spatial direction vector in the navigation star table and calculate the angular distances between nearby stars A and B, between nearby stars B and the primary star C, and between nearby stars A and the primary star C. Then compare these angular distances with the actual star points calculated. If they are within the set error range ω, then three star points have been successfully identified. Step B3: Repeat step B2 until all neighboring stars matching the ring number have been identified. If the total number of identified stars is greater than 3, the matching is successful and accurate identification is completed.

2. The star map recognition method based on radial features and inter-star angular distances according to claim 1, characterized in that: In step five, if three or more neighboring stars cannot be matched, select other stars close to the center of the field of view as the primary star and proceed to step six; Step 6: If the value of the primary star to be replaced is greater than the threshold for replacing the primary star, the recognition fails; otherwise, a radial recognition pattern is constructed based on the positions of the nearest stars within the pattern radius of the primary star, and the process returns to Step 4.

3. The star map recognition method based on radial features and inter-star angular distances according to claim 1, characterized in that: The principle for selecting several navigation stars in step one is: after removing binary stars, select stars with a magnitude of 5.8 or higher as navigation stars.

Images