Method and device for determining an attitude based on matching of features of a solar image

Abstract

The application relates to the technical field of attitude measurement, in particular to a method and device for determining an attitude based on feature matching of a sun image, wherein the method comprises the following steps: obtaining a target sun image, then extracting sunspot features meeting preset stability conditions from the target sun image, combining bidirectional consistency constraints, geometric consistency constraints, local gray consistency constraints and sub-pixel correction constraints to obtain sunspot matching feature point pairs meeting preset matching conditions, then constructing unit direction observation vector sets of two groups of target sun images, constructing a centralized direction vector set, and generating a relative attitude change amount of a spacecraft corresponding to the target sun image. Thus, the problem that related technologies use other vector information for attitude compensation, which is large in calculation amount, insufficient in real-time performance and difficult to meet the high real-time performance and high robustness requirements of an embedded space platform is solved.

Description

Technical Field

[0001] This application relates to the field of attitude measurement technology, and in particular to an attitude determination method and apparatus based on solar image feature matching. Background Technology

[0002] Accurate attitude acquisition and stable control of spacecraft are fundamental conditions for spacecraft mission execution. Sun sensors are widely used due to their simple structure and low power consumption. However, in related technologies, sun sensors typically only measure the direction of the solar vector, making it difficult to obtain the spacecraft's roll angle around the solar vector direction, and thus unable to directly acquire the spacecraft's three-axis attitude. While related technologies have attempted to utilize other vector information for attitude compensation, the computational load is high and real-time performance is insufficient, failing to meet the high real-time performance and robustness requirements of embedded aerospace platforms, and thus urgently needs to be addressed. Summary of the Invention

[0003] This application provides an attitude determination method and apparatus based on solar image feature matching to solve the problems of related technologies that use other vector information for attitude compensation, which have large computational load and insufficient real-time performance, making it difficult to meet the high real-time performance and high robustness requirements of embedded aerospace platforms.

[0004] The first aspect of this application provides an attitude determination method based on solar image feature matching, comprising the following steps: acquiring a target solar image; extracting sunspot features that satisfy preset stability conditions from the target solar image, and based on the sunspot features, combining bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, obtaining sunspot matching feature point pairs that satisfy preset matching conditions; constructing a unit direction observation vector set of the target solar image based on the sunspot matching feature point pairs, constructing a centralized direction vector set of the target solar image based on the unit direction observation vector set, and generating the spacecraft relative attitude change amount corresponding to the target solar image based on the centralized direction vector set.

[0005] Based on the above technical means, this application embodiment utilizes the spatial distribution stability of sunspots in short-term continuous observations. By extracting local constraint features for stable sunspot observations, it achieves stable extraction of sunspot region features. Combined with a multi-constraint consistency optimization strategy, it achieves highly stable feature matching between solar images. Furthermore, through direction vector construction and rotation matrix estimation, it achieves accurate calculation of the spacecraft's relative attitude change. It does not require the construction of a three-dimensional structural model of the target space, which can effectively reduce computational complexity, improve matching stability and attitude estimation robustness under complex solar observation conditions, and has good real-time performance, making it more suitable for attitude change measurement tasks in embedded aerospace platforms.

[0006] Optionally, in one embodiment of this application, extracting sunspot features that satisfy preset stability conditions from the target solar image includes: generating a solar region mask that satisfies a first preset segmentation condition based on the target solar image; generating a photosphere background image that satisfies a preset filling condition based on the solar region mask; constructing a corresponding background difference enhancement image based on the target solar image and the photosphere background image; extracting candidate sunspot regions that satisfy a second preset segmentation condition based on the background difference enhancement image, and obtaining sunspot edge regions that satisfy preset gradient response conditions; generating a sunspot region mask based on the candidate sunspot regions and the sunspot edge regions, and extracting the sunspot features based on the sunspot region mask.

[0007] Based on the above technical means, the embodiments of this application can accurately locate effective sunspot regions by segmenting the target solar image, filling and modeling the background, enhancing the background difference, extracting candidate sunspot regions and filtering sunspot edge regions, generating sunspot region masks and extracting sunspot features. This avoids the uneven illumination, gradual changes in background brightness, imaging noise and non-target texture interference that are common in solar images. It also eliminates small noise areas, weak texture pseudo-response areas and areas with insufficient gradient response, thereby improving the stability of sunspot feature extraction and providing reliable feature input for subsequent feature matching and pose calculation.

[0008] Optionally, in one embodiment of this application, the step of extracting the sunspot features based on the sunspot region mask includes: calculating the target expansion range based on the sunspot region mask; constructing a target region of interest in the sunspot region mask based on the target expansion range; and extracting the sunspot features based on the target region of interest.

[0009] Based on the above technical means, the embodiments of this application calculate the target expansion range according to the sunspot region mask, construct the target region of interest according to the target expansion range, and extract sunspot features within the target region of interest. This can dynamically select the range of feature extraction that is suitable for the size of the sunspot region, reduce the feature retrieval area, reduce the generation of redundant features, and improve the targeting, uniformity and stability of sunspot feature extraction.

[0010] Optionally, in one embodiment of this application, the step of obtaining a sunspot matching feature point pair that satisfies preset matching conditions based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, includes: determining the corresponding binary descriptor based on the sunspot features; constructing a forward matching set and a reverse matching set that satisfy preset search conditions based on the binary descriptor; constructing an initial matching point set based on the forward matching set and the reverse matching set, combined with the bidirectional consistency constraints; constructing a geometric interior point set based on the initial matching point set, combined with the geometric consistency constraints; determining the matching position that satisfies preset response conditions based on the geometric interior point set, combined with the local grayscale consistency constraints; and obtaining the sunspot matching feature point pair based on the matching position, combined with the sub-pixel correction constraints.

[0011] Based on the above technical means, this application embodiment constructs a forward matching set and a reverse matching set, and sequentially combines bidirectional consistency constraints, geometric consistency constraints, and local grayscale consistency constraints to filter matching positions. It also uses sub-pixel correction constraints to correct the accuracy of the matching positions, thereby achieving stable feature matching of the sunspot region from initial matching to high-precision positioning. This effectively reduces the false matching rate, improves the matching stability and positioning accuracy under weak textured solar background conditions, and provides reliable input for subsequent attitude change estimation.

[0012] Optionally, in one embodiment of this application, generating the spacecraft relative attitude change amount corresponding to the target solar image based on the set of centered direction vectors includes: constructing a covariance matrix corresponding to the set of centered direction vectors; performing singular value decomposition on the covariance matrix to obtain singular value decomposition results; calculating the relative attitude rotation matrix of the target solar image based on the singular value decomposition results; and generating the relative attitude change amount based on the relative attitude rotation matrix.

[0013] Based on the above technical means, the embodiments of this application construct a covariance matrix from a set of centered direction vectors and perform singular value decomposition. The inter-frame relative attitude rotation matrix is ​​solved based on the singular value decomposition results, and the relative attitude change is generated. Attitude calculation can be completed without relying on the three-dimensional geometric model of the sun, reducing the computational complexity of the algorithm and effectively improving the stability and robustness of attitude estimation under complex solar observation conditions.

[0014] A second aspect of this application provides an attitude determination device based on solar image feature matching, comprising: an acquisition module for acquiring a target solar image; an extraction module for extracting sunspot features from the target solar image that satisfy preset stability conditions, and obtaining sunspot matching feature point pairs that satisfy preset matching conditions based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints; and a determination module for constructing a unit direction observation vector set of the target solar image based on the sunspot matching feature point pairs, constructing a centralized direction vector set of the target solar image based on the unit direction observation vector set, and generating the spacecraft relative attitude change amount corresponding to the target solar image based on the centralized direction vector set.

[0015] Based on the above technical means, this application embodiment utilizes the spatial distribution stability of sunspots in short-term continuous observations. By extracting local constraint features for stable sunspot observations, it achieves stable extraction of sunspot region features. Combined with a multi-constraint consistency optimization strategy, it achieves highly stable feature matching between solar images. Furthermore, through direction vector construction and rotation matrix estimation, it achieves accurate calculation of the spacecraft's relative attitude change. It does not require the construction of a three-dimensional structural model of the target space, which can effectively reduce computational complexity, improve matching stability and attitude estimation robustness under complex solar observation conditions, and has good real-time performance, making it more suitable for attitude change measurement tasks in embedded aerospace platforms.

[0016] Optionally, in one embodiment of this application, the extraction module includes: a first generation unit, configured to generate a solar region mask satisfying a first preset segmentation condition based on the target solar image; a second generation unit, configured to generate a photosphere background image satisfying a preset filling condition based on the solar region mask; a construction unit, configured to construct a corresponding background difference enhancement image based on the target solar image and the photosphere background image; a first extraction unit, configured to extract candidate sunspot regions satisfying a second preset segmentation condition based on the background difference enhancement image, and obtain sunspot edge regions satisfying a preset gradient response condition; and a second extraction unit, configured to generate a sunspot region mask based on the candidate sunspot regions and the sunspot edge regions, so as to extract the sunspot features based on the sunspot region mask.

[0017] Based on the above technical means, the embodiments of this application can accurately locate effective sunspot regions by segmenting the target solar image, filling and modeling the background, enhancing the background difference, extracting candidate sunspot regions and filtering sunspot edge regions, generating sunspot region masks and extracting sunspot features. This avoids the uneven illumination, gradual changes in background brightness, imaging noise and non-target texture interference that are common in solar images. It also eliminates small noise areas, weak texture pseudo-response areas and areas with insufficient gradient response, thereby improving the stability of sunspot feature extraction and providing reliable feature input for subsequent feature matching and pose calculation.

[0018] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the program to implement the attitude determination method based on solar image feature matching as described in the above embodiments.

[0019] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the attitude determination method based on solar image feature matching as described above.

[0020] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, implements the above-described attitude determination method based on solar image feature matching.

[0021] This application utilizes the spatial distribution stability of sunspots in short-term continuous observations. Through local constraint feature extraction oriented towards stable sunspot observations, it achieves stable extraction of sunspot region features. Combined with a multi-constraint consistency optimization strategy, it achieves highly stable feature matching between solar images. Furthermore, through direction vector construction and rotation matrix estimation, it achieves accurate calculation of the spacecraft's relative attitude changes. This eliminates the need to construct a three-dimensional structural model of the target space, effectively reducing computational complexity and improving matching stability and attitude estimation robustness under complex solar observation conditions. It also possesses good real-time performance, making it more suitable for attitude change measurement tasks in embedded aerospace platforms. Therefore, it solves the problems of related technologies using other vector information for attitude compensation, which suffer from high computational load, insufficient real-time performance, and difficulty in meeting the high real-time and robustness requirements of embedded aerospace platforms.

[0022] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0023] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a pose determination method based on solar image feature matching provided in an embodiment of this application; Figure 2 This is a schematic diagram of local constraint feature extraction for sunspots according to an embodiment of this application; Figure 3 This is a schematic diagram of multi-constraint feature matching according to an embodiment of this application; Figure 4 This is a schematic diagram of relative attitude change estimation according to an embodiment of this application; Figure 5 This is a schematic diagram illustrating the estimation of a rotation matrix using sunspots according to an embodiment of this application; Figure 6 This is a schematic diagram of a pose determination method based on solar image feature matching according to an embodiment of this application; Figure 7 This is a block diagram of an attitude determination device based on solar image feature matching according to an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application.

[0024] Figure label: 70 - Attitude determination device based on solar image feature matching; 100 - Acquisition module, 200 - Extraction module, 300 - Determination module; 801 - Memory, 802 - Processor, 803 - Communication interface. Detailed Implementation

[0025] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0026] Spacecraft attitude measurement technology is a technique for acquiring spacecraft attitude information in space. It mainly uses attitude sensors to sense external reference information and combines this with attitude calculation algorithms to calculate the spacecraft's attitude parameters. Attitude sensors include, but are not limited to, star sensors and solar sensors.

[0027] The star sensor identifies stars by observing their distribution in a celestial coordinate system and relying on image feature extraction and star map pattern matching techniques. Mainstream identification methods include geometrically constrained star map matching and pattern-association-based feature description methods. Based on this, attitude calculation algorithms such as quaternion estimation and singular value decomposition are combined to finally calculate the spacecraft's attitude parameters. However, this type of scheme is sensitive to lighting conditions. Once subjected to strong solar light interference, it is prone to a sharp drop in the star signal-to-noise ratio, star overload, or even feature loss, directly leading to star identification failure and severely affecting the stability and reliability of attitude measurement.

[0028] Compared to star sensors, sun sensors are widely used in spacecraft systems due to their simple structure, small size, and low power consumption. Among related technologies, there are several implementation schemes for sun sensors: analog measurement schemes based on four-quadrant detectors and photodiode arrays, which calculate the solar incidence angle by collecting the solar intensity distribution and optimize measurement accuracy by combining error modeling and compensation methods; digital measurement schemes based on pinhole imaging and multi-aperture array structures, which calculate the solar vector direction based on the solar imaging position and improve accuracy by averaging multiple measurements; and correction schemes that introduce machine learning, which further improve measurement performance by modeling and correcting measurement errors. However, the above schemes can only obtain the solar vector direction and cannot obtain the roll angle of the spacecraft around the solar vector direction, nor can they directly obtain the three-axis attitude of the spacecraft. Furthermore, some schemes have high computational complexity, making it difficult to meet the high real-time performance and high robustness requirements of embedded aerospace platforms.

[0029] Therefore, under complex spatial lighting conditions, how to improve the stability of feature extraction and the robustness of matching, and improve the accuracy and reliability of attitude measurement while ensuring real-time performance, remains an urgent problem to be solved.

[0030] Therefore, this application proposes a method and apparatus for attitude determination based on solar image feature matching. The method and apparatus for attitude determination based on solar image feature matching according to embodiments of this application are described below with reference to the accompanying drawings.

[0031] Figure 1 This is a flowchart of an attitude determination method based on solar image feature matching provided according to an embodiment of this application.

[0032] like Figure 1 As shown, the attitude determination method based on solar image feature matching includes the following steps: In step S101, an image of the target sun is acquired.

[0033] In the embodiments of this application, the target solar image refers to a single or multiple frames of solar images acquired by a spacecraft, which are used as the base image for subsequent extraction of sunspot features. It should be noted that in the embodiments of this application, sunspot features are first extracted from each single frame of solar image, and then sunspot feature matching and attitude calculation are performed on two different frames of solar images to solve for the relative attitude change rotation vector between the two frames of solar images at different times.

[0034] In step S102, sunspot features that meet preset stability conditions are extracted from the target solar image. Based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, sunspot matching feature point pairs that meet preset matching conditions are obtained.

[0035] The following details how embodiments of this application extract sunspot features that meet preset stability conditions from a target solar image.

[0036] Optionally, in one embodiment of this application, extracting sunspot features that meet preset stability conditions from a target solar image includes: generating a solar region mask that meets a first preset segmentation condition based on the target solar image; generating a photosphere background image that meets preset filling conditions based on the solar region mask; constructing a corresponding background difference enhancement image based on the target solar image and the photosphere background image; extracting candidate sunspot regions that meet a second preset segmentation condition based on the background difference enhancement image, and obtaining sunspot edge regions that meet preset gradient response conditions; and generating a sunspot region mask based on the candidate sunspot regions and the sunspot edge regions, so as to extract sunspot features based on the sunspot region mask.

[0037] In the embodiments of this application, the first preset segmentation condition refers to the criterion for segmenting the effective solar region from the target solar image, which is used to generate a solar region mask. Specifically, it can be set by those skilled in the art based on the grayscale value of the effective solar region, and this application does not impose any specific restrictions.

[0038] The preset fill conditions refer to the criteria for determining the background repair and smooth filling of the solar region mask, which are used to generate the photosphere background image. Specifically, they can be set by those skilled in the art based on the characteristics of the local low-brightness areas of sunspots, and this application does not impose specific limitations.

[0039] The second preset segmentation condition refers to the criterion for segmenting candidate black spot regions in the background difference enhancement image. Specifically, it can be set by those skilled in the art based on the grayscale distribution range of the background difference enhancement image, and this application does not impose any specific restrictions.

[0040] The preset gradient response condition refers to the gradient screening criteria for judging the effectiveness of the black spot edge region. The specific criteria can be set by those skilled in the art based on the gradient response intensity of the black spot edge region. This application does not impose any specific restrictions.

[0041] The preset stability conditions refer to the comprehensive judgment criteria for the effectiveness screening of candidate black spot regions. Specifically, they can be set by those skilled in the art based on the morphology, gray level, and spatial distribution characteristics of the candidate black spot regions. This application does not impose any specific restrictions.

[0042] It is understandable that sunspots are localized low-brightness active regions on the solar photosphere, with gray values ​​significantly lower than the surrounding photosphere background. They also exhibit irregular boundaries, weak local texture variations, and dynamic scale changes. Solar images typically suffer from uneven illumination, gradual changes in background brightness, imaging noise, and interference from non-target textures. Directly employing global feature detection methods can easily generate a large number of invalid corner points in the solar photosphere background region, leading to redundant feature point distribution, increased matching computation, and a higher false-match rate. To address this, this application achieves stable feature extraction of sunspot regions through solar region constraints, background reconstruction enhancement, sunspot edge gradient response modeling, and scale-adaptive local region constraints.

[0043] by Figure 2 For example, in this embodiment of the application, the input target solar image is preprocessed, and the solar region is extracted using an adaptive threshold segmentation method to obtain a solar region mask that meets the first preset segmentation condition. It is used to eliminate interference from regions outside of solar activity background.

[0044] This embodiment of the application performs photosphere background modeling on a target solar image under the constraint of a solar region mask. Since sunspots appear as localized low-brightness areas in the image, this embodiment employs morphological closing operations to fill these low-brightness areas, resulting in a photosphere background image that meets preset filling conditions. The formula for calculating the background image of the light sphere can be, but is not limited to, the following: , in, Background image of the light sphere, For the target solar image, It is an elliptical structural element. This is a morphological dilation operation. This is a morphological erosion operation. Photosphere background modeling can effectively suppress the interference of the gradually varying brightness distribution of the solar photosphere on sunspot region detection.

[0045] This application's embodiments construct a background difference enhancement image based on the target solar image and the photosphere background image. The formula for calculating background subtraction enhancement of an image can be, but is not limited to, the following: , in, To enhance the background image through difference, The horizontal coordinate of the image. The vertical coordinates of the image are shown. After differential processing, the sunspot region exhibits high response characteristics in the background differential enhancement image, effectively highlighting the sunspot region.

[0046] This application embodiment constructs a dynamic thresholding segmentation model based on the grayscale distribution range of the background subtraction enhanced image to extract candidate black spot regions that meet a second preset segmentation condition. The expression for the dynamic thresholding segmentation model can be, but is not limited to, as follows: , in, For dynamic thresholds, The maximum grayscale value of the background subtraction enhancement image. The minimum grayscale value of the background difference enhancement image. This is a dynamically adjustable coefficient. The dynamic threshold mechanism can automatically adjust the sunspot segmentation threshold according to different imaging conditions, improving the algorithm's robustness to exposure changes and imaging noise.

[0047] Furthermore, this embodiment employs the Sobel operator to perform edge gradient response analysis on the solar region. By calculating the intensity of grayscale changes in the background difference enhanced image, a gradient response map of the sunspot edge region is constructed to determine the sunspot edge regions that meet the preset gradient response conditions. The expression for the intensity of grayscale changes can be, but is not limited to, as follows: , in, The intensity of grayscale change. For the lateral gradient component, This represents the longitudinal gradient component.

[0048] This application combines connected component analysis, area constraints, grayscale change intensity constraints, and spatial continuity constraints to jointly screen candidate black spot regions, eliminating small-area noise regions, weak texture pseudo-response regions, and regions with insufficient gradient response, thereby obtaining a black spot region mask that meets preset stability conditions.

[0049] This application embodiment, through segmentation of the target solar image, background filling modeling, background difference enhancement, extraction of candidate sunspot regions and screening of sunspot edge regions, generation of sunspot region masks and extraction of sunspot features, can accurately locate effective sunspot regions, avoid the uneven illumination, gradual changes in background brightness, imaging noise and non-target texture interference that are common in solar images, eliminate small noise areas, weak texture pseudo-response areas and areas with insufficient gradient response, improve the stability of sunspot feature extraction, and provide reliable feature input for subsequent feature matching and pose calculation.

[0050] Optionally, in one embodiment of this application, extracting sunspot features based on a sunspot region mask includes: calculating the target extension range based on the sunspot region mask; constructing a target region of interest in the sunspot region mask based on the target extension range; and extracting sunspot features based on the target region of interest.

[0051] In the embodiments of this application, the target expansion range can be dynamically adjusted by those skilled in the art based on the area of ​​the sunspot region and the scaling factor, and this application does not impose specific limitations.

[0052] The target region of interest can be dynamically adjusted by those skilled in the art based on the width of the black spot region, the height of the black spot region, and the target expansion range; this application does not impose specific limitations.

[0053] Continue with Figure 2 For example, after obtaining a black spot region mask that meets preset stability conditions, in order to improve the stability of local feature point extraction, this embodiment calculates the target expansion range based on the scale parameter of the black spot region. The expression for the target expansion range can be, but is not limited to, as follows: , in, Expand the scope of the target. This represents the area of ​​the sunspot region. This is the proportionality coefficient.

[0054] This application's embodiments construct the target region of interest based on the scale parameters of the sunspot region and the target extension range. The expression for the target region of interest can be, but is not limited to, as follows: , in, For the target region of interest, The width of the black spot region. The height of the black spot region is shown. The embodiments of this application can adaptively adjust the target region of interest according to black spot regions of different scales, ensuring complete preservation of the black spot edge region and locally stable texture regions.

[0055] This application embodiment can extract sunspot features based on a target region of interest. Specifically, within the target region of interest, combined with sunspot edge gradient response constraints, the FAST corner detection algorithm is used to extract feature points, and a non-maximum suppression algorithm is used to remove duplicate response points, resulting in an initial set of feature points. To avoid excessive clustering of feature points in local areas, this application embodiment divides the area into local grids and constrains the number of feature points within each grid to improve the spatial distribution uniformity of feature points within the sunspot region, resulting in a uniform set of feature points. Furthermore, this application embodiment uses the gray-scale centroid method to calculate the local neighborhood gray-scale distribution of feature points and calculates the principal direction of feature points through image gray-scale moments to improve the robustness of feature points to image rotation changes. In addition, this application embodiment uses an ORB (Oriented FAST and Rotated BRIEF) binary descriptor with a direction rotation compensation mechanism to encode feature points, thereby constructing a feature representation that combines rotation invariance and illumination robustness.

[0056] This application embodiment calculates the target expansion range based on the sunspot region mask, constructs the target region of interest based on the target expansion range, and extracts sunspot features within the target region of interest. It can dynamically select the range of feature extraction that is suitable for the size of the sunspot region, reduce the feature retrieval area, reduce the generation of redundant features, and improve the targeting, uniformity and stability of sunspot feature extraction.

[0057] The following details how embodiments of this application obtain sunspot matching feature point pairs that satisfy preset matching conditions based on sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints.

[0058] Optionally, in one embodiment of this application, based on sunspot features and combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, sunspot matching feature point pairs that satisfy preset matching conditions are obtained, including: determining the corresponding binary descriptor based on sunspot features; constructing a forward matching set and a reverse matching set that satisfy preset search conditions based on the binary descriptor; constructing an initial matching point set based on the forward matching set and the reverse matching set, combined with bidirectional consistency constraints; constructing a geometric interior point set based on the initial matching point set, combined with geometric consistency constraints; determining the matching position that satisfies preset response conditions based on the geometric interior point set, combined with local grayscale consistency constraints; and obtaining sunspot matching feature point pairs based on the matching position, combined with sub-pixel correction constraints.

[0059] In the embodiments of this application, the preset search conditions refer to the judgment criteria for feature matching retrieval based on the similarity of binary descriptors. The specific criteria can be set by those skilled in the art according to the matching accuracy requirements, and this application does not impose any specific limitations.

[0060] The preset response conditions refer to the criteria for judging the similarity of neighboring textures and selecting matching positions during the local grayscale consistency optimization process. The specific criteria can be set by those skilled in the art based on the image grayscale characteristics and matching accuracy requirements. This application does not impose any specific restrictions.

[0061] Bidirectional consistency constraint refers to the filtering rules for bidirectional verification of the matching relationship of sunspot features in different frames. For example, in the embodiments of this application, a forward matching set from the previous frame to the next frame and a reverse matching set from the next frame to the previous frame are constructed respectively. Only feature point pairs with corresponding bidirectional matching indices are retained, and points that do not correspond in unidirectional matching are removed to obtain an initial matching point set.

[0062] Geometric consistency constraints refer to the screening rules for examining the spatial transformation relationship of feature matching points. For example, in the embodiments of this application, the inter-frame homography matrix is ​​solved based on the initial matching point pairs, pixel projection mapping is completed through the homography matrix, the reprojection error of each matching point is calculated, and matching points whose reprojection errors do not meet the error constraints are eliminated to obtain the geometric interior point set.

[0063] Local grayscale consistency constraint refers to the screening rules for verifying the local texture matching effect of matching points. For example, in the embodiments of this application, a local image region in the neighborhood of the matching point is extracted, the similarity of the neighborhood grayscale distribution is calculated, and matching positions that meet the grayscale matching standard are screened according to preset response conditions, while invalid points with texture misalignment or grayscale mismatch are eliminated.

[0064] Subpixel correction constraints refer to precision optimization rules for matching positions at the integer pixel level. For example, in the embodiments of this application, based on the integer pixel-level matching positions, neighborhood response data is fitted to calculate the subpixel offset, offset compensation is performed on the integer pixel coordinates, and high-precision sunspot matching feature point pairs are output.

[0065] Understandably, after completing local feature extraction and ORB binary descriptor construction under the constraints of the sunspot region, it is necessary to establish a stable and reliable feature correspondence between solar images. Due to the influence of local deformation, grayscale fluctuations, and imaging noise in the sunspot region, feature matching methods based solely on descriptor distance are prone to mismatches and local drift, affecting the accuracy of subsequent image registration and attitude calculation. Therefore, this application's embodiments optimize the feature matching results step-by-step through bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, achieving high-precision and high-stability matching of the sunspot region.

[0066] by Figure 3For example, in this application embodiment, instead of directly using the one-way nearest neighbor matching result as the final correspondence, unstable matching points are gradually screened out through a multi-level constraint mechanism, and the matching result is optimized at the sub-pixel level by combining local grayscale response, thereby improving the accuracy and robustness of sunspot feature matching.

[0067] This application embodiment uses solar images. With Sun Images After extracting the ORB binary descriptor subset, first target the solar image. Descriptors corresponding to feature points In solar images The similarity of descriptors is measured by Hamming distance, and the matching descriptor corresponding to the minimum Hamming distance is searched to complete the solar image. Image pointing to the sun Positive nearest neighbor matching. The expression for single-point matching can be, but is not limited to, as follows: , in, Image of the sun The index of the matching descriptor, Image of the sun The traversal index of all descriptors in the array is , This refers to the Hamming distance operation. Image of the sun The Middle Descriptors corresponding to each feature point.

[0068] This application embodiment involves traversing solar images. After identifying and matching the feature points, all positive nearest neighbor matching results are summarized to obtain solar images that meet the preset search conditions. Image pointing to the sun positive matching set Similarly, embodiments of this application can obtain solar images that meet preset search conditions. Image pointing to the sun reverse matching set .

[0069] To mitigate the mismatch problem caused by one-way matching, this application introduces a bidirectional consistency constraint, retaining only matching pairs where bidirectional matching relationships correspond to each other, and eliminating invalid matching pairs where one-way matching is valid but reverse matching is invalid, thereby constructing an initial set of matching points. The expression for the initial set of matching points can be, but is not limited to, the following: , in, Image of the sun The Middle Feature points and solar images The Middle A matching pair consisting of feature points, Image of the sun Image pointing to the sun The set of positive matches, Image of the sun Image pointing to the sun The reverse matching set. Through the aforementioned bidirectional consistency constraints, the embodiments of this application can effectively eliminate false matching points caused by inconsistent descriptor matching.

[0070] Based on the obtained initial matching point set, this embodiment first extracts the pixel coordinates of the initial matching point set to construct a solar image. Set of matching point coordinates With Sun Images Set of matching point coordinates Since the parallax variation between solar images acquired at different times is slight, and the overall relationship approximates a planar projection, the spatial mapping relationship between two images can be characterized by homography matrix fitting. This application embodiment uses a set of matching point coordinates... Matching point coordinate set Given the conditions, iteratively solve the homography matrix using random sampling consistency constraints. Construct the initial geometric projection model. The expression for the spatial mapping relationship can be, but is not limited to, as follows: , in, Image of the sun The horizontal coordinates of the matching point Image of the sun The vertical coordinates of the matching point It is a homography matrix. Image of the sun The horizontal coordinates of the matching point Image of the sun The vertical coordinates of the matching point.

[0071] In constructing the initial geometric projection model, this embodiment of the application iteratively calculates the homography matrix parameters by randomly selecting multiple sets of matching points, continuously updating the homography matrix, and performing geometric consistency screening on the matching points by calculating the reprojection error. Specifically, for matching points whose reprojection errors meet the error constraints, this embodiment of the application classifies them as geometric interior points and constructs a set of geometric interior points based on all geometric interior points; for matching points whose reprojection errors do not meet the error constraints, this embodiment of the application classifies them as geometric exterior points and removes all geometric exterior points. It should be noted that the homography matrix that finally converges in this step is the homography matrix of the subsequent initial geometric projection model. .

[0072] After obtaining the set of geometric interior points through screening, this embodiment of the application performs local grayscale consistency optimization on the geometric interior points. This embodiment first optimizes the homography matrix of the initial geometric projection model. For each geometric interior point, its position in the solar image is calculated through spatial mapping relationships. Matching interior point locations. Solar image. The expression for matching the position of an interior point can be, but is not limited to, the following: , in, Image of the sun The matching interior point position, The homography matrix of the initial geometric projection model. Image of the sun The position of the interior point.

[0073] Subsequently, in the embodiments of this application, solar images Inner point position Define a local template region centered on the solar image. Matching interior point positions A local search region is defined around the center, and then a normalized cross-correlation function is used within this local search region to match the template with the candidate region. The expression for the normalized cross-correlation function can be, but is not limited to, as follows: , in, Image of the sun Candidate Region No. NCC response value of each pixel, Image of the sun The grayscale value of a local template region in the image. Image of the sun The grayscale value of the candidate region within the local search area. The horizontal coordinate of the pixel. The vertical coordinate of the pixel. This represents the average grayscale value of the local template area. The average grayscale value of the candidate region. This embodiment of the application traverses the local search region and calculates the NCC response value of each pixel, then selects the coordinates corresponding to the peak value of the NCC response as the optimal matching position at the integer pixel level that satisfies the preset response conditions, thus completing the local grayscale consistency optimization.

[0074] Based on obtaining the optimal matching position at the integer pixel level, this application further performs sub-pixel correction of the response peak. Since the grayscale changes of the solar image are continuous, the true optimal matching position usually falls within the pixel gaps. Therefore, this application performs parabolic interpolation fitting based on the peak neighborhood NCC response value. This application selects the peak pixel and its two left and right adjacent pixels as fitting sample points, constructs a local NCC response change curve, and calculates the sub-pixel offset. The expression for the sub-pixel offset can be, but is not limited to, as follows: , in, This is the sub-pixel offset. This represents the NCC response value of the pixel adjacent to the left of the peak pixel. This represents the NCC response value of the pixel adjacent to the right of the peak pixel. This represents the peak value of the NCC response. In this embodiment, the optimal matching position at the integer pixel level is compensated and corrected by calculating the sub-pixel offset, thereby improving the feature point matching and positioning accuracy from the pixel level to the sub-pixel level, achieving sub-pixel level feature point correction.

[0075] Furthermore, to avoid matching drift caused by localized optimization errors and to further improve overall matching stability, this application introduces a displacement anomaly constraint mechanism. This application uses the optimal matching position at the integer pixel level as a benchmark and limits the maximum offset distance between the sub-pixel corrected matching point position and the optimal matching position at the integer pixel level. If the sub-pixel offset is greater than or equal to the maximum offset distance, it is determined to be an incorrect matching point and is discarded; if the sub-pixel offset is less than the maximum offset distance, it is determined to be a correct matching point and is retained.

[0076] Furthermore, embodiments of this application can construct a high-precision matching point set, and then, based on this high-precision matching point set, use the least squares criterion to minimize the global reprojection error, iteratively solving for the optimal final homography matrix, which is used to output high-precision, high-stability sunspot matching feature point pairs. The expression for the final homography matrix can be, but is not limited to, as follows: , in, For the final homography matrix, Here are the candidate homography matrices to be iteratively optimized. Solar image after subpixel correction and displacement anomaly constraint The location of the matching feature points. Image of the sun The first in The locations of the matching feature points.

[0077] The embodiments of this application, through the above-mentioned global geometric projection model optimization, can effectively offset the problem of error accumulation caused by local single-point matching error, grayscale noise and small deformation, greatly improve the overall accuracy of the spatial mapping relationship between two solar images, and avoid the interference of local abnormal points on the global registration model.

[0078] This application embodiment constructs a forward matching set and a reverse matching set, and sequentially combines bidirectional consistency constraints, geometric consistency constraints, and local grayscale consistency constraints to filter matching positions. It also uses sub-pixel correction constraints to correct the accuracy of the matching positions, thereby achieving stable feature matching of sunspot regions from initial matching to high-precision positioning. This effectively reduces the false matching rate, improves the matching stability and positioning accuracy under weak textured solar background conditions, and provides reliable input for subsequent attitude change estimation.

[0079] In step S103, based on sunspot matching feature point pairs, a set of unit direction observation vectors for the target solar image is constructed, and a set of centered direction vectors for the target solar image is constructed based on the set of unit direction observation vectors. The spacecraft relative attitude change corresponding to the target solar image is generated based on the set of centered direction vectors.

[0080] Based on the descriptions of other embodiments, it is understood that after obtaining high-precision and high-stability sunspot matching feature point pairs, it is necessary to further calculate attitude change information. Since sunspot features essentially reflect the directional observation relationship in the camera's field of view, unlike attitude solution methods that rely on the target's three-dimensional structural model, this application embodiment does not require constructing a spatial geometric model of sunspots. Instead, it utilizes the directional consistency of sunspots during continuous observation to directly estimate the relative attitude change of the spacecraft between two frames of solar images, effectively improving the stability and robustness of attitude estimation under complex solar observation conditions.

[0081] The following details how embodiments of this application construct a set of unit-direction observation vectors for a target solar image based on sunspot matching feature point pairs, and construct a set of centered direction vectors for the target solar image based on the set of unit-direction observation vectors.

[0082] by Figure 4 For example, combined with Figure 5 As shown, this embodiment of the application acquires solar images based on sunspot matching feature point pairs. Set of matching point coordinates With Sun Images Set of matching point coordinates Then, combined with the camera imaging model, the two-dimensional pixel coordinates in the two frames of solar images are mapped to the three-dimensional observation direction vector in the camera coordinate system.

[0083] In the camera imaging model, the physical size of a pixel is Image scaling factor is The camera focal length is The coordinates of the image center are Based on the above parameters, the expression for the three-dimensional observation direction vector in the camera coordinate system can be, but is not limited to, as follows: , in, This is the 3D observation direction vector in the camera coordinate system. The x-coordinate of the image center. The ordinate of the image center. For pixel physical size, Image scaling factor This refers to the camera's focal length.

[0084] Subsequently, the three-dimensional observation direction vector is normalized in this embodiment to obtain a unit direction observation vector. The expression for the unit direction observation vector can be, but is not limited to, as follows: , in, The unit direction observation vector, Let L be the L2 norm of the three-dimensional observation direction vector.

[0085] Following the aforementioned pixel-to-vector mapping and normalization process, this application embodiment applies solar images... and images of the sun The matching feature points are calculated one by one to construct solar images. The set of unit direction observation vectors With Sun Images The set of unit direction observation vectors .

[0086] The embodiments of this application solve for the unit direction observation vector set respectively. With the set of observation vectors in the unit direction The mean vector. The expression for the mean vector can be, but is not limited to, as: , , in, The set of observation vectors in the unit direction The mean vector, The set of observation vectors in the unit direction The mean vector, The number of feature point pairs matched for sunspots. Image of the sun The Middle Unit direction observation vector of each matching feature point Image of the sun The Middle The unit direction observation vector of each matching feature point.

[0087] Furthermore, this embodiment of the application uses the mean vector to perform a centered difference calculation on the set of observation vectors in a unit direction, constructing a centered direction vector set. The expression for the centered direction vector set can be, but is not limited to, as follows: , , in, Image of the sun The set of centered direction vectors Image of the sun The set of centered direction vectors.

[0088] The following details how embodiments of this application generate the relative attitude change of the spacecraft corresponding to the target solar image based on a centralized set of direction vectors.

[0089] Optionally, in one embodiment of this application, generating the spacecraft's relative attitude change amount corresponding to the target solar image based on a centralized direction vector set includes: constructing a covariance matrix corresponding to the centralized direction vector set; performing singular value decomposition on the covariance matrix to obtain the singular value decomposition result; calculating the spacecraft's relative attitude rotation matrix corresponding to the target solar image based on the singular value decomposition result; and generating the relative attitude change amount based on the relative attitude rotation matrix.

[0090] Continue with Figure 4 For example, in this embodiment of the application, a covariance matrix is ​​constructed based on a set of centered direction vectors. The expression for the covariance matrix can be, but is not limited to, as follows: , in, Let be the covariance matrix.

[0091] Therefore, in this embodiment of the application, singular value decomposition is performed on the covariance matrix to obtain the singular value decomposition result. The decomposition expression can be, but is not limited to, as follows: , in, It is a left singular orthogonal matrix. Let covariance matrix be the variance matrix. It is a right singular orthogonal matrix.

[0092] This application embodiment, based on singular value decomposition results, employs a rigid body rotation estimation method based on orientation constraints to solve for the relative attitude rotation matrix between two frames of solar images at different times. The expression for the relative attitude rotation matrix can be, but is not limited to, as follows: , in, The relative attitude rotation matrix between two solar images at different times represents the rigid body rotation transformation relationship that minimizes the error between the observation directions of sunspots in the two images.

[0093] After obtaining the relative attitude rotation matrix, this embodiment converts the rotation matrix form into a rotation vector form to obtain the relative attitude change rotation vector. The expression for the relative attitude change rotation vector can be, but is not limited to, as follows: , in, The rotation vector is the relative attitude change between two solar images at different times. The direction of this rotation vector is the direction of the rotation axis, which is used to characterize the attitude deflection of the spacecraft during the two imaging intervals. This is a conversion operation.

[0094] This application embodiment constructs a covariance matrix from a set of centered direction vectors and performs singular value decomposition. Based on the singular value decomposition results, it solves the inter-frame relative attitude rotation matrix and generates the relative attitude change. The attitude calculation can be completed without relying on the three-dimensional geometric model of the sun, reducing the computational complexity of the algorithm and effectively improving the stability and robustness of attitude estimation under complex solar observation conditions.

[0095] Based on the above description, the embodiments of this application use the relative error of angular velocity as an evaluation index. Specifically, the embodiments of this application first calculate the observed angular velocity based on the relative attitude change rotation vector. Then, based on the Earth's rotational angular velocity... Using the reference value, calculate the relative error. The expression for the relative error can be, but is not limited to, as follows: , in, This is a relative error. To observe angular velocity, This is the theoretical angular velocity of Earth's rotation.

[0096] Calculation results from multiple sets of solar image data show that the relative error of the angular velocity in this embodiment is controlled within ±5%. This result effectively verifies that this application has high accuracy and robustness in sunspot detection, feature matching, and attitude calculation. Even when sunspot distribution is complex, local matching errors exist, or image quality is affected by noise, the reliability of the overall estimation results can still be guaranteed.

[0097] like Figure 6 As shown below, the principle of the attitude determination method based on solar image feature matching proposed in this application will be illustrated with a specific embodiment.

[0098] This embodiment of the application uses target solar images continuously acquired by a solar imaging camera on a spacecraft as input. First, the input target solar images are preprocessed, and solar regions are extracted using an adaptive threshold segmentation method to obtain a solar region mask that meets the first preset segmentation condition. Then, under the constraint of the solar region mask, photosphere background modeling is performed on the target solar image to obtain a photosphere background image that meets the preset filling condition. Then, based on the target solar image and the photosphere background image, a background difference enhancement image is constructed to complete background difference enhancement and sunspot region enhancement. Then, based on the background difference enhancement image, a sunspot region mask that meets the preset stability condition is obtained. In addition, to improve the stability of local feature point extraction, a target region of interest is constructed based on the sunspot region mask, and sunspot features are extracted within the target region of interest.

[0099] Furthermore, in this embodiment of the application, the ORB binary descriptor of sunspot features is determined to construct a forward matching set and a reverse matching set that satisfy preset search conditions. Then, the initial matching point set is obtained by cross-validation using bidirectional consistency constraints, the geometric interior point set is obtained by geometric consistency constraints, the matching position that satisfies preset response conditions is obtained by local grayscale consistency constraints, and then the integer pixel-level matching position is optimized by sub-pixel correction constraints to obtain sunspot matching feature point pairs.

[0100] Furthermore, in this embodiment, relying on the camera imaging model, the pixel coordinates of matching feature points are mapped to a three-dimensional observation direction vector in the camera coordinate system. Then, the three-dimensional observation direction vector is normalized to construct a set of unit direction observation vectors. Next, the mean vector of the unit direction observation vector set is calculated. Then, based on the mean vector, a centered difference calculation is performed on the unit direction observation vector set to construct a centered direction vector set. Next, based on the centered direction vector set, a covariance matrix is ​​constructed, and singular value decomposition is performed on the covariance matrix to obtain the singular value decomposition result. Then, based on the singular value decomposition result, the relative attitude rotation matrix between two frames of solar images at different times is solved, and the rotation matrix form is converted into a rotation vector form to obtain the relative attitude change rotation vector. This characterizes the spacecraft's attitude deflection during the two imaging intervals, thus completing the determination of the spacecraft's relative attitude.

[0101] The attitude determination method based on solar image feature matching proposed in this application utilizes the spatial distribution stability of sunspots in short-term continuous observations. Through local constraint feature extraction oriented towards stable sunspot observations, stable extraction of sunspot region features is achieved. Combined with a multi-constraint consistency optimization strategy, highly stable feature matching between solar images is realized. Furthermore, through direction vector construction and rotation matrix estimation, accurate calculation of the spacecraft's relative attitude change is achieved. This eliminates the need to construct a three-dimensional structural model of the target space, effectively reducing computational complexity and improving matching stability and attitude estimation robustness under complex solar observation conditions. It also possesses good real-time performance, making it more suitable for attitude change measurement tasks in embedded aerospace platforms. Therefore, it solves the problems of related technologies using other vector information for attitude compensation, which suffer from high computational load, insufficient real-time performance, and difficulty in meeting the high real-time and high robustness requirements of embedded aerospace platforms.

[0102] Next, with reference to the accompanying drawings, an attitude determination device based on solar image feature matching according to an embodiment of this application is described.

[0103] Figure 7 This is a block diagram of an attitude determination device based on solar image feature matching provided in an embodiment of this application.

[0104] like Figure 7 As shown, the attitude determination device 70 based on solar image feature matching includes: an acquisition module 100, an extraction module 200, and a determination module 300.

[0105] The acquisition module 100 is used to acquire the target solar image.

[0106] The extraction module 200 is used to extract sunspot features that meet preset stability conditions from the target solar image, and based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints and sub-pixel correction constraints, obtain sunspot matching feature point pairs that meet preset matching conditions.

[0107] The determination module 300 is used to construct a set of unit direction observation vectors of the target solar image based on sunspot matching feature point pairs, construct a set of centered direction vectors of the target solar image based on the set of unit direction observation vectors, and generate the spacecraft relative attitude change corresponding to the target solar image based on the set of centered direction vectors.

[0108] Optionally, in one embodiment of this application, the extraction module 200 includes: a first generation unit, a second generation unit, a first construction unit, a first extraction unit, and a second extraction unit.

[0109] The first generation unit is used to generate a solar region mask that meets the first preset segmentation conditions based on the target solar image.

[0110] The second generation unit is used to generate a photosphere background image that meets preset filling conditions based on a solar region mask.

[0111] The first building unit is used to construct a corresponding background difference enhancement image based on the target solar image and the photosphere background image.

[0112] The first extraction unit is used to extract candidate black spot regions that meet the second preset segmentation conditions based on the background difference enhancement image, and to obtain black spot edge regions that meet the preset gradient response conditions.

[0113] The second extraction unit is used to generate a sunspot region mask based on the candidate sunspot region and the sunspot edge region, so as to extract sunspot features based on the sunspot region mask.

[0114] Optionally, in one embodiment of this application, the second extraction unit includes: a calculation subunit, a construction subunit, and an extraction subunit.

[0115] The computational subunit is used to calculate the target expansion range based on the sunspot region mask.

[0116] Construct sub-units to build the target region of interest in the sunspot region mask based on the target's extended range.

[0117] Extract sub-units for extracting sunspot features based on the target region of interest.

[0118] Optionally, in one embodiment of this application, the extraction module 200 includes: a first determining unit, a second constructing unit, a third constructing unit, a fourth constructing unit, a second determining unit, and a third generating unit.

[0119] The first determining unit is used to determine the corresponding binary descriptor based on the characteristics of sunspots.

[0120] The second building unit is used to construct a positive matching set and a negative matching set that satisfy preset search conditions based on binary descriptors.

[0121] The third building unit is used to construct an initial set of matching points based on the forward matching set and the reverse matching set, combined with bidirectional consistency constraints.

[0122] The fourth building unit is used to construct a set of geometric interior points based on the initial set of matching points and geometric consistency constraints.

[0123] The second determining unit is used to determine the matching position that satisfies the preset response conditions based on the set of geometric interior points and in combination with local grayscale consistency constraints.

[0124] The third generation unit is used to obtain sunspot matching feature point pairs based on the matching position and combined with sub-pixel correction constraints.

[0125] Optionally, in one embodiment of this application, the determining module 300 includes: a fifth building unit, a decomposition unit, a calculation unit, and a fourth generating unit.

[0126] The fifth building unit is used to construct the covariance matrix corresponding to the set of centered direction vectors.

[0127] The decomposition unit is used to perform singular value decomposition on the covariance matrix to obtain the singular value decomposition result.

[0128] The computational unit is used to calculate the relative attitude rotation matrix of the target solar image based on the singular value decomposition results.

[0129] The fourth generation unit is used to generate relative attitude change based on the relative attitude rotation matrix.

[0130] It should be noted that the foregoing explanation of the attitude determination method based on solar image feature matching also applies to the attitude determination device based on solar image feature matching in this embodiment, and will not be repeated here.

[0131] The attitude determination device based on solar image feature matching proposed in this application utilizes the spatial distribution stability of sunspots during short-term continuous observations. Through local constraint feature extraction oriented towards stable sunspot observations, stable extraction of sunspot region features is achieved. Combined with a multi-constraint consistency optimization strategy, highly stable feature matching between solar images is realized. Furthermore, through direction vector construction and rotation matrix estimation, accurate calculation of the spacecraft's relative attitude change is achieved. This eliminates the need to construct a three-dimensional structural model of the target space, effectively reducing computational complexity and improving matching stability and attitude estimation robustness under complex solar observation conditions. It also possesses good real-time performance, making it more suitable for attitude change measurement tasks in embedded aerospace platforms. Therefore, it solves the problems of related technologies using other vector information for attitude compensation, which suffer from high computational load, insufficient real-time performance, and difficulty in meeting the high real-time and high robustness requirements of embedded aerospace platforms.

[0132] Figure 8 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. The electronic device may include: The memory 801, the processor 802, and the computer program stored on the memory 801 and capable of running on the processor 802.

[0133] When the processor 802 executes the program, it implements the attitude determination method based on solar image feature matching provided in the above embodiments.

[0134] Furthermore, electronic devices also include: Communication interface 803 is used for communication between memory 801 and processor 802.

[0135] The memory 801 is used to store computer programs that can run on the processor 802.

[0136] The memory 801 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0137] If the memory 801, processor 802, and communication interface 803 are implemented independently, then the communication interface 803, memory 801, and processor 802 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized into address buses, data buses, control buses, etc. For ease of representation, Figure 8 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0138] Optionally, in a specific implementation, if the memory 801, processor 802, and communication interface 803 are integrated on a single chip, then the memory 801, processor 802, and communication interface 803 can communicate with each other through an internal interface.

[0139] The processor 802 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0140] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described attitude determination method based on solar image feature matching.

[0141] This application also provides a computer program product, including a computer program that, when executed, implements the above-described attitude determination method based on solar image feature matching.

[0142] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0143] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0144] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0145] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0146] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0147] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0148] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0149] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A pose determination method based on solar image feature matching, characterized in that, Includes the following steps: Acquire target solar image; Sunspot features that meet preset stability conditions are extracted from the target solar image. Based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, sunspot matching feature point pairs that meet preset matching conditions are obtained. Based on the sunspot matching feature point pairs, a set of unit direction observation vectors for the target solar image is constructed, and a set of centered direction vectors for the target solar image is constructed based on the set of unit direction observation vectors. Based on the set of centered direction vectors, the spacecraft relative attitude change corresponding to the target solar image is generated.

2. The method according to claim 1, characterized in that, Extracting sunspot features that meet preset stability conditions from the target solar image includes: Based on the target solar image, a solar region mask that meets the first preset segmentation condition is generated; Based on the solar region mask, a photosphere background image that meets the preset filling conditions is generated; Based on the target solar image and the photosphere background image, a corresponding background difference enhancement image is constructed; Based on the background difference enhancement image, candidate black spot regions that meet the second preset segmentation conditions are extracted, and black spot edge regions that meet the preset gradient response conditions are obtained. A black spot region mask is generated based on the candidate black spot region and the black spot edge region.

3. The method according to claim 2, characterized in that, The extraction of sunspot features based on the sunspot region mask includes: Calculate the target expansion range based on the sunspot region mask; Based on the target expansion range, a target region of interest is constructed in the sunspot region mask; The sunspot features are extracted based on the target region of interest.

4. The method according to claim 1, characterized in that, The process of obtaining sunspot matching feature point pairs that satisfy preset matching conditions based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints, and sub-pixel correction constraints, includes: Based on the sunspot characteristics, the corresponding binary descriptor is determined; Based on the binary descriptor, construct a positive matching set and a negative matching set that satisfy the preset search conditions; Based on the forward matching set and the reverse matching set, an initial matching point set is constructed in conjunction with the bidirectional consistency constraint; Based on the initial set of matching points, a set of geometric interior points is constructed in conjunction with the geometric consistency constraints; Based on the set of geometric interior points, and in conjunction with the local grayscale consistency constraint, the matching position that satisfies the preset response condition is determined; Based on the matching position and combined with the sub-pixel correction constraint, the sunspot matching feature point pair is obtained.

5. The method according to claim 1, characterized in that, The process of generating the spacecraft's relative attitude change corresponding to the target solar image based on the centralized direction vector set includes: Construct the covariance matrix corresponding to the set of centered direction vectors; Singular value decomposition is performed on the covariance matrix to obtain the singular value decomposition results; Based on the singular value decomposition results, the relative attitude rotation matrix of the target solar image is calculated; The relative attitude change is generated based on the relative attitude rotation matrix.

6. An attitude determination device based on solar image feature matching, characterized in that, include: The acquisition module is used to acquire images of the target sun. The extraction module is used to extract sunspot features that meet preset stability conditions from the target solar image, and based on the sunspot features, combined with bidirectional consistency constraints, geometric consistency constraints, local grayscale consistency constraints and sub-pixel correction constraints, obtain sunspot matching feature point pairs that meet preset matching conditions. The determination module is used to construct a set of unit direction observation vectors for the target solar image based on the sunspot matching feature point pairs, construct a set of centered direction vectors for the target solar image based on the set of unit direction observation vectors, and generate the spacecraft relative attitude change amount corresponding to the target solar image based on the set of centered direction vectors.

7. The apparatus according to claim 6, characterized in that, The extraction module includes: The first generation unit is used to generate a solar region mask that meets the first preset segmentation conditions based on the target solar image; The second generation unit is used to generate a photosphere background image that meets preset filling conditions based on the solar region mask; The construction unit is used to construct a corresponding background difference enhancement image based on the target solar image and the photosphere background image; The first extraction unit is used to extract candidate black spot regions that meet the second preset segmentation conditions based on the background difference enhancement image, and to obtain black spot edge regions that meet the preset gradient response conditions. The second extraction unit is used to generate a sunspot region mask based on the candidate sunspot region and the sunspot edge region, so as to extract the sunspot features based on the sunspot region mask.

8. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and capable of running on the processor, the processor executing the program to implement the attitude determination method based on solar image feature matching as described in any one of claims 1-5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the attitude determination method based on solar image feature matching as described in any one of claims 1-5.

10. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the attitude determination method based on solar image feature matching as described in any one of claims 1-5.

Images