Space cooperative target pose vision measurement system

Abstract

The invention discloses a space cooperative target pose vision measurement system, and relates to the technical field of pose vision measurement. The invention provides a measurement system adopting a visible light passive cooperative target subsystem and a vision measurement subsystem in order to overcome the defects of an active illumination spectrum difference scheme space cooperative target pose vision measurement system. According to the system, high-performance background noise suppression, mark point extraction and pose calculation are achieved, the cooperative target subsystem is simple in structure and long in on-orbit service life, and the vision measurement subsystem can meet various space photoelectric detection task requirements. Based on the visible light passive cooperative target and the binocular vision system, real-time high-precision pose measurement of the space target is realized, the complexity and the cost of the measurement system are reduced, and the service life of the measurement system is prolonged.

Description

Technical Field

[0001] The present application relates to the technical field of posture visual measurement, and in particular to a spatial cooperative target posture visual measurement system. Background Art

[0002] To address background interference, existing cooperative target position and visual measurement systems typically employ active illumination spectral differentiation. This approach requires the measurement system to utilize two spectral illumination sources. The cooperative target is configured to absorb light from one spectral band and reflect light from the other. This allows for controlled acquisition of cooperative target images illuminated by both spectral bands. Image differentiation removes background interference, retaining valid landmark information for target resolution. While this approach offers a simple data processing workflow, the dual-spectrum active illumination system is complex and costly, making it unsuitable for widespread adoption. Summary of the Invention

[0003] The purpose of this application is to provide a space cooperative target posture visual measurement system, which can realize real-time high-precision posture measurement of space targets based on visible light passive cooperative targets and binocular vision systems, thereby reducing the complexity and cost of the measurement system.

[0004] To achieve the above objectives, this application provides the following solutions.

[0005] In a first aspect, the present application provides a space cooperative target pose visual measurement system, comprising: a cooperative target subsystem and a visual measurement subsystem, wherein the cooperative target subsystem is arranged on a target aircraft, and the visual measurement subsystem is arranged on a service aircraft;

[0006] The visual measurement subsystem includes a visible light binocular camera and an information processor;

[0007] The visible light binocular camera is used to obtain a binocular measurement image of the cooperative target subsystem;

[0008] The information processor is used to perform landmark extraction, target recognition and pose calculation on the binocular measurement image.

[0009] Optionally, the target pattern of the cooperative target subsystem includes four coplanar A-type target components and one B-type target component, wherein three A-type target components and one B-type target component are distributed at the four vertices of a square, and another A-type target component is located at the center of the square.

[0010] Optionally, the type A target assembly and the type B target assembly each include a black central circle and a plurality of black and white rings arranged outside the black central circle; wherein the ring closest to the black central circle is a white ring;

[0011] The diameter of the black center circle of the B-type target assembly is larger than the diameter of the black center circle of the A-type target assembly. The circular ring of the A-type target assembly corresponds one-to-one to the circular ring of the B-type target assembly and has the same diameter.

[0012] Optionally, the binocular measurement image includes a left camera measurement image and a right camera measurement image. In terms of extracting landmark points from the binocular measurement image, the information processor is specifically configured to:

[0013] Performing connected region extraction on the left camera measurement image and the right camera measurement image respectively to obtain a plurality of first connected regions in the left camera measurement image and a plurality of second connected regions in the right camera measurement image;

[0014] Screening the plurality of first connected regions in the left camera measurement image and the plurality of second connected regions in the right camera measurement image respectively to determine a plurality of first candidate marker point connected regions in the left camera image and a plurality of second candidate marker point connected regions in the right camera image;

[0015] Calculating characteristic parameters of the connected areas of the plurality of first candidate marker points and the connected areas of the plurality of second candidate marker points respectively to obtain characteristic parameters of the plurality of first candidate marker points and characteristic parameters of the plurality of second candidate marker points;

[0016] The plurality of first candidate marker points and the plurality of second candidate marker points are matched according to the characteristic parameters of the plurality of first candidate marker points and the characteristic parameters of the plurality of second candidate marker points to obtain matching marker point pairs.

[0017] Optionally, the screening of the plurality of first connected areas in the left camera measurement image and the plurality of second connected areas in the right camera measurement image to determine the plurality of first candidate marker point connected areas in the left camera image and the plurality of second candidate marker point connected areas in the right camera image specifically includes:

[0018] Calculate the area, aspect ratio, concavity, area ratio and circularity of the target connected region;

[0019] When the area, aspect ratio, concavity, area ratio and roundness of the target connected area meet the first preset condition, the target connected area is determined to be the candidate marker point connected area; the target connected area is the first connected area or the second connected area, when the target connected area is the first connected area, the candidate marker point connected area is the first candidate marker point connected area, and when the target connected area is the second connected area, the candidate marker point connected area is the second candidate marker point connected area.

[0020] Optionally, the first preset condition is: σ Amin ≤A IPx ≤σAmax &k ab min ≤k ab &d concavity min ≤d concavity &k area min ≤k area &d circularity min ≤d circularity ;

[0021] Among them, A IPx is the area of the target connected region, σ Amin and σ Amax are the minimum threshold for area screening and the maximum threshold for area screening, k ab is the width-to-length ratio of the target connected area, k ab min is the minimum threshold for aspect ratio screening, d concavity is the concavity of the target connected region, d concavity min is the minimum threshold for concavity screening, k area is the area ratio of the target connected region, k area min is the minimum threshold for area ratio screening, d circularity is the circularity of the target connected area, d circularity min Filter the circularity threshold for the target connected region.

[0022] Optionally, matching the plurality of first candidate marker points with the plurality of second candidate marker points according to the characteristic parameters of the plurality of first candidate marker points and the characteristic parameters of the plurality of second candidate marker points to obtain matching marker point pairs specifically includes:

[0023] Calculating the epipolar distance between the first target candidate landmark point and each second candidate landmark point; the first target candidate landmark point is any first candidate landmark point;

[0024] Determine the second candidate landmark point with the smallest epipolar distance as the second target candidate landmark point;

[0025] Calculating a radius deviation rate, an area deviation rate, and an area ratio deviation rate between the first target candidate marker point and the second target candidate marker point based on characteristic parameters of the first target candidate marker point and characteristic parameters of the second target candidate marker point;

[0026] When the epipolar distance, the radius deviation rate, the area deviation rate, and the area ratio deviation rate meet the second preset condition, the first target candidate landmark point and the second target candidate landmark point are combined into a matching landmark point pair.

[0027] Optionally, the second preset condition is:

[0028] d LA ≤d LAmax &Δr T ≤Δ max &ΔATPx ≤Δ max &Δk Tarea ≤Δ max ;

[0029] Among them, d LA , Δr T , ΔA TPx , Δk Tarea are the epipolar distance, radius deviation rate, area deviation rate, and area ratio deviation rate between the first target candidate landmark point and the second target candidate landmark point, respectively. LA max is the epipolar distance matching screening threshold, Δ max Filter threshold for the maximum deviation rate of feature parameters.

[0030] Optionally, in the aspect of target identification, the information processor is specifically used to:

[0031] The landmark points included in the matching landmark point pairs are used as target landmark points and form a target landmark point set;

[0032] Calculating the three-dimensional spatial position and area ratio of each target marker point in the target marker point set;

[0033] Determining a target marker point corresponding to the type B target component in the target marker point set according to the area ratio as a benchmark reference point, and obtaining position coding information of the benchmark reference point;

[0034] Calculating reference distances of target marker points other than the benchmark reference point in the target marker point set based on the three-dimensional spatial position of the benchmark reference point and the three-dimensional spatial positions of the target marker points other than the benchmark reference point in the target marker point set;

[0035] Determine the target marker point with the smallest reference distance as the first angle reference point, and obtain the position coding information of the first angle reference point; the first angle reference point is the target marker point of the A-type target assembly located at the center of the square;

[0036] Determine the target marker point with the largest reference distance as the second angle reference point, and obtain the position coding information of the second angle reference point; the vertex of the square where the type A target component corresponding to the second angle reference point is located is located on the same diagonal line as the vertex where the type B target component is located;

[0037] determining the first angle reference point or the second angle reference point as the target angle reference point;

[0038] According to the three-dimensional spatial position of the reference reference point and the spatial position of the target angle reference point, a vector from the reference reference point to the target angle reference point is determined as a reference vector;

[0039] According to the three-dimensional spatial position of the reference reference point and the spatial position of the target marker point to be identified, a vector from the reference reference point to the target marker point to be identified is determined as a reference vector; the target marker point to be identified is any target marker point in the target marker point set except the reference reference point, the first angle reference point and the second angle reference point;

[0040] Calculating the angle between the reference vector and the reference vector as a reference angle of the target marker to be identified;

[0041] The position coding information of the target marker to be identified is determined according to the reference angle.

[0042] Optionally, in terms of pose solution, the information processor is specifically used to:

[0043] Define the objective function:

[0044]

[0045] Q Ti =P Ti -P T ;

[0046]

[0047] Among them, W is the objective function, Q Ti is the relative position of target landmark point i and the centroid of all target points in the target coordinate system, Q Wi is the relative position of target marker point i and the centroid of all target points in the measurement coordinate system, n is the number of target marker points, and the superscript T indicates transposition; P Ti is the theoretical position of target marker point i in the target coordinate system, P T is the theoretical position of the center of mass of all target points in the target coordinate system, P Wi is the measured position of target point i in the measurement coordinate system, P W is the measured position of the centroid of all target points in the measurement coordinate system;

[0048] Perform singular value decomposition on the objective function to obtain the singular value decomposition result: W = UΣV T , where Σ is the diagonal matrix composed of the singular values of W, U and V are the first and second diagonal matrices in the singular value decomposition results respectively;

[0049] According to the singular value decomposition results, the rotation matrix is obtained: R TW =UV T , where R TW is the rotation matrix;

[0050] According to the rotation matrix, the translation vector is determined as: TW=P W -R TW P T , where t TW is the translation vector;

[0051] The position and posture of the target aircraft relative to the visible light binocular camera are determined according to the rotation matrix, the translation vector and the position and posture of the target aircraft in the binocular measurement image.

[0052] According to the specific embodiments provided in this application, this application has the following technical effects.

[0053] The present application provides a space cooperative target posture visual measurement system. In response to the shortcomings of the space cooperative target posture visual measurement system using the active illumination spectrum difference scheme, the present application proposes a measurement system that uses a visible light passive and passive cooperative target subsystem and a visual measurement subsystem. The system achieves high-performance background noise suppression, landmark point extraction, and posture solution. The cooperative target subsystem has a simple structure and a long on-orbit lifespan, and the visual measurement subsystem can meet the needs of various space optoelectronic detection missions. Based on the visible light passive and passive cooperative target and binocular vision system, the present application achieves real-time, high-precision posture measurement of space targets, reducing the complexity and cost of the measurement system. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0055] Figure 1 A schematic structural diagram of a spatial cooperative target pose visual measurement system provided in one embodiment of the present application.

[0056] Figure 2 A structural diagram of a target assembly provided in one embodiment of the present application.

[0057] Figure 3 This is a target pattern design diagram of the cooperative target subsystem provided in one embodiment of the present application.

[0058] Figure 4 A schematic diagram of the coordinate system definition of the visual measurement subsystem provided in one embodiment of the present application.

[0059] Figure 5 A schematic diagram of the pose solution of a spatial cooperative target pose visual measurement system provided in one embodiment of the present application.

[0060] Figure 6A flowchart of connected region extraction provided in one embodiment of the present application.

[0061] Figure 7 A schematic diagram of morphological feature parameters for screening landmark points provided in one embodiment of the present application.

[0062] Figure 8 This is a diagram of the definition of cooperative target position encoding information algorithm identification parameters provided in one embodiment of the present application.

[0063] Figure 9 A schematic diagram of the 6-DOF real-time pose measurement results of a space target provided in one embodiment of the present application. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0065] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0066] In an exemplary embodiment, a spatial cooperative target pose visual measurement system is provided, such as Figure 1 As shown, it includes: a cooperative target subsystem and a visual measurement subsystem, the cooperative target subsystem is set on the target aircraft, and the visual measurement subsystem is set on the service aircraft; the visual measurement subsystem includes a visible light binocular camera and an information processor; the visible light binocular camera is used to obtain binocular measurement images of the cooperative target subsystem; the information processor is used to perform landmark point extraction, target recognition and pose solution on the binocular measurement images.

[0067] The space cooperative target posture visual measurement system in the above embodiment includes a cooperative target subsystem and a visual measurement subsystem, which is used for real-time high-precision posture measurement of space targets with 6 degrees of freedom. Figure 1As shown in the figure, the cooperative target subsystem is installed on the target aircraft for pose measurement, and the relative positional relationships of the various target components are known. The visual measurement subsystem is installed on the service aircraft and mainly includes a visible light binocular camera and an information processor (including a measurement algorithm). The system's pose measurement principle is as follows: the binocular camera images the target aircraft, and the algorithm extracts the image position of the cooperative target point from the measurement images of the left and right cameras. The target point is identified based on the known target point pattern coding features and position coding features. The position and attitude of the target aircraft relative to the measurement system are calculated based on the known positional relationship of the target point in the target coordinate system and its position in the measurement image.

[0068] In another exemplary embodiment, the cooperative target subsystem is composed of two target components (type A and type B), such as Figure 2 As shown, Figure 2 (a) is a schematic diagram of the A-type target assembly. Figure 2 (b) in the figure shows a schematic diagram of a Type B target assembly. Types A and B target assemblies are numerically encoded using black circular patterns of varying diameters in their centers. The algorithm identifies the target assembly based on the area ratio of the central black region. The target assembly is machined from aluminum alloy. The black portion is treated with a black anodized surface, while the white portion is sprayed with white paint, resulting in a high-contrast circular outline in the image.

[0069] The target pattern of the cooperative target subsystem consists of five coplanar mounted target assemblies, including four A-type target assemblies (respectively Figure 3 A1, A2, A3, A4) and 1 B-type target component (for Figure 3 B), the target pattern design is as follows Figure 3 As shown, the four targets A1, A2, A3, and B are located at the four vertices of a square with a side length of 255 mm, and the target A4 is located at the center of the square. T The target is defined as the center of the target pattern square, O T -Y T The axis is parallel to the line connecting A1 to A2. T -Z T The axis is parallel to the line connecting A2 to A3, O T -X T The axis is perpendicular to the target mounting plane and complies with the right-hand rule.

[0070] In another exemplary embodiment, the visual measurement subsystem is a binocular vision measurement system consisting of two visible light cameras. The baseline B of the binocular vision measurement system is defined as the distance between the optical centers of the two cameras. In this system, the baseline distance is designed to be 1.3m. The baseline direction is aligned with the O coordinate system of the left and right cameras. C -Y C Axis vertical. C-Z C The angle between the axis and the baseline is the camera's adduction angle α, which is designed to be 76° in this system. W is the midpoint of the baseline, O W -Y W The axis is parallel to the baseline and points from the left camera to the right camera, O W -Z W Axis and camera coordinate system O C -Y C Axis parallel, O W -X W The axis follows the right-hand rule, such as Figure 4 shown.

[0071] In another exemplary embodiment, the binocular camera measurement image is processed by a measurement algorithm to extract the target information of the cooperative target and calculate the relative position and posture of the cooperative target. The main processing flow is as follows: Figure 5 As shown in the figure, the measurement images of the two cameras are first processed separately, and the feature parameters of the candidate landmark points in the image are obtained by connecting area extraction and landmark point screening and positioning; then, the feature parameters of the candidate landmark points obtained by the two cameras are combined to perform dual-target landmark matching and coded information recognition, and determine the mapping relationship between the landmark points in the image and the known cooperative target model; finally, the six-degree-of-freedom relative pose of the cooperative target is solved based on the landmark point feature parameters and the mapping relationship.

[0072] 1) Connected region extraction.

[0073] Threshold segmentation is used to quickly remove most of the background noise in the measurement image, resulting in a binary image containing the white circular region of the cooperative target. Eight-neighborhood connected regions are extracted for the foreground pixels in the binary image. During the calculation process, each foreground pixel in the binary image is scanned one by one, and connected regions are marked based on their neighborhoods. Ultimately, all independent connected regions are identified.

[0074] Processing flow such as Figure 6 As shown, the main steps include the following steps 101 to 103.

[0075] Step 101: Mark connected areas.

[0076] Scan the binary image pixel by pixel along the row direction from left to right and from top to bottom. If a foreground pixel is scanned, the connected region is marked based on the status of its four upper-left neighboring pixels. The process proceeds one by one in the order of left, upper-left, upper, and upper-right. If the first neighboring pixel is a foreground pixel, the connected region label value of the currently scanned pixel is marked as the connected region label value of the foreground neighboring pixel. If all four upper-left neighboring pixels are background pixels, the currently scanned pixel is marked as the new connected region label value.

[0077] Step 102: Update the equivalence table.

[0078] The equivalence table records the relationships between connected domains. The first row of the equivalence table contains the label values of the connected domains in ascending order. The second row indicates the label values of the equivalent connected domains that are connected to the label value region in the first row during the scan. When a new label value is required for a scanned pixel, a new column of data is added to the end of the equivalence table. When the label value of a scanned pixel needs to be determined based on neighboring pixels, the label value of the equivalent domain corresponding to the neighboring label value in the equivalence table is used for the assignment.

[0079] During pixel-by-pixel scanning, only equivalence operations between two label values can be established. This can result in complex connected regions consisting of multiple label values, but appearing as implicit multi-layer equivalence relationships in the equivalence table. To simplify subsequent labeling of truly connected regions, the equivalence table must be refreshed and sorted after the pixel-by-pixel scan to ensure that only one layer of mapping exists. The equivalence table is traversed, and each equivalent label value is mapped iteratively using the label values in the second row, ensuring that each equivalent label value in the table is sorted to the minimum value.

[0080] Step 103: Count the true connected areas.

[0081] Based on the updated equivalence table, the label matrix is scanned a second time, and all the original label values in the label matrix are updated to their corresponding equivalent label values in the equivalence table. Based on the final label matrix, the number of connected regions, the area of each connected region, and the row and column ranges are counted.

[0082] 2) Landmark point screening and positioning.

[0083] The extracted connected areas are screened by image morphological feature parameters, and the circular areas are retained as candidate landmarks in the image. Figure 7 As shown, the graphical morphological characteristic parameters used for screening include:

[0084] A1. Area of connected regions:

[0085] The area of the connected region A TPx Defined as the number of foreground pixels in the extracted connected region. TPx The result obtained when precalculating the parameters each time For comparison, the maximum area screening coefficient σ is set Amax and the minimum area screening coefficient σ Amin To adjust the connected area screening range.

[0086] A2. Aspect Ratio:

[0087] Fit the minimum circumscribed rectangle of the connected area with a width-to-length ratio of k abIt is defined as the ratio of the short side length a to the long side length b of the minimum circumscribed rectangle of the connected area, which is:

[0088] k ab =a / b

[0089] From the definition, we know that k ab ≤1, when the connected area is circular k ab = 1. By setting the minimum aspect ratio threshold k abmin To adjust the connected area screening range.

[0090] A3, concavity:

[0091] The outer contour of the connected area is extracted and the convex hull contour is screened, and the concavity d concavity Defined as the length of the convex hull contour of the connected region l convex The length of the outer contour of the connected area l blob The ratio of is:

[0092] d concavity =l convex / l blob

[0093] From the definition, we know that d concavity ≤1, when the connected area is circular concavity =1, when there is a defect on the circular edge d concavity <1. By setting the minimum concavity threshold d concavitymin To adjust the connected area screening range.

[0094] A4. Area ratio:

[0095] Calculate the minimum circumscribed circle of the connected area, and the area ratio k area Defined as the area of the connected region A TPx Its minimum circumscribed circle area A Tcircle The ratio of is:

[0096] k area =A TPx / A Tcircle

[0097] From the definition, we know that k area ≤1, k area The closer it is to 1, the closer the connected area is to a circle. By setting the minimum area ratio threshold k areamin To adjust the connected area screening range.

[0098] A5. Roundness:

[0099] Roundness d circularity Defined as the length of the convex hull contour of the connected region l convex Its minimum circumference C Tcircle The ratio of is:

[0100] d circularity =l convex / C Tcircle

[0101] From the definition, we know that d circularity ≤1, the roundness is insensitive to small-scale edge defects of the connected area contour and has good robustness to the defects of the marker pattern caused by shadows, glare and other phenomena. By setting the minimum roundness threshold d circularitymin To adjust the connected area screening range.

[0102] The computational complexity of the morphological feature parameters of the connected regions increases gradually, and the computational speed is improved by comparing and calculating the feature parameters of different priorities layer by layer.

[0103] In the embodiment of the present application, a first preset condition is set for screening, and the first preset condition is:

[0104] σ Amin ≤A IPx ≤σ A max &k ab min ≤k ab &d concavity min ≤d concavity &k area min ≤k area &d circularity min ≤d circularity ;

[0105] Among them, A IPx is the area of the target connected region, σ Amin and σ Amax are the minimum threshold for area screening and the maximum threshold for area screening, k ab is the width-to-length ratio of the target connected area, k ab min is the minimum threshold for aspect ratio screening, d concavity is the concavity of the target connected region, d concavity min is the minimum threshold for concavity screening, k area is the area ratio of the target connected region, k area min is the minimum threshold for area ratio screening, d circularity is the circularity of the target connected area, d circularity min Filter the circularity threshold for the target connected region.

[0106] The connected region morphological feature parameter screening thresholds used in the embodiment of the present application are σ Amax =1.5,σ Amin =0.7, k abmin =0.75, d concavitymin =0.85, k areamin =0.75, d circularitymin =0.85.

[0107] The connected regions of candidate landmark points are obtained by screening the morphological feature parameters. The convex hull contours of these connected regions are fitted with elliptical shapes to calculate the feature parameters of candidate landmark points required for target recognition and pose calculation, including:

[0108] B1, the image position of the candidate landmark point (u T ,v T ), defined as the sub-pixel position of the center of the fitted ellipse in the image;

[0109] B2, image radius r of candidate landmark points T , defined as the sub-pixel length of the semi-major axis of the fitted ellipse;

[0110] B3. Area A of the alternative landmark point TPx , defined as the foreground pixel area of the connected region;

[0111] B4, area ratio k of alternative landmark points Tarea , defined as the ratio of the foreground pixel area of the connected region to the area of the fitted ellipse.

[0112] 3) Dual-target landmark matching.

[0113] The candidate landmark information obtained by the left and right cameras needs to be matched and screened through the epipolar geometry relationship of the binocular vision measurement system. Suppose the candidate landmark point p is extracted from the left camera image. L (u L ,v L ), extract candidate landmark points p from the right camera image R (u R ,v R ), then the epipolar distance d between the two candidate landmarks in the right camera coordinate system LR It can be expressed as:

[0114]

[0115] Where A LR 、B LR 、C LR ——Click p L The corresponding epipolar equation parameters in the right camera coordinate system can be calculated by the following formula:

[0116] [A LR B LR C LR ] T =Fp L =F[u L v L 1] T

[0117] Where F is the basic matrix of the binocular vision measurement system, which is obtained by calibrating the internal and external parameters of the system.

[0118] If the candidate landmarks in the left and right camera images correspond to the same cooperative target, the theoretical epipolar distance d LR Should be 0. In the algorithm processing, the candidate markers of one camera are used as the benchmark, and the limit distance with the candidate markers of the other camera is calculated according to the above formula. The candidate marker matching pair with the smallest epipolar distance is selected, and its limit distance is set as d LRmin At the same time, the characteristic parameter deviation rate of the candidate landmark matching pair is calculated, including the radius deviation rate Δr T , area deviation rate ΔA TPx , area ratio deviation rate Δk Tarea , the specific calculation formula is as follows:

[0119]

[0120] Set the epipolar distance matching screening threshold d LRmax , the maximum deviation rate screening threshold of characteristic parameters Δ max , when the candidate landmark matching pairs satisfy d LA ≤d LAmax &Δr T ≤Δ max &ΔA TPx ≤Δ max &Δk Tarea ≤Δ max When , it is recorded as a valid matching landmark point pair. In the embodiment of the present application, the matching screening threshold is set to d LRmax =15,Δ max =0.2, where d LRmax The value of is related to the calibration accuracy of the visual measurement subsystem.

[0121] According to the binocular vision measurement principle, the three-dimensional spatial position P of the target component corresponding to the effective matching mark point in the measurement system coordinate system is calculated. T (x T ,y T ,z T ). According to the camera imaging model:

[0122]

[0123] Where M L ,M R ——The projection matrix of the left and right cameras relative to the visual measurement system coordinate system, including the camera's intrinsic parameter matrix and extrinsic parameter matrix, obtained by system calibration, z T For depth.

[0124] The above formulas can be combined to solve PT The four linear equations can be written in matrix form as follows:

[0125] AP T =G

[0126] Solving using the least squares method yields:

[0127] P T =(A T A) -1 A T G

[0128] At the same time, the characteristic parameters of the target landmark are calculated, including: radius area and area ratio They are:

[0129]

[0130] The three-dimensional spatial position P of the target landmark obtained by binocular matching screening T , image radius area Area ratio Will be used for target encoding information identification.

[0131] 4) Identification of coded information.

[0132] The number mapping relationship between image landmarks and cooperative target models is determined by the graphic coding and position coding information of the cooperative target. The graphic coding information is represented by characteristic parameters such as the image radius, area, and area ratio of the landmark. The main task is to identify the B-type target component in the image.

[0133] The target landmark point set after dual-target landmark point matching can be expressed as:

[0134] S T ={T1,T2,…,T i}

[0135] Where, T i ——The i-th target landmark in the set.

[0136] From the set S T Find the B-type cooperative target marker T B The judgment condition can be expressed as:

[0137]

[0138] Where k TBarea ——Theoretical area ratio of type B cooperative targets,

[0139] Δk TBarea——Screening deviation value of type B cooperative target area ratio.

[0140] The theoretical area ratio of the B-type cooperative target in the embodiment of the present application is k TBarea =0.75, the area ratio screening deviation value is set to Δk TBarea =0.1.

[0141] The target position coding information is based on the B-type target as the reference point, and includes two characteristic parameters: reference distance and reference angle. i Reference distance D(T i ) indicates T i With T B The Euclidean distance in three-dimensional space is:

[0142] D(T i )=|P T (T i )-P T (T B )|

[0143] Through cooperative target pattern design, the reference distance of each target point has obvious and distinguishable differences. The algorithm can calculate the reference distance of each target landmark by measuring the three-dimensional coordinates, and match it with the theoretical reference distance to achieve number recognition of most of the landmarks.

[0144] The cooperative target subsystem selects the target component with obvious distinction (such as the largest or smallest distance) as the angle reference point, and sets it as T C ,like Figure 8 In the embodiment of the present application, the target A2 or A4 in the cooperative target subsystem is selected as the angle reference point.

[0145] Determine the reference point T B and angle reference point T C After that, the target landmark point T i The reference angle θ(T i ) using vector and vector The angle between them is expressed as:

[0146]

[0147] Except T B and T C Other target landmarks can be matched with the known coding position coding information of the cooperative target subsystem through the two characteristic parameters of reference distance and reference angle to complete the target point number mapping and identification process.

[0148] 5) Solving cooperative target pose.

[0149] According to the algorithm, the three-dimensional position information of the target marker and the definition of the cooperative target subsystem coordinate system are identified to solve the relative position and attitude of the cooperative target. Let the attitude rotation matrix from the target coordinate system to the measurement coordinate system be R TW , translation vector is t TW , then the pose solution error term of the i-th target point can be expressed as:

[0150] e i =P Wi -(R TW P Ti +t TW )

[0151] Where, P Ti ——theoretical position of target point i in the target coordinate system;

[0152] P Wi ——The measured position of target point i in the measurement coordinate system.

[0153] The cooperative target pose solution is to construct a least squares problem to minimize the sum of square errors of all target points. The objective function is:

[0154]

[0155] To solve the above formula, first define the centroid of the two sets of landmarks:

[0156]

[0157] The objective function can be simplified as:

[0158]

[0159] By simplifying the objective function, it becomes two independent optimization terms. The first term of the objective function is only related to the rotation matrix R. TW Related, the rotation matrix can be constructed to calculate the objective function as:

[0160]

[0161] Where Q Wi =P Wi -P W , Q Ti =P Ti -P T .

[0162] In the above formula, only the third term is related to the rotation matrix, so the objective function can be further optimized as:

[0163]

[0164] To solve the rotation matrix, define W in the objective function as:

[0165]

[0166] Performing SVD decomposition on W yields:

[0167] W=UΣV T

[0168] Where, Σ is a diagonal matrix composed of the singular values of the W matrix, and the diagonal elements are arranged from large to small;

[0169] U, V——diagonal matrices.

[0170] When W is of full rank, the rotation matrix can be expressed as:

[0171] R TW =UV T

[0172] After obtaining the rotation matrix, set the second term of the objective function to 0, and then calculate the translation vector as:

[0173] t TW =P W -R TW P T

[0174] In another exemplary embodiment, the following test is provided to test and verify the system in the above embodiment.

[0175] The space cooperative target posture visual measurement system in the embodiment of the present application was built in the laboratory, the robot arm guide rail simulated the relative motion, and the solar simulator was used to simulate the space lighting environment. The space target 6-DOF real-time high-precision posture measurement test was carried out in the laboratory. The posture measurement results of the system are shown in Figure 2. Figure 9 As shown, Figure 9 (a)-(f) are the measurement results of the space target's x-axis relative position, y-axis relative position, z-axis relative position, relative roll angle, relative pitch angle, and relative yaw angle, respectively. During the entire process, the system outputs a correct and stable trend in the measurement pose results.

[0176] The system adopts a visible light passive cooperative target binocular vision measurement solution, which has the following advantages compared with the currently commonly used spectral differential cooperative target measurement system:

[0177] (1) The cooperative target subsystem does not require additional spectral gating characteristic layers, and its on-orbit service life is longer.

[0178] (2) The visual measurement subsystem uses a conventional wide-band visible light binocular camera, which does not require active illumination of a specific spectral band. It has a simple hardware structure, high reliability, and flexible expansion capabilities.

[0179] (3) The system's binocular vision cooperative target pose measurement algorithm can achieve high-precision positioning of cooperative target landmarks in complex interference backgrounds, and realize rapid landmark recognition through target coding information detection and binocular cross-validation, with high robustness and high-precision pose measurement performance.

[0180] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0181] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A space cooperative target position visual measurement system, characterized by: include: A cooperative targeting subsystem and a visual measurement subsystem, wherein the cooperative targeting subsystem is arranged on the target aircraft, and the visual measurement subsystem is arranged on the service aircraft; The visual measurement subsystem includes a visible light binocular camera and an information processor; The visible light binocular camera is used to obtain a binocular measurement image of the cooperative target subsystem; The information processor is used to perform landmark extraction, target recognition and pose calculation on the binocular measurement image.

2. The space cooperative target pose visual measurement system according to claim 1, characterized in that: The target pattern of the cooperative target subsystem includes four coplanar A-type target components and one B-type target component, wherein three A-type target components and one B-type target component are distributed at the four vertices of a square, and another A-type target component is located at the center of the square.

3. The space cooperative target posture visual measurement system according to claim 1, characterized in that: The type A target assembly and the type B target assembly both include a black central circle and a plurality of black and white rings arranged outside the black central circle; wherein the ring closest to the black central circle is a white ring; The diameter of the black center circle of the B-type target assembly is larger than the diameter of the black center circle of the A-type target assembly. The circular ring of the A-type target assembly corresponds one-to-one to the circular ring of the B-type target assembly and has the same diameter.

4. The space cooperative target pose visual measurement system according to claim 1, characterized in that: The binocular measurement image includes a left camera measurement image and a right camera measurement image. In terms of extracting landmark points from the binocular measurement image, the information processor is specifically used to: Performing connected region extraction on the left camera measurement image and the right camera measurement image respectively to obtain a plurality of first connected regions in the left camera measurement image and a plurality of second connected regions in the right camera measurement image; Screening the plurality of first connected regions in the left camera measurement image and the plurality of second connected regions in the right camera measurement image respectively to determine a plurality of first candidate marker point connected regions in the left camera image and a plurality of second candidate marker point connected regions in the right camera image; Calculating characteristic parameters of the connected areas of the plurality of first candidate marker points and the connected areas of the plurality of second candidate marker points respectively to obtain characteristic parameters of the plurality of first candidate marker points and characteristic parameters of the plurality of second candidate marker points; The plurality of first candidate marker points and the plurality of second candidate marker points are matched according to the characteristic parameters of the plurality of first candidate marker points and the characteristic parameters of the plurality of second candidate marker points to obtain matching marker point pairs.

5. The space cooperative target posture visual measurement system according to claim 4, characterized in that: The screening of the plurality of first connected areas in the left camera measurement image and the plurality of second connected areas in the right camera measurement image to determine the plurality of first candidate marker point connected areas in the left camera image and the plurality of second candidate marker point connected areas in the right camera image specifically includes: Calculate the area, aspect ratio, concavity, area ratio and circularity of the target connected region; When the area, aspect ratio, concavity, area ratio and roundness of the target connected area meet the first preset condition, the target connected area is determined to be the candidate marker point connected area; the target connected area is the first connected area or the second connected area, when the target connected area is the first connected area, the candidate marker point connected area is the first candidate marker point connected area, and when the target connected area is the second connected area, the candidate marker point connected area is the second candidate marker point connected area.

6. The space cooperative target posture visual measurement system according to claim 5, characterized in that: The first preset condition is: Amin ≤A IPx ≤σ Amax &k abmin ≤k ab &d concavitymin ≤d concavity &k areamin ≤k area &d circularitymin ≤d circularity ; Among them, A IPx is the area of the target connected region, σ Amin and σ Amax are the minimum threshold for area screening and the maximum threshold for area screening, k ab is the width-to-length ratio of the target connected area, k abmin is the minimum threshold for aspect ratio screening, d concavity is the concavity of the target connected region, d concavitymin is the minimum threshold for concavity screening, k area is the area ratio of the target connected region, k areamin is the minimum threshold for area ratio screening, d circularity is the circularity of the target connected area, d circularitymin Filter the circularity threshold for the target connected region.

7. The space cooperative target posture visual measurement system according to claim 4, characterized in that: Matching the plurality of first candidate marker points with the plurality of second candidate marker points according to the characteristic parameters of the plurality of first candidate marker points and the characteristic parameters of the plurality of second candidate marker points to obtain matching marker point pairs specifically includes: Calculating the epipolar distance between the first target candidate landmark point and each second candidate landmark point; the first target candidate landmark point is any first candidate landmark point; Determine the second candidate landmark point with the smallest epipolar distance as the second target candidate landmark point; Calculating a radius deviation rate, an area deviation rate, and an area ratio deviation rate between the first target candidate marker point and the second target candidate marker point based on characteristic parameters of the first target candidate marker point and characteristic parameters of the second target candidate marker point; When the epipolar distance, the radius deviation rate, the area deviation rate, and the area ratio deviation rate meet the second preset condition, the first target candidate landmark point and the second target candidate landmark point are combined into a matching landmark point pair.

8. The space cooperative target position visual measurement system according to claim 7, characterized in that: The second preset condition is: d LA ≤d LA max &Δr T ≤Δ max &ΔA TPx ≤Δ max &Δk Tarea ≤Δ max ; Among them, d LA , Δr T , ΔA TPx , Δk Tarea are the epipolar distance, radius deviation rate, area deviation rate, and area ratio deviation rate between the first target candidate landmark point and the second target candidate landmark point, respectively. LAmax is the epipolar distance matching screening threshold, Δ m ax is the maximum deviation rate screening threshold of the characteristic parameter.

9. The space cooperative target posture visual measurement system according to claim 1, characterized in that: In the aspect of target identification, the information processor is specifically used for: The landmark points included in the matching landmark point pairs are used as target landmark points and form a target landmark point set; Calculating the three-dimensional spatial position and area ratio of each target marker point in the target marker point set; Determining a target marker point corresponding to the type B target component in the target marker point set according to the area ratio as a benchmark reference point, and obtaining position coding information of the benchmark reference point; Calculating reference distances of target marker points other than the benchmark reference point in the target marker point set based on the three-dimensional spatial position of the benchmark reference point and the three-dimensional spatial positions of the target marker points other than the benchmark reference point in the target marker point set; Determine the target marker point with the smallest reference distance as the first angle reference point, and obtain the position coding information of the first angle reference point; the first angle reference point is the target marker point of the A-type target assembly located at the center of the square; Determine the target marker point with the largest reference distance as the second angle reference point, and obtain the position coding information of the second angle reference point; the vertex of the square where the type A target component corresponding to the second angle reference point is located is located on the same diagonal line as the vertex where the type B target component is located; determining the first angle reference point or the second angle reference point as the target angle reference point; According to the three-dimensional spatial position of the reference reference point and the spatial position of the target angle reference point, a vector from the reference reference point to the target angle reference point is determined as a reference vector; According to the three-dimensional spatial position of the reference reference point and the spatial position of the target marker point to be identified, a vector from the reference reference point to the target marker point to be identified is determined as a reference vector; the target marker point to be identified is any target marker point in the target marker point set except the reference reference point, the first angle reference point and the second angle reference point; Calculating the angle between the reference vector and the reference vector as a reference angle of the target marker to be identified; The position coding information of the target marker to be identified is determined according to the reference angle.

10. The space cooperative target pose visual measurement system according to claim 1, characterized in that: In terms of pose calculation, the information processor is specifically used to: Define the objective function: Q Ti =P Ti -P T ; Q Wi =P Wi -P W ; Among them, W is the objective function, Q Ti is the relative position of target landmark point i and the centroid of all target points in the target coordinate system, Q Wi is the relative position of target marker point i and the centroid of all target points in the measurement coordinate system, n is the number of target marker points, and the superscript T indicates transposition; P Ti is the theoretical position of target marker point i in the target coordinate system, P T is the theoretical position of the center of mass of all target points in the target coordinate system, P Wi is the measured position of target point i in the measurement coordinate system, P W is the measured position of the centroid of all target points in the measurement coordinate system; Perform singular value decomposition on the objective function to obtain the singular value decomposition result: W = UΣV T , where Σ is the diagonal matrix composed of the singular values of W, U and V are the first and second diagonal matrices in the singular value decomposition results respectively; According to the singular value decomposition results, the rotation matrix is obtained: R TW =UV T , where R TW is the rotation matrix; According to the rotation matrix, the translation vector is determined as: TW =P W -R TW P T , where t TW is the translation vector; The position and posture of the target aircraft relative to the visible light binocular camera are determined according to the rotation matrix, the translation vector and the position and posture of the target aircraft in the binocular measurement image.

Images