Space target SFM three-dimensional reconstruction method based on multiple frames of ISAR images

Abstract

The invention is suitable for the technical field of radar signal processing, and provides a space target SFM three-dimensional reconstruction method based on a multi-frame ISAR image, and the method comprises the steps: carrying out the preprocessing and imaging of large-angle inverse synthetic aperture radar ISAR echo data of a space target, and obtaining a multi-frame ISAR image sequence; extracting scattering points in each frame of ISAR image sequence according to an orthogonal matching pursuit algorithm, and performing feature matching on adjacent frames of ISAR image sequences to obtain a scattering point set; obtaining instantaneous radar sight line information, and constructing a projection matrix from a space target scattering center three-dimensional coordinate to an ISAR two-dimensional pixel plane according to the instantaneous radar sight line information; and obtaining a three-dimensional reconstruction result according to the scattering point set and the projection matrix in combination with a triangulation principle in an incremental motion recovery structure SFM method. According to the method, the robustness and precision of three-dimensional reconstruction can be improved.

Description

Technical Field

[0001] This application belongs to the technical field of radar signal processing, and particularly relates to a method for three-dimensional reconstruction of spatial targets by SFM based on multi-frame ISAR images. Background Art

[0002] Inverse Synthetic Aperture Radar (ISAR) imaging technology is one of the important means to obtain information of spatial targets. By transmitting electromagnetic waves and receiving echoes, ISAR can generate high-resolution two-dimensional images, revealing the structural information of the target from the radar perspective. However, the two-dimensional images obtained by ISAR imaging only reflect the projection of the spatial target on the radar imaging plane and cannot directly provide the three-dimensional geometric structure of the target. Therefore, how to reconstruct the three-dimensional structure of the spatial target from multi-frame ISAR images has become an important research direction in the current field of ISAR imaging technology.

[0003] Existing ISAR three-dimensional reconstruction methods establish a projection matrix between the three-dimensional geometric model of the spatial target and the ISAR image, and fit the projection matrix through iterative optimization (such as particle swarm optimization, etc.) to obtain the three-dimensional structure of the spatial target. However, when the motion of the target has large uncertainties or is difficult to accurately model, the above methods cannot construct an accurate projection matrix, resulting in problems of reduced robustness and accuracy of the reconstruction results. Summary of the Invention

[0004] The embodiments of this application provide a method for three-dimensional reconstruction of spatial targets by SFM based on multi-frame ISAR images, which can solve the problems of reduced robustness and accuracy of the existing ISAR three-dimensional reconstruction methods.

[0005] In a first aspect, the embodiments of this application provide a method for three-dimensional reconstruction of spatial targets by SFM based on multi-frame ISAR images, including: Step 1, preprocessing and imaging the ISAR echo data of the spatial target with a large angle to obtain a multi-frame ISAR image sequence; Step 2, extracting scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm, and performing feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set; Step 3, obtaining the instantaneous radar line-of-sight information, and constructing a projection matrix from the three-dimensional coordinates of the spatial target scatter center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information; Step 4, obtaining the three-dimensional reconstruction result according to the scatter point set and the projection matrix, in combination with the triangulation principle in the incremental Structure from Motion (SFM) method.

[0006] In a possible implementation manner of the first aspect, the above Step 1 specifically includes:

[0007] Divide the ISAR echo data into multiple sub-aperture data;

[0008] Perform envelope alignment and autofocus processing on each sub-aperture data respectively to complete translational compensation;

[0009] Use the RD algorithm to image the ISAR echo data after translational compensation to obtain a multi-frame ISAR image sequence.

[0010] Optionally, in another possible implementation manner of the first aspect, extracting the scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm in step 2 above includes:

[0011] Set the input parameters of the orthogonal matching pursuit algorithm as the over-complete dictionary matrix observation vector and the sparsity K of the signal to be recovered;

[0012] Set the initial state of the orthogonal matching pursuit algorithm as: sparse signal x0 = 0, residual vector r0 = y - Ax0 = y, index set iteration number t = 1;

[0013] Calculate the current residual r t-1 and each column of the dictionary matrix A correlation, and select the index j of the dictionary element most relevant to the current residual t :

[0014]

[0015] Add the index j t to the index set Λ t :

[0016] Λ t = Λ t-1 ∪{j t}

[0017] On the current index set Λ t update the sparse coefficient x by the least squares method t :

[0018]

[0019] where, is the sub-matrix composed of the column vectors corresponding to the index set Λ t in the dictionary matrix A;

[0020] Update the residual vector r t :

[0021]

[0022] Determine whether the residual vector reaches the preset residual threshold or whether the number of scatter points reaches the sparsity K of the signal to be restored. If it reaches, stop the iteration; if not, repeat the above orthogonal matching pursuit algorithm;

[0023] Obtain the sparse signal to be estimated, and extract the scatter points in each frame of the ISAR image sequence according to the sparse signal to be estimated.

[0024] Optionally, in another possible implementation manner of the first aspect, the feature matching of the adjacent frame ISAR image sequences in step 2 above to obtain the scatter point set includes:

[0025] Extract the two-dimensional coordinates of the scatter points from each frame of the ISAR image;

[0026] Take the current frame of the ISAR image as the reference frame;

[0027] According to the projection matrix of each frame of the ISAR image, convert the scatter point coordinates of the adjacent frame image to the reference coordinate system of the current frame image;

[0028] For each scatter point in the current frame, calculate its distance from all scatter points in the adjacent frame;

[0029] Find the point in the adjacent frame that is closest to the scatter point in the current frame, and record its coordinates and distance;

[0030] Repeat the above nearest neighbor matching process until all scatter points in the current frame find corresponding matching points;

[0031] Select the majority of points with a distance less than the preset threshold to form the scatter point set after matching.

[0032] Optionally, in another possible implementation manner of the first aspect, the obtaining of the instantaneous radar line-of-sight information in step 3 above includes:

[0033] Establish a space target orbital plane coordinate system O-XYZ, where OZ points to the geocentric direction, the plane formed by OZ and the space target motion direction is called the orbital plane, OX is in the orbital plane and points to the space target motion direction, and the OY direction is determined by the right-hand rule;

[0034] Obtain the instantaneous radar line-of-sight information according to the established space target orbital plane coordinate system O-XYZ:

[0035] L LOS =[cosβ(t)sinα(t),cosβ(t)cosα(t),sinβ(t)] T

[0036] where α(t) and β(t) are the radar line-of-sight azimuth angle and elevation angle in the orbital plane coordinate system respectively, and t is the slow time during the observation period.

[0037] Optionally, in another possible implementation of the first aspect, constructing the projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information in step 3 above includes:

[0038] For the i-axis in the radar ray direction and the j-axis in the Doppler direction, the projection vectors i and j can be expressed as:

[0039]

[0040] where α(t0) and β(t0) are the radar line-of-sight observation angles at the initial time t0, and are the first-order derivatives of α(t) and β(t);

[0041] According to the projection vectors i and j, obtain the projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional imaging plane:

[0042]

[0043] Optionally, in another possible implementation of the first aspect, obtaining the three-dimensional reconstruction result according to the scattering point set and the projection matrix, in combination with the triangulation principle in the incremental structure from motion (SFM) method in step 4 above includes:

[0044] According to the projection matrix P from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane, obtain the transformation matrix M from the three-dimensional world coordinate system to the two-dimensional pixel coordinate system:

[0045]

[0046] where H and W represent the height and width of the ISAR image, Δr = c / 2B and Δf a = λ / 2W are the range resolution and azimuth resolution respectively, c is the speed of light, B is the radar bandwidth, λ is the wavelength, and W is the rotation angle of the radar ray during the observation period;

[0047] According to the transformation matrices of the two images and the corresponding matching scattering point sets, construct a constraint equation using the triangulation principle:

[0048]

[0049] where x i = [u i , v i , 1] T represents the coordinates of the scattering point matched in the i-th image, and M iThe transformation matrix representing the i-th image, which is a 3-row and 4-column matrix, and X = [x, y, z, 1] T Represents the coordinates of a point in three-dimensional space;

[0050] Let

[0051]

[0052] where M ij Represents the j-th row of the transformation matrix corresponding to the i-th image, then the above formula can be expressed as:

[0053]

[0054] Cross-multiply both sides of the above equation by x i , to obtain:

[0055]

[0056] The reconstructed three-dimensional point coordinates X can be obtained using the SVD algorithm.

[0057] A method for three-dimensional reconstruction of a spatial target SFM based on multiple-frame ISAR images provided by this application preprocesses and images the inverse synthetic aperture radar (ISAR) echo data of the spatial target at large angles to obtain a sequence of multiple-frame ISAR images; extracts scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm, and performs feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set; obtains the instantaneous radar line-of-sight information, and constructs a projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information; according to the scatter point set and the projection matrix, combines the triangulation principle in the incremental structure from motion (SFM) method to obtain the three-dimensional reconstruction result. Through the above method, the robustness and accuracy of the three-dimensional reconstruction can be improved. Brief Description of the Drawings

[0058] In order to more clearly illustrate the technical solutions in the embodiments of this application, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings in the following description are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0059] Figure 1 It is a schematic flow chart of the method for three-dimensional reconstruction of a spatial target SFM based on multiple-frame ISAR images provided in an embodiment of this application;

[0060] Figure 2 It is a result diagram of extracting scatter points using the OMP algorithm provided in an embodiment of this application;

[0061] Figure 3 The result graph of matching scatter points by the nearest neighbor method provided by an embodiment of this application;

[0062] Figure 4 The result graph of three-dimensional reconstruction of space targets by the traditional factorization algorithm and the algorithm proposed by the present invention provided by an embodiment of this application. Detailed implementation manners

[0063] In the following description, specific details such as specific system structures and technologies are presented for the purpose of illustration rather than limitation, so as to thoroughly understand the embodiments of this application. However, those skilled in the art should clearly understand that this application can also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid unnecessary details from interfering with the description of this application.

[0064] It should be understood that when used in the specification and appended claims of this application, the term "comprising" indicates the presence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations.

[0065] It should also be understood that the term "and / or" as used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0066] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when", "once", "in response to determining", or "in response to detecting" according to the context. Similarly, the phrase "if determined" or "if [the described condition or event] is detected" can be interpreted as meaning "once determined", "in response to determining", "once [the described condition or event] is detected", or "in response to detecting [the described condition or event]" according to the context.

[0067] In addition, in the description of the specification and appended claims of this application, the terms "first", "second", "third", etc. are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance.

[0068] References to "one embodiment" or "some embodiments" etc. described in the specification of the present application mean that specific features, structures or characteristics described in connection with that embodiment are included in one or more embodiments of the present application. Thus, statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments" etc. that appear in different places in this specification do not necessarily all refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized. The terms "comprising", "including", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized.

[0069] The method for three-dimensional reconstruction of a spatial target SFM based on multi-frame ISAR images provided by the present application will be described in detail below with reference to the accompanying drawings.

[0070] Figure 1 The flowchart of a method for three-dimensional reconstruction of a spatial target SFM based on multi-frame ISAR images provided by an embodiment of the present application is shown.

[0071] As Figure 1 shown, the method for three-dimensional reconstruction of a spatial target SFM based on multi-frame ISAR images includes the following steps:

[0072] Step 101: Preprocess and image the ISAR echo data of the spatial target at a large angle to obtain a multi-frame ISAR image sequence.

[0073] Among them, the ISAR echo data may refer to forming a two-dimensional high-resolution image through the relative displacement between the radar and the moving target, and the echo data includes: range direction (target size) and azimuth direction (motion characteristics) information.

[0074] Further, in a possible implementation manner of an embodiment of the present application, the above step 101 includes:

[0075] Divide the ISAR echo data into multiple sub-aperture data;

[0076] Perform envelope alignment and autofocus processing on each sub-aperture data respectively to complete translational compensation;

[0077] Use the RD algorithm to image the ISAR echo data after translational compensation to obtain a multi-frame ISAR image sequence.

[0078] Among them, the RD algorithm is one of the core algorithms in synthetic aperture radar processing and is used to generate high-resolution radar images. It realizes high-resolution imaging through two-dimensional processing: range compression and azimuth compression, and by using pulse compression technology and Doppler frequency shift analysis.

[0079] In the embodiments of the present application, preprocessing and imaging obtain a multi-frame ISAR image sequence, thereby improving the image quality, ensuring the accuracy of subsequent three-dimensional reconstruction, and providing a data basis.

[0080] Step 102: Extract the scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm, and perform feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set.

[0081] Among them, the orthogonal matching pursuit algorithm may refer to a sparse signal reconstruction algorithm based on greedy iteration. By successively selecting the atom (i.e., the basis vector in the dictionary matrix) most relevant to the current residual and using orthogonal projection to update the support set, the sparse representation of the signal under the over-complete dictionary is realized.

[0082] Furthermore, in a possible implementation manner of the embodiments of the present application, extracting the scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm in the above step 102 includes:

[0083] Set the input parameters of the orthogonal matching pursuit algorithm as the over-complete dictionary matrix observation vector and the sparsity K of the signal to be restored;

[0084] Set the initial state of the orthogonal matching pursuit algorithm as: sparse signal x0 = 0, residual vector r0 = y - Ax0 = y, index set iteration times t = 1;

[0085] Calculate the current residual r t-1 and each column of the dictionary matrix A correlation, and select the index j of the dictionary element most relevant to the current residual t :

[0086]

[0087] Add the index j t to the index set Λ t :

[0088] Λ t = Λ t-1 ∪{j t}

[0089] On the current index set Λ t update the sparse coefficient x by the least squares method t :

[0090]

[0091] Among them, is the index set Λ in the dictionary matrix A tA sub-matrix composed of corresponding column vectors;

[0092] Update the residual vector r t :

[0093]

[0094] Determine whether the residual vector reaches the preset residual threshold or the number of scattering points reaches the sparsity K of the signal to be restored. If it reaches, stop the iteration. If it does not reach, repeat the above orthogonal matching pursuit algorithm;

[0095] Obtain the sparse signal to be estimated, and extract the scattering points in each frame of the ISAR image sequence according to the sparse signal to be estimated.

[0096] Furthermore, in a possible implementation manner of the embodiment of the present application, the feature matching of adjacent frames of the ISAR image sequence in step 102 above to obtain a scattering point set includes:

[0097] Extract the two-dimensional coordinates of the scattering points from each frame of the ISAR image;

[0098] Take the current frame of the ISAR image as the reference frame;

[0099] According to the projection matrix of each frame of the ISAR image, convert the scattering point coordinates of the adjacent frame image to the reference coordinate system of the current frame image;

[0100] For each scattering point in the current frame, calculate its distance from all scattering points in the adjacent frame;

[0101] Find the point in the adjacent frame that is closest to the scattering point in the current frame, and record its coordinates and distance;

[0102] Repeat the above nearest neighbor matching process until all scattering points in the current frame find corresponding matching points;

[0103] Select the majority of points with a distance less than the preset threshold to form a set of matched scattering points.

[0104] In the embodiment of the present application, strong scattering centers are accurately extracted from each frame of the ISAR image through the orthogonal matching pursuit algorithm, and its sparse reconstruction characteristic is used to suppress noise interference and retain weak scattering points; then, based on the motion consistency constraint, cross-frame correlation matching is performed on the scattering points of adjacent frames to solve the problem of feature breakage caused by target pose changes or occlusions. Finally, a spatio-temporally aligned set of scattering points is generated, providing high-confidence input data for subsequent three-dimensional projection calculation and avoiding three-dimensional structure distortion caused by false detection / missing detection in the traditional threshold segmentation method.

[0105] Step 103: Obtain the instantaneous radar line-of-sight information, and construct a projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane based on the instantaneous radar line-of-sight information.

[0106] Further, in a possible implementation manner of the embodiment of the present application, obtaining the instantaneous radar line-of-sight information in the above step 103 includes:

[0107] Establish a spatial target orbital plane coordinate system O-XYZ, where OZ points to the direction of the earth's center. The plane formed by OZ and the moving direction of the spatial target is called the orbital plane. OX is in the orbital plane and points to the moving direction of the spatial target, and the direction of OY is determined by the right-hand rule;

[0108] Obtain the instantaneous radar line-of-sight information according to the established spatial target orbital plane coordinate system O-XYZ:

[0109] L LOS = [cosβ(t)sinα(t), cosβ(t)cosα(t), sinβ(t)] T

[0110] where α(t) and β(t) are the azimuth angle and elevation angle of the radar line of sight in the orbital plane coordinate system respectively, and t is the slow time during the observation period.

[0111] Further, in a possible implementation manner of the embodiment of the present application, constructing a projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane in the above step 103 includes:

[0112] For the i-axis in the radar ray direction and the j-axis in the Doppler direction, the projection vectors i and j can be expressed as:

[0113]

[0114] where α(t0) and β(t0) are the radar line-of-sight observation angles at the initial time t0, and are the first-order derivatives of α(t) and β(t);

[0115] According to the projection vectors i and j, obtain the projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional imaging plane:

[0116]

[0117] Step 104: According to the scatter point set and the projection matrix, and combining the triangulation principle in the incremental structure from motion (SFM) method, obtain the three-dimensional reconstruction result.

[0118] Among them, the triangulation principle may refer to the principle based on multi-view geometric intersection. By matching the two-dimensional projection coordinates of feature points in at least two frames of images and the corresponding projection matrices of the cameras (constituted by pose parameters), an overdetermined system of equations is constructed to solve the three-dimensional spatial coordinates of the target points.

[0119] Further, in a possible implementation manner of the embodiment of the present application, the above step 104 includes:

[0120] According to the projection matrix P from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane, the transformation matrix M from the three-dimensional world coordinate system to the two-dimensional pixel coordinate system is obtained:

[0121]

[0122] Where H and W represent the height and width of the ISAR image, Δr = c / 2B and Δf a = λ / 2W are the range resolution and azimuth resolution respectively, c is the speed of light, B is the radar bandwidth, λ is the wavelength, and W is the rotation angle of the radar ray during the observation period;

[0123] According to the transformation matrices of the two images and the corresponding matching scatter point sets, constraint equations are constructed using the triangulation principle:

[0124]

[0125] Where x i = [u i , v i , 1] T represents the coordinates of the scatter points matched in the i-th image, M i represents the transformation matrix of the i-th image, which is a 3-row and 4-column matrix, and X = [x, y, z, 1] T represents the coordinates of a point in three-dimensional space;

[0126] Let

[0127]

[0128] Where M ij represents the j-th row of the transformation matrix corresponding to the i-th image, then the above formula can be expressed as:

[0129]

[0130] Multiply both sides of the above equation by x i simultaneously to obtain:

[0131]

[0132] The reconstructed three-dimensional point coordinates X can be obtained using the SVD algorithm.

[0133] The method for three-dimensional reconstruction of space target SFM based on multi-frame ISAR images provided by this application preprocesses and images the inverse synthetic aperture radar (ISAR) echo data of a space target at a large angle to obtain a multi-frame ISAR image sequence; extracts scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm, and performs feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set; obtains the instantaneous radar line-of-sight information, and constructs a projection matrix from the three-dimensional coordinates of the space target scatter center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information; and obtains the three-dimensional reconstruction result according to the scatter point set and the projection matrix, in combination with the triangulation principle in the incremental structure from motion (SFM) method. Through the above method, the robustness and accuracy of three-dimensional reconstruction can be improved.

[0134] To verify the beneficial effects of the present invention, the following experiments are carried out:

[0135] Use STK to simulate a low-earth orbit satellite. The orbital altitude of the target satellite is 300 km. Use the satellite orbital data, combine with the Jason-3 satellite model for electromagnetic simulation, use the range-Doppler algorithm for high-resolution ISAR imaging, divide it into 16 frames of data, use different algorithms for three-dimensional reconstruction, and compare the reconstruction quality.

[0136] The experimental results are as follows: Refer to the schematic Figure 2 , Figure 2 is the result of extracting the scatter points of the ISAR images of frames 6, 8, and 10 using the OMP algorithm. Refer to Figure 3 , Figure 3 is the scatter point matching result of the ISAR images of frames 6, 8, and 10. Refer to Figure 4 , Figure 4 is a comparison chart of the reconstruction results of the factorization algorithm and the algorithm of the present invention. The present invention can clearly reconstruct the satellite structure, the structure is more obvious, the points are denser, and the reconstruction effect is more accurate.

[0137] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of this application.

[0138] The above embodiments are only used to illustrate the technical solutions of this application, rather than to limit them; although this application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included in the protection scope of this application.

Claims

1. A 3D reconstruction method for space target SFM based on multi-frame ISAR images, characterized in that, It includes the following steps: Step 1: Preprocess and image the inverse synthetic aperture radar (ISAR) echo data of the space target at a large angle to obtain a multi-frame ISAR image sequence; Step 2: Extract the scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm, and perform feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set; Step 3: Obtain the instantaneous radar line-of-sight information, and construct a projection matrix from the three-dimensional coordinates of the space target scatter center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information; Step 4: According to the scatter point set and the projection matrix, and combining the triangulation principle in the incremental structure from motion (SFM) method, obtain the three-dimensional reconstruction result.

2. The method for three-dimensional reconstruction of a space target SFM based on multi-frame ISAR images according to claim 1, wherein The said Step 1 includes: Divide the ISAR echo data into multiple sub-aperture data; Perform envelope alignment and autofocus processing on each sub-aperture data respectively to complete translational compensation; Use the range-doppler (RD) algorithm to image the ISAR echo data after translational compensation to obtain a multi-frame ISAR image sequence.

3. The method for three-dimensional reconstruction of a space target SFM based on multi-frame ISAR images according to claim 1, characterized in that, The extracting of the scatter points in each frame of the ISAR image sequence according to the orthogonal matching pursuit algorithm in the said Step 2 includes: Set the input parameters of the orthogonal matching pursuit algorithm as the over-complete dictionary matrix Observation vector and the sparsity K of the signal to be recovered; Set the initial state of the orthogonal matching pursuit algorithm as: the sparse signal \(x_0 = 0\), the residual vector \(r_0=y - Ax_0=y\), and the index set The number of iterations \(t = 1\); Calculate the current residual r t-1 and each column of the dictionary matrix A for correlation, and select the index j of the dictionary element that is most correlated with the current residual t : Add index j t to the index set Λ t : Λ t = Λ t-1 ∪ {j t} On the current index set Λ t update the sparse coefficient x by the least squares method t : Among them, is the submatrix composed of the column vectors corresponding to the index set Λ t in the dictionary matrix A; Update the residual vector r t : Judge whether the residual vector reaches the preset residual threshold or whether the number of scatter points reaches the sparsity K of the signal to be recovered. If it reaches, stop the iteration; if not, repeat the above orthogonal matching pursuit algorithm; Obtain the sparse signal to be estimated, and extract the scatter points in each frame of the ISAR image sequence according to the sparse signal to be estimated.

4. The method for three-dimensional reconstruction of a spatial target SFM based on multi-frame ISAR images according to claim 1, wherein, The performing of feature matching on adjacent frames of the ISAR image sequence to obtain a scatter point set in the said Step 2 includes: Extract the two-dimensional coordinates of the scatter points from each frame of the ISAR image; Take the current frame of the ISAR image as the reference frame; According to the projection matrix of each frame of the ISAR image, convert the scatter point coordinates of the adjacent frame image to the reference coordinate system of the current frame image; For each scatter point in the current frame, calculate its distance from all scatter points in the adjacent frame; Find the point in the adjacent frame that is closest to the scatter point in the current frame, and record its coordinates and distance; Repeat the above nearest neighbor matching process until all scatter points in the current frame find corresponding matching points; Select the majority of points with a distance less than the preset threshold to form a scatter point set after matching.

5. The method for three-dimensional reconstruction of a space target SFM based on multi-frame ISAR images according to claim 1, characterized in that The obtaining of the instantaneous radar line-of-sight information in the said Step 3 includes: Establish a space target orbital plane coordinate system O-XYZ, where OZ points to the geocentric direction, the plane formed by OZ and the space target motion direction is called the orbital plane, OX is in the orbital plane and points to the space target motion direction, and the OY direction is determined by the right-hand rule; Obtain the instantaneous radar line-of-sight information according to the established space target orbital plane coordinate system O-XYZ: L LOS = [cosβ(t)sinα(t), cosβ(t)cosα(t), sinβ(t)] T where α(t) and β(t) are the azimuth angle and elevation angle of the radar line of sight in the orbital plane coordinate system respectively, and t is the slow time during the observation period.

6. The method for three-dimensional reconstruction of a spatial target SFM based on multi-frame ISAR images according to claim 5, wherein The constructing of the projection matrix from the three-dimensional coordinates of the space target scatter center to the ISAR two-dimensional pixel plane according to the instantaneous radar line-of-sight information in the said Step 3 includes: For the i-axis in the radar ray direction and the j-axis in the Doppler direction, the projection vectors i and j can be expressed as: where α(t0) and β(t0) are the radar line-of-sight observation angles at the initial time t0, and are the first-order derivatives of α(t) and β(t); According to the projection vectors i and j, obtain the projection matrix from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional imaging plane:

7. The 3D reconstruction method of the space target SFM based on multi-frame ISAR images according to claim 1, characterized in that The obtaining of the three-dimensional reconstruction result according to the scattering point set and the projection matrix in step 4, in combination with the triangulation principle in the incremental structure from motion (SFM) method, includes: According to the projection matrix P from the three-dimensional coordinates of the spatial target scattering center to the ISAR two-dimensional pixel plane, obtain the transformation matrix M from the three-dimensional world coordinate system to the two-dimensional pixel coordinate system: where H and W represent the height and width of the ISAR image, Δr = c / 2B and Δf a = λ / 2W are the range resolution and azimuth resolution respectively, c is the speed of light, B is the radar bandwidth, λ is the wavelength, and W is the rotation angle of the radar beam during the observation period; According to the transformation matrices of two images and the corresponding matching scattering point sets, construct a constraint equation using the triangulation principle: where x i = [u i , v i , 1] T represents the coordinates of the scattering points matched by the i-th image, M i represents the transformation matrix of the i-th image, which is a 3-row and 4-column matrix, X = [x, y, z, 1] T represents the coordinates of a point in three-dimensional space; Let where M ij represents the j-th row of the transformation matrix corresponding to the i-th image, and the above formula can be expressed as: Cross multiply both sides of the above equation by x i , to obtain: The reconstructed three-dimensional point coordinates X can be obtained using the SVD algorithm.

Images