Space target combined imaging and attitude inversion method based on satellite-borne platform

Abstract

The invention discloses a space target combined imaging and attitude inversion method based on a spaceborne platform, and relates to the field of inverse synthetic aperture radar imaging, in particular to the space target combined imaging and attitude inversion method based on the spaceborne platform. The objective of the invention is to solve the problem that high-precision imaging is difficult to realize due to three-dimensional high-order space-variant range migration and phase error caused by complex relative motion between a target and a platform in the process of space target imaging by the existing satellite-borne platform; and an existing attitude estimation algorithm needs to realize attitude inversion on the premise of rectangular part segmentation and complete orbit error compensation. The method comprises the following steps: 1, a narrow-band tracking system of a radar obtains a target slant distance and a radar line-of-sight unit vector; 2, obtaining a signal after wedge transformation; 3, acquiring spatial Doppler parameter distribution of the dominant display point; 4, obtaining an imaging result; and 5, completing accurate estimation of target attitude parameters through plane coefficient fitting.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of inverse synthetic aperture radar imaging (ISAR), in particular to a joint imaging and attitude inversion method for space targets based on a spaceborne platform. BACKGROUND

[0002] The imaging of space targets has always been a hot spot in the field of ISAR imaging. Due to atmospheric interference and orbital occlusion, traditional ground-based radars and optical systems face inherent limitations in tracking space targets. Spaceborne inverse synthetic aperture radar (ISAR) has become a key technology for describing the characteristics of space targets due to its ability to provide all-weather, long-time observation. In order to achieve more comprehensive situation awareness of space targets, high-precision imaging and attitude inversion are required. Existing high-precision imaging algorithms do not analyze the imaging characteristics of space targets, and the mapping relationship between the space target scattering points and the Doppler parameters needs further analysis. Meanwhile, the attitude inversion algorithm relies on rectangular component segmentation, and in the subsequent planar coefficient estimation, the orbital errors and the complex Doppler effects caused by high imaging resolution are often ignored. The present application proposes a joint high-resolution imaging and attitude inversion algorithm for high-dynamic Doppler space targets by estimating the spatial distribution of Doppler parameters, regionalized uniform motion compensation, and directly fitting the planar coefficients to invert the target attitude using principal component analysis. SUMMARY

[0003] The purpose of the present application is to solve the problems of existing spaceborne platforms in the imaging process of space targets, such as the three-dimensional high-order space-varying range migration and phase errors caused by complex relative motion between the target and the platform, which makes it difficult to achieve high-precision imaging, and the existing attitude estimation algorithm needs to be completely segmented into rectangular components and compensated for orbital errors before attitude inversion can be achieved. Therefore, the present application proposes a joint imaging and attitude inversion method for space targets based on a spaceborne platform.

[0004] The joint imaging and attitude inversion method for space targets based on a spaceborne platform includes the following specific process:

[0005] Step one: the narrowband tracking system of the radar obtains the target slant range and the radar line-of-sight unit vector;

[0006] Step two: based on the target slant range in step one, the received signal is compensated for translation to obtain a translation-compensated signal, and the translation-compensated signal is subjected to wedgelet transformation to obtain a wedgelet-transformed signal;

[0007] Step three: the wedgelet-transformed signal obtained in step two is subjected to time clipping to obtain a clipped signal, and the clipped signal is subjected to imaging processing to obtain a coarse focusing image 1;

[0008] The spatial Doppler parameter distribution of the highlight point is obtained by using the adaptive template matching method;

[0009] Step four: the spatial Doppler parameter distribution of the highlight point obtained in step three is interpolated to obtain the complete spatial Doppler parameter distribution on the image projection plane;

[0010] Based on the complete spatial Doppler parameter distribution, the real-time phase change of each region of the image is obtained;

[0011] The wedge-shaped transformed signal obtained in step two is imaged to obtain a coarse focusing image 2, and the coarse focusing image 2 is divided into sub-regions based on the real-time phase change to obtain the coordinate range of each region in the image domain;

[0012] Based on the coordinate range of each region, the corresponding local target signal of each region is obtained and compensated to obtain the compensated local target signal, and the compensated complete target signal is obtained based on the compensated local target signal of each region;

[0013] Based on the compensated complete target signal, the imaging result is obtained;

[0014] Step five: the radar line-of-sight unit vector Taylor expansion result obtained in step one, the spatial Doppler parameter distribution of the highlight point obtained in step three, and the imaging result obtained in step four are used to complete the accurate estimation of the target attitude parameter through plane coefficient fitting.

[0015] Preferably, the narrowband tracking system of the radar obtains the target slant range and the radar line-of-sight unit vector in step one; the specific process is as follows:

[0016] According to the narrowband tracking system of the radar

[0017] The narrowband tracking system of the radar obtains the target slant range R T (t m ), the radar line-of-sight elevation angle azimuth angle η l (t m );

[0018] Where t m is the azimuth time domain time, i.e. the azimuth dimension;

[0019] t m = mT r , T r represents the pulse repetition period, M is the number of accumulated pulses, and m = 0, 1, …, M-1;

[0020] The target slant range is represented by Taylor expansion as follows:

[0021] The target slant range is represented by Taylor expansion as follows:

[0022] wherein,

[0023] T0is the 0th Taylor series expansion coefficient of R T (t m );

[0024] T1is the 1st Taylor series expansion coefficient of R T (t m );

[0025] T2is the 2nd Taylor series expansion coefficient of R T (t m );

[0026] O T (t m ) represents a negligible higher order term;

[0027] The radar line-of-sight unit vector i los (t m ) is represented as:

[0028]

[0029] The radar line-of-sight unit vector i los (t m ) is Taylor expanded to obtain:

[0030]

[0031] wherein,

[0032] i0is the 0th Taylor series expansion coefficient of i los (t m );

[0033] i1is the 1st Taylor series expansion coefficient of i los (t m );

[0034] i2is the 2nd Taylor series expansion coefficient of i los (t m );

[0035] O los (t m ) represents a negligible higher order term.

[0036] Preferably, the step two is based on the target slant range in step one to compensate the received signal for the plane motion, to obtain a plane compensation signal, and to perform a wedge transform on the plane compensation signal to obtain a wedge transformed signal; the specific process is as follows:

[0037] Step two one:

[0038] The ISAR system transmits a linear frequency modulation signal, pulse compresses the linear frequency modulation signal, and obtains a pulse compressed target echo spectrum s(f r ,t m ), the pulse compressed target echo spectrum s(f r ,t m ) is expressed as follows:

[0039]

[0040] wherein,

[0041] N represents the number of scattering points on the target;

[0042] σ p represents the amplitude of the linear frequency modulation signal transmitted by the ISAR system;

[0043] f r represents the frequency of the range dimension;

[0044] B represents the bandwidth of the linear frequency modulation signal transmitted by the ISAR system;

[0045] f c represents the carrier frequency;

[0046] λ represents the radar wavelength;

[0047] rect(·) represents a rectangular window;

[0048] j represents an imaginary unit, j 2 =-1;

[0049] R p (t m ) represents the range migration of the scattering point p relative to the radar, and is expressed as:

[0050] R p (t m ) = R T (t m ) + R Rot,p (t m )

[0051] R Rot,p (t m ) = i los (t m ) · p T

[0052] wherein,

[0053] R Rot,p (t m ) represents the rotational range migration of the scattering point p relative to the radar;

[0054] • represents the vector product;

[0055] p represents the three-dimensional coordinate vector of the scattering point p of the target;

[0056] The upper index T represents the transpose;

[0057] Step two: constructing the compensation term H(f r , t m ) based on the Taylor series expansion coefficient of the target slant range R T (t m ) in step one;

[0058] Step two: multiplying the compensation term H(f r , t m ) with the pulse-compressed target echo spectrum s(f r , t m ) to obtain the signal s'(f r , t m ) after the compensation of the translational motion;

[0059] Step two: performing the wedge transform on the signal s'(f r , t m ) after the compensation of the translational motion to obtain the signal s'(f r , τ m ) after the wedge transform.

[0060] Preferably, the step two constructs the compensation term H(f T , t m ) based on the Taylor series expansion coefficient of the target slant range R r (t m ) in step one; the specific process is as follows:

[0061]

[0062] wherein,

[0063] c represents the speed of light.

[0064] Preferably, the step two multiplies the compensation term H(f r , t m ) with the pulse-compressed target echo spectrum s(f r , t m ) to obtain the signal s'(f r , t m ) after the compensation of the translational motion; which is expressed as:

[0065] s'(f r , t m ) = s(f r , t m ) × H(f r , tm ).

[0066] Preferably, in step 24, the signal s'(f r ,t m ) is wedge-shaped transformed to obtain the wedge-shaped transformed signal s'(f r ,τ m );

[0067] in,

[0068] τ m is the time component after wedge transformation, τ m =(f r +f c ) / f c t m .

[0069] Preferably, in the step 3, the wedge-shaped transformed signal obtained in the step 2 is time-intercepted to obtain an intercepted signal, and imaging processing is performed on the intercepted signal to obtain a coarse focus image 1;

[0070] Using the adaptive template matching method, the prominent points on the coarse focus image 1 are extracted and the spatial Doppler parameter distribution of the prominent points is obtained;

[0071] The specific process is:

[0072] Step 31:

[0073] The criterion is that the second-order distance migration within the interception time period is less than c / 2B;

[0074] According to the criterion, the signal s'(f r ,τ m ) intercept the first time period and obtain the intercepted signal s' p (f r ,τ m );

[0075] Step 32:

[0076] For the intercepted signal s' p (f r ,τ m ) Perform inverse Fourier transform of the distance dimension and Fourier transform of the azimuth dimension in sequence to obtain a coarse focus image 1S' p (t r ,f m );

[0077] Among them, t r represents the time of the distance dimension, f m represents the frequency of the azimuthal dimension;

[0078] Step 33:

[0079] extracting the salient points in the coarse focus image IS' p r m obtaining spatial Doppler parameter distribution of all salient points;

[0080] The spatial Doppler parameter distribution of all salient points is the zeroth and first order Doppler parameter set {(D 0,p ,D 1,p}, p = 1, 2,..., N L of all extracted salient points;

[0081] N L is the number of extracted salient points;

[0082] D 0,p is the zeroth order Doppler parameter set of all extracted salient points;

[0083] D 1,p is the first order Doppler parameter set of all extracted salient points;

[0084] Step three four:

[0085] extracting the azimuth dimension signal corresponding to each salient point extracted in step three three through an image domain filter; the specific process is:

[0086] extracting the azimuth dimension signal corresponding to the pth salient point extracted in step three three through an image domain filter, denoted as s' p (τ m );

[0087] Step three five:

[0088] estimating the frequency modulation k p of s' m (τ 2,p ) using ICPF;

[0089] obtaining the second order Doppler parameter D 2,p corresponding to the salient point p based on the frequency modulation k 2,p of s' p (τ m ); denoted as:

[0090]

[0091] wherein f r,p is the fast time frequency corresponding to the distance dimension unit where the salient point p is located,

[0092] wherein T l ​​represents the pulse width of the transmitted chirp signal of the ISAR system;

[0093] Step 36: repeat steps 34 to 35 to obtain the second-order Doppler parameters D 2,p , p = 1, 2,..., N L ;

[0094] Step 37: based on the zero-order and first-order Doppler parameter sets {(D 0,p , D 1,p ), p = 1, 2,..., N L extracted in step 33 and the second-order Doppler parameters D 2,p , p = 1, 2,..., N L obtained in step 36, obtain the spatial Doppler parameter set {(D 0,p , D 1,p , D 2,p ), p = 1, 2,..., N L .

[0095] Preferably, the spatial Doppler parameter distribution of the feature points obtained in step 33 is interpolated in step 4 to obtain the complete spatial Doppler parameter distribution on the image projection plane;

[0096] Based on the complete spatial Doppler parameter distribution, the real-time phase change of each region of the image is obtained;

[0097] The wedge-transformed signal obtained in step 2 is imaged to obtain a coarse focusing image 2, and the coarse focusing image 2 is divided into sub-regions based on the real-time phase change to obtain the coordinate range of each region in the image domain;

[0098] Based on the coordinate range of each region, the local target signal corresponding to each region is obtained and compensated to obtain the compensated local target signal, and the compensated complete target signal is obtained based on the compensated local target signal of each region;

[0099] Based on the compensated complete target signal, the imaging result is obtained;

[0100] The specific process is as follows:

[0101] Step 41:

[0102] The second-order Doppler parameters D 2,p , p = 1, 2,..., N L corresponding to all feature points obtained in step 3 are interpolated based on the natural neighbor interpolation method of triangulation to obtain the interpolated second-order Doppler parameters D2;

[0103] Step four two: based on the interpolated second order Doppler parameter D2, obtain real-time phase change is expressed as:

[0104]

[0105] Step four three: construct the coarse focusing image 2S'(t r ,f m ); the specific process is:

[0106] The wedge-shaped transformed signal s'(f r ,τ m ) obtained in step two is sequentially subjected to IFT in the distance dimension and FT in the azimuth dimension, to obtain the coarse focusing image 2S'(t r ,f m );

[0107] Step four four: when the real-time phase change is less than π / 4, the coarse focusing image 2S'(t r ,f m ) is divided into Q sub-regions according to the adaptive partitioning principle;

[0108] The image corresponding to the qth sub-region is expressed as S' q (t r ,f m );

[0109] S' q (t r ,f m ) is sequentially subjected to FT in the distance dimension and IFT in the azimuth dimension, to obtain the signal s' q (f r ,t m ) corresponding to the qth sub-region;

[0110] Step four five: construct the compensation term h q (f r ,τ m ) of the signal corresponding to the qth sub-region; it is expressed as:

[0111]

[0112] wherein, is the average value of the second order Doppler parameter in the qth sub-region;

[0113] Step four six:

[0114] The compensation term h q (f r ,τ m ) of the signal corresponding to the qth sub-region constructed in step four five is combined with the signal s'q (f r ,t m ) multiplication, to obtain the compensated signal of the qth sub-region;

[0115] Step four seven:

[0116] Repeat steps four five to four six to obtain the complete target signal after compensation of all sub-regions:

[0117]

[0118] Step four eight:

[0119] Perform IFT in the distance dimension and FT in the azimuth dimension on the complete target signal s t (f r ,τ m ) in turn to obtain the imaging result of the target.

[0120] Preferably, the step five utilizes the radar line-of-sight unit vector Taylor expansion result obtained in step one, the spatial Doppler parameter distribution result of the characteristic point obtained in step three, and the imaging result obtained in step four to complete accurate estimation of the target attitude parameter through plane coefficient fitting; the specific process is as follows:

[0121] Step five one:

[0122] The Doppler parameter distribution result {(D 0,p ,D 1,p ,D 2,p )}, p = 1, 2,..., N L Generate point cloud data D, and the three-dimensional coordinates of the point cloud data D are D 0,p , p = 1, 2,..., N L , D 1,p , p = 1, 2,..., N L , D 2,p , p = 1, 2,..., N L ;

[0123] D 0,p is the x-axis coordinate, D 1,p is the y-axis coordinate, and D 2,p is the z-axis coordinate;

[0124] Step five two: let the iteration number i = 1;

[0125] Step five three: perform plane fitting on the point cloud data D by using principal component analysis to obtain the fitting plane and plane coefficients C 0,i , C 1,i , 1, C bias,i of the i-th cycle;

[0126] wherein,

[0127] C 0,i is the plane coefficient of D 0,p ;

[0128] C 1,i is the plane coefficient of D 1,p ;

[0129] 1 is the plane coefficient of D 2,p ;

[0130] C bias,i is the plane coefficient constant term;

[0131] Step five four: calculate the Euclidean distance d L of the N p feature points obtained in step three to the fitting plane p = 1, 2,..., N L ; represented as:

[0132]

[0133] Step five five: calculate the standard deviation σ of the Euclidean distance; represented as:

[0134]

[0135] wherein, is the mean of the Euclidean distance, N i is the number of feature points in the i-th loop;

[0136] Step five six: retain the feature points d p < σ, let the iteration number i = i + 1, and update the number of feature points and the point cloud data, repeat steps five three to five five, until the number of remaining feature points is less than three or the Euclidean distance of each retained feature point satisfies d p ≥ σ;

[0137] Obtain the plane coefficients fitted by all feature points, represented as {(C 0,i , C 1,i , 1, C bias,i )}, i = 1, 2,..., N C ;

[0138] wherein,

[0139] N C is the number of planes fitted;

[0140] Step five seven: calculate the normal vector direction n i corresponding to the i-th plane:

[0141] ​

[0142] Step fifty-eight:

[0143] In the N C selected normal vector of the solar wing plane

[0144] wherein,

[0145] The superscript T represents transposition;

[0146] Step fifty-nine:

[0147] Label the unit vector l 1,proj of the long side and the unit vector l 2,proj of the short side of the solar wing on the high-precision imaging result of the target obtained in step four 1,proj ;

[0148] Step fifty: based on the unit vector l 2,proj of the long side and the unit vector l 1,proj of the short side of the solar wing, calculate the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing.

[0149] At this point, the target attitude estimation result is represented by the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing, and the direction of the normal vector .

[0150] Preferably, the step fifty based on the unit vector l 1,proj of the long side and the unit vector l 2,proj of the short side of the solar wing, calculating the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing; is expressed as:

[0151]

[0152] wherein,

[0153] |·|2 represents the two-norm of the solution vector;

[0154] l 1,proj (1) represents the first element of l 1,proj ; l 1,proj (2) represents the second element of l 1,proj ; l 2,proj (1) represents the first element of l 2,proj ; l 2,proj (2) represents the second element of l 2,proj ;

[0155] inv() represents the inverse of the matrix;

[0156] The superscript T represents transposition.

[0157] The beneficial effects of the present application are:

[0158] The present application proposes a space target joint imaging and attitude inversion method based on a spaceborne platform, and relates to the field of inverse synthetic aperture radar imaging (ISAR), in particular to an ISAR imaging and attitude inversion method of a spaceborne platform. The present application mainly aims at the key problems of high-order three-dimensional space-varying range migration and phase error caused by complex relative motion when the spaceborne platform images space targets, and comprehensive utilization of image phase information. The main contents of the present application include: firstly, a mapping model of target Doppler parameters and image domain scatter points is established, and the spatial distribution of space-varying Doppler parameters is further estimated and reconstructed; then, an adaptive region segmentation mechanism is designed to realize regionalized consistency compensation of three-dimensional space-time error, and high-resolution imaging results are obtained; at the same time, an explicit association between Doppler parameters and satellite attitude is obtained based on the characteristics of satellite plane components, and the target attitude is directly fitted and inverted by using principal component analysis method.

[0159] The present application proposes a space target joint imaging and attitude inversion method based on a spaceborne platform, and relates to the field of inverse synthetic aperture radar imaging (ISAR), in particular to an ISAR imaging and attitude inversion method of a spaceborne platform. The present application mainly aims at the key problems of high-order three-dimensional space-varying range migration and phase error caused by complex relative motion when the spaceborne platform images space targets, and comprehensive utilization of image phase information. The main contents of the present application include: firstly, a mapping model of target Doppler parameters and image domain scatter points is established, and the spatial distribution of space-varying Doppler parameters is further estimated and reconstructed; then, an adaptive region segmentation mechanism is designed to realize regionalized consistency compensation of three-dimensional space-time error, and high-resolution imaging results are obtained; at the same time, an explicit association between Doppler parameters and satellite attitude is obtained based on the characteristics of satellite plane components, and the target attitude is directly fitted and inverted by using principal component analysis method. BRIEF DESCRIPTION OF DRAWINGS

[0160] Figure 1 The flowchart of the present application is shown in the figure;

[0161] Figure 2 The relative running track diagram is shown in the figure;

[0162] Fig. 3(a) is a RD image after translation compensation (attitude 1)

[0163] Fig. 3(b) is a RD image after KT transformation (attitude 1);

[0164] Fig. 3(c) is a high-precision imaging result (attitude 1);

[0165] Fig. 3(d) is a RD image after translation compensation (attitude 2);

[0166] Fig. 3(e) is a RD image after KT transformation (attitude 2);

[0167] Fig. 3(f) is a high-precision imaging result (attitude 2);

[0168] Fig. 4(a) is a RD image after translation compensation (case 1);

[0169] Fig. 4(b) RD image after KT transform (case 1);

[0170] Fig. 4(c) high-precision imaging result (case 1);

[0171] Fig. 4(d) RD image after translation compensation (case 2);

[0172] Fig. 4(e) RD image after KT transform (case 2);

[0173] Fig. 4(f) high-precision imaging result (case 2). DETAILED DESCRIPTION

[0174] Embodiment one: the embodiment is based on a spaceborne platform, and a specific process of a joint imaging and attitude inversion method for a space target is as follows:

[0175] Step one: a narrowband tracking system of the radar obtains a target slant range and a radar line-of-sight unit vector;

[0176] Step two: based on the target slant range in step one, a translation compensation is performed on a received signal to obtain a translation-compensated signal, and a wedgelet transform is performed on the translation-compensated signal to obtain a wedgelet-transformed signal;

[0177] Step three: a time clipping is performed on the wedgelet-transformed signal obtained in step two to obtain a clipped signal, and an imaging processing is performed on the clipped signal to obtain a coarse focusing image 1;

[0178] An adaptive template matching method (such as a multi-scale Gaussian Laplacian (LOG) detector) is used to extract a special point on the coarse focusing image 1, and a spatial Doppler parameter distribution of the special point is obtained;

[0179] Step four: an interpolation processing is performed on the spatial Doppler parameter distribution result of the special point obtained in step three to obtain a complete spatial Doppler parameter distribution result on an image projection plane;

[0180] Based on the complete spatial Doppler parameter distribution result, a real-time phase change of each region position of the image is obtained;

[0181] An imaging processing is performed on the wedgelet-transformed signal obtained in step two to obtain a coarse focusing image 2, a sub-region division is performed on the coarse focusing image 2 based on the real-time phase change, and a coordinate range of each region in an image domain is obtained;

[0182] Based on the coordinate range of each region, a local target signal corresponding to each region is obtained and compensated to obtain a compensated local target signal, and a compensated complete target signal is obtained based on the compensated local target signal of each region;

[0183] Based on the compensated complete target signal, a high-precision imaging result is obtained;

[0184] Step five: using the Taylor expansion result of the radar line-of-sight unit vector obtained in step one, the spatial Doppler parameter distribution result of the feature point obtained in step three, and the high-precision imaging result obtained in step four, the accurate estimation of the target attitude parameter is completed through plane coefficient fitting.

[0185] Specific implementation method two: the difference between this implementation method and the specific implementation method one is that the narrowband tracking system of the radar obtains the target slant range, the radar line-of-sight unit vector in step one;

[0186] The specific process is as follows:

[0187] According to the narrowband tracking system of the radar

[0188] The narrowband tracking system of the radar obtains the target slant range R T (t m ), the radar line-of-sight elevation angle azimuth angle η l (t m ), and the like;

[0189] Wherein, t m is the azimuth time domain time, that is, the azimuth dimension;

[0190] t m =mT r , T r represents the pulse repetition period, M is the number of accumulated pulses, and m=0, 1, …, M-1;

[0191] The target slant range is expressed by Taylor expansion as follows:

[0192]

[0193] Wherein,

[0194] T0 is the 0-order Taylor series expansion coefficient of R T (t m );

[0195] T1 is the 1-order Taylor series expansion coefficient of R T (t m );

[0196] T2 is the 2-order Taylor series expansion coefficient of R T (t m );

[0197] O T (t m ) represents a high-order term that can be ignored;

[0198] The radar line-of-sight unit vector ilos (t m ) is expressed as:

[0199]

[0200] The radar line of sight unit vector i los (t m ) and perform Taylor expansion to obtain:

[0201]

[0202] in,

[0203] i0 is i los (t m )'s 0th-order Taylor series expansion coefficient;

[0204] i1 is i los (t m )'s first-order Taylor series expansion coefficient;

[0205] i2 is i los (t m )'s second-order Taylor series expansion coefficients;

[0206] O los (t m ) represents higher-order terms that can be ignored.

[0207] Other steps and parameters are the same as those in the first embodiment.

[0208] Specific embodiment three: This embodiment differs from specific embodiment one or two in that, in step two, translational motion compensation is performed on the received signal based on the target slant range in step one to obtain a translationally compensated signal, and wedge transformation is performed on the translationally compensated signal to obtain a wedge transformed signal;

[0209] The specific process is:

[0210] Step 21:

[0211] The ISAR system transmits a linear frequency modulation (LFM) signal, performs pulse compression on the linear frequency modulation (LFM) signal, and obtains the target echo spectrum s(f r ,t m ), the target echo spectrum after pulse compression s(f r ,t m ) is represented as follows:

[0212]

[0213] in,

[0214] N represents the number of scattering points on the target;

[0215] σ p denotes the amplitude of the linear frequency modulated signal transmitted by the ISAR system;

[0216] f r denotes the frequency in the range dimension;

[0217] B denotes the bandwidth of the linear frequency modulated signal transmitted by the ISAR system;

[0218] f c denotes the carrier frequency;

[0219] λ denotes the radar wavelength;

[0220] rect(·) denotes the rectangular window;

[0221] j denotes the imaginary unit, j 2 = -1;

[0222] R p (t m ) denotes the range migration of the scatterer p with respect to the radar, expressed as:

[0223] R p (t m ) = R T (t m ) + R Rot,p (t m )

[0224] R Rot,p (t m ) = i los (t m ) · p T

[0225] wherein,

[0226] R Rot,p (t m ) denotes the rotational range migration of the scatterer p with respect to the radar;

[0227] · denotes the vector product;

[0228] p denotes the three-dimensional coordinate vector of the scatterer p of the target;

[0229] the upper index T denotes the transposition;

[0230] Step two two: constructing a compensation term H(f r , t m ) based on the Taylor series expansion coefficients of the slant range R T (t m ) of the target in step one;

[0231] Step two three: substituting the compensation term H(f r , tm ) and the pulse compressed target echo spectrum s(f r , t m ) to obtain the signal s'(f r , t m ) after the compensation of the translational motion;

[0232] Step two four: wedge transform is performed on the signal s'(f r , t m ) after the compensation of the translational motion to obtain the signal s'(f r , τ m ) after the wedge transform.

[0233] The other steps and parameters are the same as those in the first or second embodiment.

[0234] The fourth embodiment is different from the first to third embodiments in that the compensation term H(f T , t m ) is constructed based on the Taylor series expansion coefficients of the target slant range R r (t m ) in step two two of the first embodiment; the specific process is as follows:

[0235]

[0236] wherein,

[0237] c represents the speed of light.

[0238] The other steps and parameters are the same as those in the first to third embodiments.

[0239] The fifth embodiment is different from the first to fourth embodiments in that the compensation term H(f r , t m ) is multiplied with the pulse compressed target echo spectrum s(f r , t m ) to obtain the signal s'(f r , t m ) after the compensation of the translational motion; it is expressed as:

[0240] s'(f r , t m ) = s(f r , t m ) × H(f r , t m )

[0241] The other steps and parameters are the same as those in the first to fourth embodiments.

[0242] Specific implementation six: different from one of the specific implementations one to five, the step two four is that the signal s'(f r ,τ m ) after panning compensation is wedge-shaped transformed to obtain the signal s'(f r ,τ m ) after wedge-shaped transformation;

[0243] Wherein,

[0244] τ m is the time component after wedge-shaped transformation, τ m = (f r +f c ) / f c t m .

[0245] The other steps and parameters are the same as one of the specific implementations one to five.

[0246] Specific implementation seven: different from one of the specific implementations one to six, the step three is that the signal obtained in step two is time-truncated to obtain a truncated signal, and the truncated signal is subjected to imaging processing to obtain a coarse focus image 1;

[0247] The adaptive template matching method (such as a multi-scale Gaussian Laplacian (LOG) detector) is used to extract a special point on the coarse focus image 1, and the spatial Doppler parameter distribution of the special point is obtained;

[0248] The specific process is as follows:

[0249] Step three one:

[0250] After the processing in step two, there is still high-order space-varying range migration in the echo. In order to avoid the influence of the error on the subsequent parameter estimation, the range of the second-order range migration can be estimated according to the change of the radar line of sight and the average size of the spatial target, so that the second-order range migration in the time period is less than c / 2B as the criterion;

[0251] The signal s'(f r ,τ m ) after wedge-shaped transformation is truncated in the first time period (the signal s'(f r ,τ m ) after wedge-shaped transformation is only truncated in the first time period according to the criterion), and the truncated signal s' p (f r ,τ m ) is obtained;

[0252] Step three two:

[0253] to the intercepted signal s p (f r ,τ m ) is sequentially subjected to Inverse Fourier transform (IFT) in the range dimension and Fourier transform (FT) in the azimuth dimension, to obtain a coarse focus image 1S p (t r ,f m );

[0254] where t r represents the time in the range dimension, and f m represents the frequency in the azimuth dimension;

[0255] Step three:

[0256] For a single highlight point in the coarse focus image 1S p (t r ,f m ), the distribution of the echo intensity of each highlight point in the coarse focus image 1S p (t r ,f m ) usually follows a specific pattern in the coarse focus image 1S p (t r ,f m ).

[0257] Therefore, the highlight points in the coarse focus image 1S p (t r ,f m ) are extracted by using an adaptive template matching method (such as a multi-scale Gaussian Laplacian (LOG) detector), and the spatial Doppler parameter distribution of all the highlight points is obtained;

[0258] The spatial Doppler parameter distribution of all the highlight points is a zeroth-order and first-order Doppler parameter set {(D 0,p ,D 1,p ), p = 1, 2,..., N L} of all the extracted highlight points;

[0259] N L is the number of the extracted highlight points;

[0260] D 0,p is a zeroth-order Doppler parameter set of all the extracted highlight points;

[0261] D 1,p is a first-order Doppler parameter set of all the extracted highlight points;

[0262] Step three four:

[0263] According to the results of the LOG detector, the azimuth dimension signal corresponding to each prominent point extracted in step 33 can be extracted by an image domain filter; the specific process is as follows:

[0264] Taking the prominent point p as an example, the azimuth dimension signal corresponding to the pth prominent point extracted in step 3 is extracted by the image domain filter, expressed as s' p (τ m );

[0265] Steps 3 and 5:

[0266] Using ICPF (Integrated cubic phase function) to estimate s' p (τ m ) frequency modulation k 2,p ;

[0267] Based on s' p (τ m ) frequency modulation k 2,p , obtain the second-order Doppler parameter D corresponding to the prominent point p 2,p ; expressed as:

[0268]

[0269] Among them, f r,p is the fast time frequency corresponding to the distance dimension unit where the special point p is located,

[0270] Among them, T l Indicates the pulse width of the linear frequency modulation signal transmitted by the ISAR system;

[0271] Step 36: Repeat steps 34 to 35 to obtain the second-order Doppler parameters D corresponding to all the highlighted points. 2,p ,p=1,2,...,N L ;

[0272] Step 37: Based on the zero-order and first-order Doppler parameter sets of all the prominent points extracted in step 33 {(D 0,p ,D 1,p )},p=1,2,...,N L The second-order Doppler parameters D corresponding to all the prominent points obtained in steps 3 and 6 2,p ,p=1,2,...,N L ; Get the spatial Doppler parameter set of all the prominent points {(D 0,p ,D 1,p ,D 2,p )},p=1,2,...,N L .

[0273] Other steps and parameters are the same as one of embodiments one to six.

[0274] Embodiment eight is different from one of embodiments one to seven in that the spatial Doppler parameter distribution result of the special points obtained in step three is interpolated in step four to obtain a complete spatial Doppler parameter distribution result on the image projection plane.

[0275] Based on the complete spatial Doppler parameter distribution result, real-time phase changes of positions of each region of the image are obtained.

[0276] The wedge-shaped transformed signal obtained in step two is subjected to imaging processing to obtain a coarse focusing image 2, the coarse focusing image 2 is subjected to sub-region division based on the real-time phase changes, and the coordinate range of each region in the image domain is obtained.

[0277] Based on the coordinate range of each region, a local target signal corresponding to each region is obtained and compensated to obtain a compensated local target signal, and a compensated complete target signal is obtained based on the compensated local target signal of each region.

[0278] Based on the compensated complete target signal, a high-precision imaging result is obtained.

[0279] The specific process is as follows:

[0280] Step four one:

[0281] The spatial distribution result of the Doppler parameters of the special points obtained in step three is {(D 0,p ,D 1,p ,D 2,p}, p = 1, 2,..., N L The second-order Doppler parameter {D 2,p}, p = 1, 2,..., N L in the above formula can be used to represent the local second-order Doppler information near each special point; the second-order Doppler parameter of the remaining positions except the special points still needs to be estimated, and the second-order Doppler parameter D 2,p ,p = 1, 2,..., N L corresponding to all the special points obtained in step three is interpolated based on the natural neighbor interpolation method of triangulation to obtain the interpolated second-order Doppler parameter D2.

[0282] Step four two: based on the interpolated second-order Doppler parameter D2, the real-time phase change is obtained, which is expressed as:

[0283]

[0284] Step four three: a coarse focusing image 2S'(t rf m ); the specific process is as follows:

[0285] The wedge-shaped transformed signal s'(f r ,τ m ) obtained in step two is subjected to IFT in the distance dimension and FT in the azimuth dimension in sequence to obtain a coarse focusing image 2S'(t r ,f m );

[0286] Although D2 is obtained by estimating parameters from the intercepted signal, considering the smoothness of the orbit of the space target in the attitude-stable state, the estimation result D2 in the current short time can still represent the spatial distribution result of the second-order Doppler parameter in a long time, and is applied to subsequent fine motion compensation. When the phase change caused by the second-order Doppler parameter is less than π / 4, the error term caused by the second-order Doppler parameter can be ignored. Therefore, the coarse focusing image 2 can be divided into multiple sub-regions according to the ignoring condition, and the average value of the second-order parameter in the current sub-region is used for compensation, wherein the coarse focusing image 2 can be obtained from the wedge-shaped transformed signal s'(f r ,τ m ) in step one, and the coarse focusing image 2S'(t r ,f m ) can be obtained by performing IFT in the distance dimension and FT in the azimuth dimension on s'(f r ,τ m ) respectively.

[0287] When the real-time phase change is less than π / 4, the error term caused by the second-order Doppler parameter can be ignored. Therefore, the coarse focusing image 2S'(t q ,f r ) can be divided into Q sub-regions according to the adaptive partitioning principle according to the ignoring condition;

[0288] The image corresponding to the qth sub-region is represented as S' m (t q ,f r );

[0289] S' m (t q ,f r ) is subjected to FT in the distance dimension and IFT in the azimuth dimension in sequence to obtain the signal s' m (f q ,t r ) corresponding to the qth sub-region;

[0290] Step four five: constructing a compensation term h m (f q ,τ r) represents:

[0291]

[0292] wherein, is the average value of the second-order Doppler parameter in the qth sub-region;

[0293] Step four six:

[0294] The compensation term h q (f r ,τ m ) of the signal corresponding to the qth sub-region constructed in step four five is multiplied by the signal s' q (f r ,t m ) corresponding to the qth sub-region obtained in step four four, to obtain the compensated signal of the qth sub-region;

[0295] Step four seven:

[0296] Steps four five to four six are repeatedly executed to obtain the complete target signal after compensation of all sub-regions:

[0297]

[0298] Step four eight:

[0299] The complete target signal s t (f r ,τ m ) after compensation of all sub-regions is sequentially subjected to IFT in the range dimension and FT in the azimuth dimension, to obtain the high-precision imaging result of the target.

[0300] The other steps and parameters are the same as one of the first to seventh embodiments.

[0301] Specifically, the step five utilizes the Taylor expansion result of the radar line-of-sight unit vector obtained in step one, the spatial Doppler parameter distribution result of the feature point obtained in step three, and the high-precision imaging result obtained in step four, to complete accurate estimation of the target attitude parameter through plane coefficient fitting.

[0302] The specific process is as follows:

[0303] Step five one:

[0304] The Doppler parameter distribution result {(D 0,p ,D 1,p ,D 2,p )}, p = 1, 2,..., N L is generated, and the three-dimensional coordinates of the point cloud data D are D 0,pp = 1, 2,..., N L , D 1,p p = 1, 2,..., N L , D 2,p p = 1, 2,..., N L ;

[0305] D 0,p is the x-axis coordinate, D 1,p is the y-axis coordinate, D 2,p is the z-axis coordinate;

[0306] Step five two: let the iteration number i = 1;

[0307] Step five three: use principal component analysis to fit the plane of point cloud data D, and obtain the fitting plane and plane coefficients C 0,i , C 1,i , 1, C bias,i of the i-th cycle;

[0308] wherein,

[0309] C 0,i is the plane coefficient of D 0,p ;

[0310] C 1,i is the plane coefficient of D 1,p ;

[0311] 1 is the plane coefficient of D 2,p ;

[0312] C bias,i is the plane coefficient constant term;

[0313] Step five four: calculate the Euclidean distance d L of the N p feature points obtained in step three to the fitting plane, p = 1, 2,..., N L ; represented as:

[0314]

[0315] Step five five: calculate the standard deviation σ of the Euclidean distance; represented as:

[0316]

[0317] wherein, is the mean of the Euclidean distance, N i is the number of feature points in the i-th cycle (the initial value is the number of all feature points N L obtained in step three) ;

[0318] Step five six: retain d pLet i = i + 1, and update the number of salient points and the point cloud data, repeat steps 53-55 until the number of remaining salient points is less than three or the Euclidean distance of each salient point meets d p ≥ σ;

[0319] Obtain the plane coefficients fitted by all salient points, denoted as {(C 0,i , C 1,i , 1, C bias,i ), i = 1, 2,..., N C ;

[0320] Wherein,

[0321] N C is the number of fitted planes;

[0322] Step 57: Calculate the normal vector direction n i of the i-th plane:

[0323]

[0324] Step 58: In the N C fitted planes, select the normal vector of the solar wing plane

[0325] Wherein, the superscript T represents transposition;

[0326] Step 59:

[0327] Label the unit vector l 1,proj of the long side and the unit vector l 2,proj of the short side of the solar wing on the high-precision imaging result of the target obtained in step 4;

[0328] Step 50: Based on the unit vector l 1,proj of the long side and the unit vector l 2,proj of the short side of the solar wing, calculate the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing;

[0329] At this point, the target attitude estimation result is represented by the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing, and the direction of the normal vector .

[0330] The other steps and parameters are the same as one of the first to eighth embodiments.

[0331] The tenth embodiment is different from one of the first to ninth embodiments in that in the step 50, based on the unit vector l 1,proj of the long side and the unit vector l 2,proj of the short side of the solar wing, the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side of the solar wing are calculated., calculate the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side; expressed as:

[0332]

[0333] in,

[0334] |·|2 means the binary norm of the solution vector;

[0335] l 1,proj (1) indicates l 1,proj The first element of l 1,proj (2) indicates l 1,proj The second element of l 2,proj (1) indicates l 2,proj The first element of l 2,proj (2) indicates l 2,proj The second element of

[0336] inv() means to invert the matrix;

[0337] The superscript T means to find the transpose;

[0338] The other steps and parameters are the same as those in Specific Embodiments 1 to 9.

[0339] Example:

[0340] The joint imaging and attitude inversion method of space targets based on a satellite-borne platform designed in the present invention is verified by simulation, and the results are explained.

[0341] Simulation 1 is used to verify the effectiveness of the present invention in situational awareness of space targets with different postures. Figure 2 Figure 3 shows the high-precision imaging results of the target at different poses obtained by the situational awareness algorithm proposed in this invention. The corresponding image entropy and pose estimation error are shown in Table 1, where the pose estimation error is represented by the angle between the actual normal vector of the target solar panel and the estimated result. As can be seen from Figure 3 and Table 1, the algorithm proposed in this invention can obtain high-precision imaging results for different target poses and achieve effective target pose estimation.

[0342] Table 1. Space target situation awareness estimation results

[0343]

[0344] Simulation 2 is used to verify the effectiveness of the present invention in the situational awareness of space targets with different orbital errors. Figure 2The orbit error is shown in Table 2. Fig. 4 shows the high-precision imaging results of the situation awareness algorithm proposed by the present application in the presence of different orbit errors, and the corresponding image entropy and attitude estimation error are shown in Table 3. As can be seen from Fig. 4 and Table 3, the algorithm proposed by the present application can obtain high-precision imaging results in the presence of different orbit errors and effectively estimate the target attitude.

[0345] Table 2 Orbit error

[0346]

[0347] Table 3 Space target situation awareness estimation results

[0348]

[0349] The present application can also have other various embodiments, and those skilled in the art can make various corresponding changes and modifications according to the present application without departing from the spirit and essence of the present application, but these corresponding changes and modifications should all belong to the protection scope of the claims attached to the present application.

Claims

1. A method for joint imaging and attitude inversion of space targets based on a spaceborne platform, characterized by: The specific process of the method is: Step 1: The radar's narrowband tracking system obtains the target slant range and radar line of sight unit vector; Step 2: performing translation compensation on the received signal based on the target slant range in step 1 to obtain a translation-compensated signal, and performing wedge transformation on the translation-compensated signal to obtain a wedge-transformed signal; Step 3: performing time interception on the wedge-shaped transformed signal obtained in step 2 to obtain an intercepted signal, performing imaging processing on the intercepted signal to obtain a coarse focus image 1; Using the adaptive template matching method, the prominent points on the coarse focus image 1 are extracted and the spatial Doppler parameter distribution of the prominent points is obtained; Step 4: interpolate the spatial Doppler parameter distribution results of the prominent points obtained in step 3 to obtain the complete spatial Doppler parameter distribution results on the image projection plane; Based on the complete spatial Doppler parameter distribution results, the real-time phase change of each area of ​​the image is obtained; Performing imaging processing on the wedge-shaped transformed signal obtained in step 2 to obtain a coarse focus image 2, dividing the coarse focus image 2 into sub-regions based on the real-time phase change to obtain the coordinate range of each region in the image domain; Based on the coordinate range of each area, a local target signal corresponding to each area is obtained and compensated to obtain a compensated local target signal, and a compensated complete target signal is obtained based on the compensated local target signal of each area; Obtain imaging results based on the compensated complete target signal; Step 5: Using the Taylor expansion result of the radar line of sight unit vector obtained in step 1, the spatial Doppler parameter distribution result of the prominent point obtained in step 3, and the imaging result obtained in step 4, the target attitude parameters are accurately estimated through plane coefficient fitting.

2. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 1, characterized in that: In step 1, the narrowband tracking system of the radar obtains the target slant range and the radar line of sight unit vector; the specific process is: Acquired based on the radar's narrowband tracking system The radar's narrowband tracking system obtains the target slant range R T (t m ), the elevation angle of the radar line of sight Azimuth angle η l (t m )parameter; Among them, t m is the azimuth time domain time, i.e., the azimuth dimension; t m =mT r , T r represents the pulse repetition period, M is the number of accumulated pulses, m=0,1,…,M-1; The target slant range is expressed by Taylor expansion as: in, T0 is R T (t m )'s 0th-order Taylor series expansion coefficient; T1 is R T (t m )'s first-order Taylor series expansion coefficient; T2 is R T (t m )'s second-order Taylor series expansion coefficients; O T (t m ) represents the higher-order terms that can be ignored; Radar line of sight unit vector i los (t m ) is expressed as: The radar line of sight unit vector i los (t m ) and perform Taylor expansion to obtain: in, i0 is i los (t m )'s 0th-order Taylor series expansion coefficient; i1 is i los (t m )'s first-order Taylor series expansion coefficient; i2 is i los (t m )'s second-order Taylor series expansion coefficients; O los (t m ) represents higher-order terms that can be ignored.

3. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 2, characterized in that: In the step 2, the received signal is subjected to translational motion compensation based on the target slant range in the step 1 to obtain a translationally compensated signal, and the translationally compensated signal is subjected to wedge transformation to obtain a wedge transformed signal. The specific process is as follows: Step 21: The ISAR system transmits a linear frequency modulation signal and performs pulse compression on the linear frequency modulation signal to obtain the target echo spectrum s(f r ,t m ), the target echo spectrum after pulse compression s(f r ,t m ) is represented as follows: in, N represents the number of scattering points on the target; σ p Represents the amplitude of the linear frequency modulation signal transmitted by the ISAR system; f r represents the frequency of the distance dimension; B represents the bandwidth of the linear frequency modulation signal transmitted by the ISAR system; f c Indicates the carrier frequency; λ represents the radar wavelength; rect(·) represents a rectangular window; j represents the imaginary unit, j 2 =-1; R p (t m ) represents the range migration of the scattering point p relative to the radar, which can be expressed as: R p (t m )=R T (t m )+R Rot,p (t m ) R Rot,p (t m )=i los (t m )·p T in, R Rot,p (t m ) represents the rotational range migration of the scattering point p relative to the radar; Represents vector product; p represents the three-dimensional coordinate vector of the scattering point p of the target; The superscript T means to find the transpose; Step 22: Based on the target slant range R in step 1 T (t m ) constructs the compensation term H(f r ,t m ); Step 23: Compensation term H(f r ,t m ) and the target echo spectrum s(f r ,t m ) to obtain the signal s'(f r ,t m ); Step 24: Compensate the signal s'(f r ,t m ) is wedge-shaped transformed to obtain the wedge-shaped transformed signal s'(f r ,τ m ).

4. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 3, characterized in that: In step 22, based on the target slant distance R in step 1 T (t m ) constructs the compensation term H(f r ,t m ); the specific process is: in, c represents the speed of light.

5. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 4, characterized in that: In the steps 2 and 3, the compensation term H(f r ,t m ) and the target echo spectrum s(f r ,t m ) to obtain the signal s'(f r ,t m ); expressed as: s'(f r ,t m )=s(f r ,t m )×H(f r ,t m )。 6. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 5, characterized in that: In step 24, the signal s'(f r ,t m ) is wedge-shaped transformed to obtain the wedge-shaped transformed signal s'(f r ,τ m ); in, τ m is the time component after wedge transformation, τ m =(f r +f c ) / f c t m .

7. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 6, characterized in that: In the step 3, the wedge-shaped transformed signal obtained in the step 2 is time-intercepted to obtain an intercepted signal, and the intercepted signal is image-processed to obtain a coarse focus image 1; Using the adaptive template matching method, the prominent points on the coarse focus image 1 are extracted and the spatial Doppler parameter distribution of the prominent points is obtained; The specific process is: Step 31: The criterion is that the second-order distance migration within the interception time period is less than c / 2B; According to the criterion, the signal s'(f r ,τ m ) intercept the first time period and obtain the intercepted signal s' p (f r ,τ m ); Step 32: For the intercepted signal s' p (f r ,τ m ) Perform inverse Fourier transform of the distance dimension and Fourier transform of the azimuth dimension in sequence to obtain a coarse focus image 1S' p (t r ,f m ); Among them, t r represents the time of the distance dimension, f m represents the frequency of the azimuthal dimension; Step 33: Adaptive template matching method is used to match the coarse focus image 1S' p (t r ,f m ) to extract the prominent points and obtain the spatial Doppler parameter distribution of all the prominent points; The spatial Doppler parameter distribution of all the prominent points is the set of zero-order and first-order Doppler parameters of all the prominent points extracted {(D 0,p ,D 1,p )},p=1,2,...,N L ; N L is the number of extracted distinctive points; D 0,p is the set of zero-order Doppler parameters of all the extracted prominent points; D 1,p is the set of first-order Doppler parameters of all the extracted distinctive points; Steps 3 and 4: The azimuth dimension signal corresponding to each prominent point extracted in step 33 is extracted through the image domain filter; the specific process is as follows: The azimuth dimension signal corresponding to the pth prominent point extracted in step 3 is extracted by the image domain filter and expressed as s' p (τ m ); Steps 3 and 5: Using ICPF to estimate s' p (τ m ) frequency modulation k 2,p ; Based on s' p (τ m ) frequency modulation k 2,p , obtain the second-order Doppler parameter D corresponding to the prominent point p 2,p ; expressed as: Among them, f r,p is the fast time frequency corresponding to the distance dimension unit where the special point p is located, Among them, T l Indicates the pulse width of the linear frequency modulation signal transmitted by the ISAR system; Step 36: Repeat steps 34 to 35 to obtain the second-order Doppler parameters D corresponding to all the highlighted points. 2,p ,p=1,2,...,N L ; Step 37: Based on the zero-order and first-order Doppler parameter sets of all the prominent points extracted in step 33 {(D 0,p ,D 1,p )},p=1,2,...,N L The second-order Doppler parameters D corresponding to all the prominent points obtained in steps 3 and 6 2,p ,p=1,2,...,N L ; Get the spatial Doppler parameter set of all the prominent points {(D 0,p ,D 1,p ,D 2,p )},p=1,2,...,N L .

8. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 7, characterized in that: In the step 4, interpolation processing is performed on the spatial Doppler parameter distribution result of the prominent point obtained in the step 3 to obtain a complete spatial Doppler parameter distribution result on the image projection plane; Based on the complete spatial Doppler parameter distribution results, the real-time phase change of each area of ​​the image is obtained; Performing imaging processing on the wedge-shaped transformed signal obtained in step 2 to obtain a coarse focus image 2, dividing the coarse focus image 2 into sub-regions based on the real-time phase change to obtain the coordinate range of each region in the image domain; Based on the coordinate range of each area, a local target signal corresponding to each area is obtained and compensated to obtain a compensated local target signal, and a compensated complete target signal is obtained based on the compensated local target signal of each area; Obtain imaging results based on the compensated complete target signal; The specific process is: Step 41: The second-order Doppler parameters D corresponding to all the prominent points obtained in step 3 are calculated based on the natural neighbor interpolation method of triangulation. 2,p ,p=1,2,...,N L Perform interpolation to obtain the interpolated second-order Doppler parameter D2; Step 42: Obtain real-time phase change based on the interpolated second-order Doppler parameter D2 Expressed as: Step 43: Construct coarse focus image 2S'(t r ,f m ); the specific process is: The wedge-shaped transformed signal s'(f r ,τ m ) and perform IFT of distance dimension and FT of azimuth dimension in sequence to obtain the coarse focusing image 2S'(t r ,f m ); Step 44: When the real-time phase changes When it is less than π / 4, the coarse focus image 2S'(t r ,f m ) is divided into Q sub-areas; The image corresponding to the qth sub-region is represented as S' q (t r ,f m ); To S' q (t r ,f m ) Perform the distance dimension FT and the azimuth dimension IFT in sequence to obtain the signal s' corresponding to the qth sub-region q (f r ,t m ); Step 4 and 5: Construct the compensation term h of the signal corresponding to the qth sub-region q (f r ,τ m ); expressed as: in, is the average value of the second-order Doppler parameter in the qth sub-region; Steps 4 and 6: The compensation term h of the signal corresponding to the qth sub-region constructed in steps 4 and 5 is q (f r ,τ m ) and the signal s' corresponding to the qth sub-region obtained in step 44 q (f r ,t m ) to obtain the compensated signal of the qth sub-region; Step 47: Repeat steps 4-5 to 4-6 to obtain the complete target signal after all sub-areas are compensated: Step 48: The complete target signal s after compensation for all sub-areas t (f r ,τ m ) Perform IFT of the distance dimension and FT of the azimuth dimension in sequence to obtain the imaging result of the target.

9. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 8, characterized in that: In step 5, the Taylor expansion result of the radar line of sight unit vector obtained in step 1, the spatial Doppler parameter distribution result of the prominent point obtained in step 3, and the imaging result obtained in step 4 are used to accurately estimate the target attitude parameters through plane coefficient fitting. The specific process is as follows: Step 51: The Doppler parameter distribution results {(D 0,p ,D 1,p ,D 2,p )},p=1,2,...,N L Generate point cloud data D, the three-dimensional coordinates of point cloud data D are D 0,p ,p=1,2,...,N L , D 1,p ,p=1,2,...,N L , D 2,p ,p=1,2,...,N L ; D 0,p is the x-axis coordinate, D 1,p is the y-axis coordinate, D 2,p is the z-axis coordinate; Step 52: Set the number of iterations i = 1; Step 53: Use principal component analysis to perform plane fitting on the point cloud data D to obtain the fitting plane and plane coefficient C of the i-th cycle 0,i ,C 1,i ,1,C bias,i ; in, C 0,i D 0,p The plane coefficient of C 1,i D 1,p The plane coefficient of 1 is D 2,p The plane coefficient of C bias,i is the plane coefficient constant term; Step 54: Calculate N obtained in step 3 L The Euclidean distance d from a prominent point to the fitting plane p ,p=1,2,...,N L ; expressed as: Step 55: Calculate the standard deviation σ of the Euclidean distance; expressed as: in, is the mean of the Euclidean distance, N i is the number of highlighted points in the i-th cycle; Steps 5 and 6: Keep d p <σ, set the number of iterations i = i + 1, and update the number of outstanding points and point cloud data, repeat steps 53 to 55 until the number of remaining outstanding points is less than three or the Euclidean distance of each outstanding point satisfies d p ≥σ; Obtain the plane coefficients fitted by all the prominent points, expressed as {(C 0,i ,C 1,i ,1,C bias,i )},i=1,2,...,N C ; in, N C is the number of fitted planes; Step 57: Calculate the normal vector direction n corresponding to the i-th plane i : Step 58: In N C From the fitted planes, select the normal vector corresponding to the solar wing plane in, The superscript T means to find the transpose; Step 59: Mark the unit vector l of the long side of the solar wing on the high-precision imaging result of the target obtained in step 4 1,proj and the unit vector l of the short side 2,proj ; Step 50: Unit vector l based on the long side of the solar wing 1,proj and the unit vector l of the short side 2,proj , calculate the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side; At this point, the target attitude estimation result is composed of the three-dimensional vector l1 corresponding to the long side of the solar wing, the three-dimensional vector l2 corresponding to the short side, and the normal vector Direction indication.

10. The method for joint imaging and attitude inversion of space targets based on a satellite-borne platform according to claim 9, characterized in that: The unit vector l based on the long side of the solar wing in step 50 1,proj and the unit vector l of the short side 2,proj , calculate the three-dimensional vector l1 corresponding to the long side of the solar wing and the three-dimensional vector l2 corresponding to the short side; expressed as: in, |·|2 means the binary norm of the solution vector; l 1,proj (1) indicates l 1,proj The first element of l 1,proj (2) indicates l 1,proj The second element of l 2,proj (1) indicates l 2,proj The first element of l 2,proj (2) indicates l 2,proj The second element of inv() means to invert the matrix; The superscript T means transpose.

Images