Handheld sar three-dimensional fast near-field imaging method based on local spectrum compression

Abstract

This invention relates to the field of millimeter-wave near-field imaging technology, and particularly to a handheld SAR three-dimensional fast near-field imaging method based on local spectral compression, comprising: acquiring raw echo data, dividing the complete synthetic array into 2 M‑1 1. Calculate a first-order subarray; for each first-order subarray, compute a uniform sampling grid with a uniform sampling rate in the (u,v,n) coordinate system, and convert it into a first-order non-uniform sampling grid in the (x,y,z) coordinate system; then, in the first-order non-uniform sampling grid, pair 2... M‑1 Reconstructing the first-order subarrays yields 2 M‑1 One first-level sub-image; for 2 M‑1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether these sub-images meet preset conditions. If not, the pairwise coherent superposition process is iteratively executed until the preset conditions are met, yielding the 3D reconstruction result. This solves the problems of low imaging quality and excessive computational complexity inherent in traditional imaging methods.

Description

Technical Field

[0001] This invention relates to the field of millimeter-wave near-field imaging technology, and in particular to a handheld SAR three-dimensional fast near-field imaging method based on local spectral compression applicable to arbitrary trajectories. Background Technology

[0002] Near-field active array-based millimeter-wave imaging technology can achieve high spatial resolution while maintaining large apertures and wide operating bandwidths. Combined with the non-ionizing and penetrating properties of millimeter waves, this technology has been widely studied and applied in various fields, including security inspection, medical diagnostics, through-wall imaging, and non-destructive testing.

[0003] In applications such as security inspection and non-destructive testing, 3D millimeter-wave imaging systems require not only high performance but also compact size, portability, and reasonable cost. Handheld SAR imaging systems, due to their unique advantages such as high flexibility, ease of operation, and cost-effectiveness, have become a highly competitive solution and have attracted widespread research attention. Compared to traditional SAR systems, handheld systems employ manual scanning of physical arrays. While this method introduces uncertainty and non-uniformity in the scanning path, it also brings significant operational convenience and adaptability. This path uncertainty places higher demands on image reconstruction methods, making accurate radar system motion tracking and the development of fast and accurate 3D imaging methods adaptable to arbitrary scanning trajectories key challenges. Although accurate motion tracking of radar systems has been extensively studied, accurate and efficient 3D imaging methods for handheld SAR systems are still lacking. Many application scenarios have extremely high requirements for imaging speed and quality, urgently requiring accurate real-time imaging technology.

[0004] However, existing traditional imaging methods, such as back projection in the time domain (BPA) and Kirchhoff migration method (KMA), achieve high-precision image reconstruction by effectively compensating for trajectory deviations at each scanning position and can be directly applied to non-uniform composite arrays. However, their computational complexity is too high, hindering their application in real-time imaging systems. Most spatial wavenumber domain imaging methods are designed for uniform arrays. Although existing research has proposed methods suitable for non-uniform sparse arrays, such as the distance migration method based on non-uniform fast Fourier transform (NUFFT-based RMA) and the equivalent phase center (EPC) approximation, their imaging quality and computational efficiency are not high. Therefore, this invention addresses these problems by innovatively proposing a handheld, fast decomposition back projection 3D imaging method that ensures both good imaging quality and high computational efficiency. Summary of the Invention

[0005] This invention provides a fast three-dimensional near-field imaging method for handheld SAR based on local spectral compression, in order to solve the problem that imaging accuracy and computational efficiency cannot be guaranteed simultaneously in near-field handheld SAR system imaging methods.

[0006] A first aspect of this invention provides a method for rapid three-dimensional near-field imaging using a handheld SAR system based on local spectral compression, comprising the following steps: acquiring raw echo data emitted by a preset handheld SAR system; dividing the synthetic aperture of the preset handheld SAR system according to the target level to obtain 2 M-1 The system extracts the echo signal of each first-level subarray from the original echo data, where M is the target level and M≥2. It then uses LLT to calculate a uniform sampling grid with a uniform sampling rate for each first-level subarray in the (u,v,n) coordinate system, and converts this uniform sampling grid into a first-level non-uniform sampling grid in the (x,y,z) coordinate system. Finally, it uses the echo signal of each first-level subarray to perform a sampling on the 2... M-1 Reconstructing the first-order subarrays yields 2 M-1 One first-level sub-image; for the 2 M-1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether the at least two second-level sub-images meet a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, and a three-dimensional reconstruction result is obtained.

[0007] Optionally, the solution formula for the uniform sampling grid is:

[0008]

[0009] Where (u,v,n) is the coordinate system after LLT coordinate transformation, k max k min x′ represents the lower and upper limits of the wavenumber of millimeter-wave signals, respectively. min,SA 、x′ max,SA y′ min,SA and y′ max,SA These represent the upper and lower limits of x′ and y′ for each first-order subarray, where x, y, and z are the coordinates of the scattering points within the imaging region, and x′... 0,SA y′ 0,SA Let x′ and y′ be the closest points to x and y in each first-order subarray, respectively, and r be the sum of the distances between the coordinates of the scattering point p in the imaging region and the four vertices of the subarray. min 、x′ max y′ min y′ maxLet x′ and y′ be the upper and lower limits of the coordinates of the inner array element of the integrated aperture, respectively, and t be the sum of the distances between the coordinates of the scattering point p in the imaging region and the four vertices of the subarray. The formula is as follows:

[0010]

[0011] Optionally, the 2 M-1 The formula for solving each first-order sub-image is:

[0012]

[0013] Among them, f 1,n (p) is the sub-image of each first-level subarray, s(p′,τ) is the echo signal of each first-level subarray, τ is the two-way propagation delay between the array element and the scattering point in the imaging region, p is the coordinate of the scattering point in the imaging region, p′ is the coordinate of the array element in the synthetic aperture, and c is the speed of light.

[0014] Optionally, the 2 M-1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether the at least two second-level sub-images satisfy a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, resulting in the sub-image of the target level, including:

[0015] Calculate the 2 M-2 The non-uniform sampling grid corresponding to each secondary sub-image;

[0016] The first-level sub-image of each sub-array is spatially down-converted using the SDC method to obtain 2 M-1 One compressed sub-image;

[0017] The 2 M-1 Each compressed sub-image is interpolated into its corresponding non-uniform sampling grid, and then each compressed sub-image is multiplied by its conjugate space down-conversion function to obtain 2 M-1 One original spectrum sub-image;

[0018] The 2 M-1 The original spectral sub-images are coherently superimposed pairwise to obtain 2 M-2 Sub-images;

[0019] Determine the 2 M-2 If a sub-image does not meet the preset conditions, a new non-uniform sampling grid is iteratively calculated and pairwise coherent superposition is performed until the preset conditions are met, and the three-dimensional reconstruction result is obtained.

[0020] Optionally, the formula for solving the compressed sub-image is:

[0021] f′ m,n(p)=f m,n (p)e -jΦ

[0022]

[0023] Where, f′ m,n (p) is the compressed sub-image of the current level sub-image of the subarray, where m is the current level, 2≤m≤M, and n=2. M-m The number of subarrays at the current level, p is the coordinate of the scattering point within the imaging region, and f is the number of subarrays at the current level. m,n (p) represents the current level sub-image of the subarray, e is the natural logarithm, j is the imaginary unit, Φ is the phase term, and k max k min Let x' be the lower and upper bounds of the millimeter-wave signal wavenumber, r be the sum of the distances between the spatial point p and the four vertices of the subarray, and x' be the upper and lower bounds of the wavenumber. min 、x′ max y′ min y′ max These are the upper and lower limits of the x′ and y′ coordinates of the inner array elements of the integrated aperture, respectively.

[0024] Optionally, the preset condition is the 2 M-2 Whether each sub-image is the final imaging result.

[0025] A second aspect of the present invention provides a handheld SAR three-dimensional fast near-field imaging device based on local spectral compression, comprising: an acquisition module for acquiring raw echo data emitted by a preset handheld SAR system; and a segmentation module for segmenting the synthetic aperture of the preset handheld SAR system according to the target level to obtain 2 M-1 The system comprises three first-level subarrays, and extracts the echo signal of each first-level subarray from the original echo data, where M is the target level, M≥2; a network conversion module is used to calculate a uniform sampling grid with a uniform sampling rate in the (u,v,n) coordinates for each first-level subarray using LLT, and convert the uniform sampling grid into a first-level non-uniform sampling grid in the (x,y,z) coordinates; a reconstruction module is used to reconstruct the 2 first-level subarrays using the echo signal of each first-level subarray in the first-level non-uniform sampling grid. M-1 Reconstructing the first-order subarrays yields 2 M-1 One first-level sub-image; an iterative overlay module, used for iteratively overlaying the two... M-1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether the at least two second-level sub-images meet a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, and a three-dimensional reconstruction result is obtained.

[0026] A third aspect of the present invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression as described in the above embodiments.

[0027] A fourth aspect of the present invention provides a computer program product, which, when executed by a processor, implements the above-described handheld SAR three-dimensional fast near-field imaging method based on local spectral compression.

[0028] A fifth aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described handheld SAR three-dimensional fast near-field imaging method based on local spectral compression.

[0029] The present invention proposes a handheld SAR 3D fast near-field imaging method based on local spectrum compression. This method divides the complete integrated array into multiple subarrays, reconstructs downsampled sub-images of each subarray using temporal imaging methods, and obtains subsequent sub-images by merging pairs of subarrays, ultimately achieving accurate 3D reconstruction results. Simultaneously, by analyzing the characteristics of the local spectrum, analytical expressions for SDC and LLT are derived, which are then used for spectral compression of sub-images, thereby reducing the number of sampling points. This overcomes the problem of excessive time complexity in traditional temporal imaging methods, enabling accurate and efficient 3D imaging of handheld SAR systems with arbitrary trajectories.

[0030] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0031] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0032] Figure 1 This is a flowchart illustrating a handheld SAR three-dimensional fast near-field imaging method based on local spectral compression provided in an embodiment of the present invention.

[0033] Figure 2 The geometric structure of the close-range handheld SAR imaging system provided in the embodiments of the present invention;

[0034] Figure 3 This is a schematic diagram illustrating the specific execution of the handheld SAR three-dimensional fast near-field imaging method based on local spectrum compression provided in an embodiment of the present invention.

[0035] Figure 4This is a schematic diagram of the local spectral support domain of a sub-image at any three locations in the imaging region, provided in an embodiment of the present invention.

[0036] Figure 5 This is a comparative diagram of the imaging effect provided in the embodiment of the present invention, wherein (a) is the imaging result of the BPA imaging method, and (b) is the imaging result of the embodiment of the present invention;

[0037] Figure 6 This is a block diagram of a handheld SAR three-dimensional fast near-field imaging device based on local spectrum compression provided in an embodiment of the present invention.

[0038] Figure 7 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0039] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0040] The following describes an embodiment of the present invention, a handheld SAR three-dimensional fast near-field imaging method based on local spectral compression, with reference to the accompanying drawings. Addressing the problem mentioned in the background section of the present invention that traditional imaging methods cannot simultaneously achieve both imaging quality and computational efficiency, the present invention provides a handheld SAR three-dimensional fast near-field imaging method based on local spectral compression. In this method, the complete integrated array is divided into multiple subarrays, the sub-image spectrum is compressed using SDC and LLT, the downsampled sub-images of each subarray are reconstructed using a temporal imaging method, and subsequent sub-images are obtained by merging them pairwise. This achieves accurate and efficient three-dimensional imaging of a handheld SAR system with arbitrary trajectories, thereby solving the problem of the inability to simultaneously achieve both imaging quality and computational efficiency in traditional imaging methods.

[0041] To facilitate those skilled in the art to understand the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression of this application, the parameter symbols involved in this application are explained in Table 1 below, and the geometric structure of the near-field handheld SAR imaging system is given to analyze the spectral characteristics of SAR images. The derivation process and results of LLT and SDC are given to provide theoretical support for reducing the computational load of 3D imaging in the future.

[0042] Table 1

[0043]

[0044]

[0045] like Figure 1 As shown in the figure, the linear single-input single-output (SISO) array forms a synthetic aperture by manual scanning and illuminates the target located in front of the aperture. Compared with traditional SAR systems that use mechanical scanning, the handheld system exhibits significant uncertainty in the scanning trajectory, leading to fluctuations in the positions of the array elements within the synthetic aperture. The coordinates of the origin in the SISO array are represented as p′=(x′,y′,z′), while the coordinates of the scattering points within the imaging region are represented as p=(x,y,z).

[0046] The propagation process of millimeter-wave signals follows the first-order Born approximation, and propagation attenuation is ignored. The echo signal can be expressed as:

[0047] s(p′,k)=∫∫∫f(p)e -j2k||p-p′|| dp (1)

[0048] Where s is the echo signal, k = 2πf / c, c is the speed of light, ||·|| is the amplitude of the vector, ||pp′|| is the distance between the array element and the scattering point, f(·) is the reflectivity distribution function of the target, which is also the objective function in the reconstruction process, e is the natural logarithm, and j is the imaginary unit.

[0049] Among existing image reconstruction methods, BPA (Block Image Processing) is widely considered the gold standard due to its excellent imaging quality. First, the measurement signal is transformed to the time domain using a Fast Inverse Fourier Transform (FFT), and then BPA is used for reconstruction. The formula is as follows:

[0050]

[0051] In the formula, k min k max τ represents the lower and upper limits of the wavenumber of the millimeter wave signal, respectively. The integral region of formula (3) is the composite aperture, and τ is the two-way propagation delay between the array element and the scattering point of the imaging region.

[0052] Substituting formula (2) into formula (3), the overall formula for BPA is:

[0053]

[0054] In Equation (4), BPA reconstructs the imaging result pixel by pixel by performing spatial wavenumber integration at each location within the imaging region. Therefore, the computational cost of BPA is proportional to the number of sampling points in the imaging result, and the number of sampling points is determined by the spectral bandwidth characteristics of the imaging result.

[0055] For close-range SAR images, the spectral characteristics, including carrier frequency and bandwidth, exhibit strong spatial variability. Therefore, local spectral analysis is required to better simulate the spectral characteristics of handheld SAR images.

[0056] The spatial wavenumber domain representation of handheld SAR images can be directly calculated using 3D Fourier transform, i.e.

[0057]

[0058] Where, k = (k x ,k y ,k z ) represents the spatial wavenumber domain variable of the image spectrum, k x ,k y and k z These are the corresponding terms in the spatial wavenumber domains for x, y, and z, respectively.

[0059] Then, the integrals over x, y, and z are solved using the stationary phase principle (POSP), as follows:

[0060]

[0061] Here, α and p0 are the slowly changing amplitude term and the stationary point in POSP, respectively.

[0062] The stationary position meets the following requirements:

[0063]

[0064] This can be further simplified to:

[0065]

[0066] Therefore, any k = k0 = (k x ,k y ,k z The expression for ) is:

[0067]

[0068] Alternatively, k0 can be expressed as a function of p′, k, and p0:

[0069]

[0070] Equations (9) and (10) determine the relationship between wavenumber domain variables and system parameter variables. For any wavenumber domain variable k... x ,k y ,k z Only when conditions (8) and (9) are satisfied will the asymptotic integral of formula (6) be non-zero, which is equivalent to a finite set of wavenumber domain points. This property can be used to determine the spectral support range of the imaging result, as shown in the formula:

[0071]

[0072] Where D is the imaging region containing the imaging target, and A is the composite aperture. For any position p0∈D in the imaging region, the corresponding spectral support range can be calculated using formulas (8) and (9):

[0073]

[0074] This is defined as the local spectral support domain at p0. Therefore, the spectral support domain of the entire image can be represented as the joint of the local spectral support domains at each location in the imaging region, i.e.:

[0075]

[0076] Equations (11)-(13) describe the characteristics supported by the spectral density of handheld SAR images. The Nyquist sampling rate for each dimension can now be derived accordingly, i.e.

[0077]

[0078] Where, k xmin k xmax k ymin k ymax k zmin and k zmax k are determined by formula (11) x k y and k z The lower and upper limits.

[0079] To simplify the formula, ignoring the minute fluctuations along the z-axis of the array elements, the above equation can be further extended to:

[0080]

[0081] Where, x min x max y min y max and z min Let x and y be the lower and upper bounds of the imaging, respectively, and let z be the lower bound, x′. min ,x′ max ,y′ min ,y′ max To define the upper and lower limits of the combined aperture x′, y′, x d,max and y d,max The formula for calculating the maximum value of |xx′| and |yy′| is as follows:

[0082]

[0083] Finally, the number of sampling points required for the imaging results is:

[0084]

[0085] Where V(D) is the volume of the imaging region D. As can be seen from the above discussion, handheld SAR images can be modeled as spatially varying band-limited 3D signals, and the spectral support domain of the image at any location can be analyzed. This allows for the design of an efficient spectral compression method combining decomposition techniques, reducing sampling requirements and thus decreasing the computational load of temporal reconstruction methods.

[0086] Furthermore, such as Figure 2 As shown, the outline of the local spectral support domain is a fan-shaped region, and its basis is a curved quadrilateral when projected onto the origin. In order to maximize the spectral occupancy of the sub-image after local spectral transformation, key points in the spatial wavenumber domain can be used to represent the approximate minimum volume bounding box of the local spectrum at any position p in the subarray SA and the imaging region, which is defined as follows:

[0087] k1=k0((x′ min,SA ,y′ 0,SA ,0),p,k max )

[0088] k2=k0((x′ max,SA ,y′ 0,SA ,0),p,k max )

[0089] k3=k0((x′ 0,SA ,y′ min,SA ,0),p,k max )

[0090] k4=k0((x′ 0,SA ,y′ max,SA ,0),p,k max )

[0091] k5=(k0((x′ min,SA ,y′ min,SA ,0),p,k min 0+k0((x′ min,SA ,y′ max,SA ,0),p,k min )+k0((x′ max,sA ,y′ min,sA ,0),p,k min )+k0((x′ max,SA ,y′ max,SA ,0),p,k min )) / 4

[0092]

[0093] Where, x′ min,SA、x′ max,SA y′ min,SA and y′ max,SA These are the upper and lower limits of x′ and y′ of the subarray SA, respectively. 0,SA y′ 0,SA Let x′ and y′ be the closest values ​​to x and y in the subarray SA, and their formulas are as follows:

[0094]

[0095]

[0096] The key points (k1,k2), (k3,k4), and (k5,k6) in the spatial wavenumber domain describe the span of the local spectral support domain along the azimuth and range directions, respectively. Therefore, the approximate minimum volume bounding box of the local spectral support domain is represented by a parallelepiped formed by three vectors: k2-k1, k4-k3, and k6-k5. However, in these six key points, the curvature of the vector field formed by k6 is not zero relative to p. Therefore, the integral factor method is used to transform k6 into an irrotational field through spatial scaling:

[0097] k7=α′(x,y,z)k6=α(x,y,z)k5 (20)

[0098]

[0099] Where α′(x,y,z) is the space-variable scaling factor, α(x,y,z)=2k max α′(x,y,z) / |k5|, the solution is:

[0100] α(x,y,z)=β(r)(22)

[0101] Where β(·) is any continuously differentiable function, and r is the sum of the distances between the spatial point p and the four vertices of the subarray, its formula is:

[0102]

[0103] Formulas (20), (22) and (23) describe the general solution of converting k6 into an irrotational vector field by the integral factor method, and then solving the arbitrary function β(·) in the following manner so that the obtained k7 is the same as k6 in the axial view direction of the sub-aperture.

[0104]

[0105] like Figure 2As shown, the positions of key points k1, k2, k3, k4, k5, and k7 at different locations p are illustrated. Furthermore, the parallelepipeds formed by vectors k2-k1, k4-k3, and k7-k5 are also represented by wireframes. Based on this result, SDC and LLT are subsequently used to convert the local spectral bounding box into a unit box located at the origin of the spatial wavenumber domain, achieving spectral compression.

[0106] SDC is achieved by multiplying a sub-image in the spatial domain with the phase term of the spatial variable, assuming f m,n For the imaging result of any subarray, the SDC operation is as follows:

[0107] f′ m,n (p)=f m,n (p)e -jΦ (25)

[0108] Wherein, the gradient of Φ should be the center k of the local spectral support domain. c ,Right now:

[0109]

[0110] Where, k c Represented using key points in the spatial wavenumber domain:

[0111]

[0112] The solution to Φ is:

[0113]

[0114] like Figure 2 As shown, after SDC, the local spectral support domain is moved to the origin of the spatial wavenumber domain, thereby significantly reducing the sampling rate required for the sub-image.

[0115] Subsequently, LLT is used to transform the bounding box of the local spectral support domain into a unit box at the origin of the spatial wavenumber domain, further improving spectral occupancy. The formulas for the three generating vectors of the parallelogram of the local spectral support domain are:

[0116] v1=k2-k1; v2=k4-k3; v3=k7-k5 (29)

[0117] The linear transformations (including scaling, rotation, and shearing transformations) used in the spatial wavenumber domain for the local spectral support domain are incorporated into the linear coordinate transformation matrix T. k In the space wavenumber domain, the LLT operation can be represented as:

[0118] 2πE=T k [v1v2v3] (30)

[0119] Here, E is the identity matrix. Therefore, after SDC and LLT, the spectrum of the sub-image is confined within a cube centered at the origin with sides of length 2π, corresponding to a unit sampling rate in each dimension. T k Therefore, the solution is:

[0120] T k =2π[v1v2v3] -1

[0121] Based on the properties of the Fourier transform, the required local linear coordinate transformation T in the spatial domain is derived. s for:

[0122]

[0123] To implement LLT on sub-images in the spatial domain, a coordinate transformation needs to be established. And locally linearize it into the desired local linear transformation:

[0124]

[0125] According to formula (32), u(x,y,z), v(x,y,z) and n(x,y,z) are the scalar potential functions of the vector fields v1 / (2π), v2 / (2π) and v3 / (2π), respectively, and are solved according to formula (32):

[0126]

[0127]

[0128] The support range of the local spectrum after SDC and LLT is as follows Figure 3 As shown, this is the spectral distribution within the unit frame at the origin in the spatial wavenumber domain. After the above processing, SDC can be performed using formula (25) and sampling can be performed on a uniform sampling grid in the transformed coordinates (u,v,n) (formula (33)) to reconstruct the sub-image of each subarray with the fewest sampling points.

[0129] Based on the above theoretical analysis, such as Figure 3 As shown, the handheld SAR three-dimensional fast near-field imaging method based on local spectrum compression proposed in this embodiment of the invention specifically includes the following steps:

[0130] In step S301, the raw echo data transmitted by the preset handheld SAR system is acquired.

[0131] Specifically, the inconsistencies in amplitude and time delay between different channels of the preset handheld SAR system are corrected, and the amplitude imbalance between different frequencies of each channel of the preset handheld SAR system is corrected to obtain the original echo data s(p′,k).

[0132] In step S302, the composite aperture of the preset handheld SAR system is divided according to the target level to obtain 2 M -1 There are 1 first-level subarrays, and the echo signal of each first-level subarray is extracted from the original echo data, where M is the target level and M≥2.

[0133] Specifically, based on the target level M, the complete synthetic aperture of the pre-defined handheld SAR system is divided into 2. M-1 Each first-level subarray is used, and the corresponding subarray SA is extracted from the original echo data. 1,n The echo data.

[0134] Taking a target level of m=3 as an example, the raw echo data is divided into 4 first-level subarrays, namely SA. 1,1 SA 1,2 SA 1,3 and SA 1,4 And extract the corresponding subarray SA from the original echo data. 1,1 SA 1,2 SA 1,3 and SA 1,4 The echo data.

[0135] In step S303, LLT is used to calculate a uniform sampling grid with a uniform sampling rate for each first-level subarray in the (u,v,n) coordinate system, and the uniform sampling grid is converted into a first-level non-uniform sampling grid in the (x,y,z) coordinate system.

[0136] Specifically, LLT is used to calculate the uniform sampling grid with a uniform sampling rate for each first-level subarray in the (u,v,n) coordinate system. Then, formula (33) is used to transform the uniform sampling grid in the (u,v,n) coordinate system to a non-uniform sampling grid in the (x,y,z) coordinate system, thus obtaining the first-level non-uniform sampling grid G ​​for each first-level subarray. 1,n Taking target level m=3 as an example, the first-level non-uniform sampling grid of each first-level subarray is G. 1,1 G 1,2 G 1,3 and G 1,4 .

[0137] In step S304, the echo signal of each first-level subarray is used to sample 2 pairs of samples in a first-level non-uniform sampling grid. M-1 Reconstructing the first-order subarrays yields 2M-1 A first-level sub-image.

[0138] Specifically, such as Figure 4 As shown, the BPA imaging method is used in the first-level non-uniform sampling grid G ​​of each first-level subarray. 1,n Reconstruct the sub-image f of each subarray 1,n (p), the reconstruction formula is:

[0139]

[0140] The integration region of formula (34) corresponds to the subarray SA. 1,n s(p′,τ) is the echo signal of each first-order subarray, τ is the two-way propagation delay between the array element and the scattering point in the imaging region, p is the coordinate of the scattering point in the imaging region, p′ is the coordinate of the array element in the synthetic aperture, and c is the speed of light.

[0141] As can be seen from formulas (14)-(15), due to the reduction in aperture, the resolution of the sub-image decreases. This reduction makes the resolution of the sub-image f of each sub-array decrease. 1,n (p) First-level non-uniform sampling grid G ​​in each first-level subarray 1,n The sampling rate is reduced, thereby alleviating the overall computational burden of BPA.

[0142] Taking target level M=3 as an example, based on the BPA imaging method, the subarray SA is used. 1,1 The echo data in its first-order non-uniform sampling grid G 1,1 In the middle, the SA of the subarray 1,1 The echo data is reconstructed to obtain the first-level sub-image f. 1,1 (p), using subarray SA 1,2 The echo data in its first-order non-uniform sampling grid G 1,2 In the middle, the SA of the subarray 1,2 The echo data is reconstructed to obtain the first-level sub-image f. 1,2 (p), using subarray SA 1,3 The echo data in its first-order non-uniform sampling grid G 1,3 In the middle, the SA of the subarray 1,3 The echo data is reconstructed to obtain the first-level sub-image f. 1,3 (p), using subarray SA 1,4 The echo data in its first-order non-uniform sampling grid G 1,4 In the middle, the SA of the subarray 1,4 The echo data is reconstructed to obtain the first-level sub-image f. 1,4 (p).

[0143] In step S305, for 2 M-1Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether the at least two second-level sub-images meet the preset conditions. If not, the pairwise coherent superposition process is iteratively executed until the preset conditions are met, and the 3D reconstruction result is obtained.

[0144] In some embodiments, for 2 M-1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether these at least two second-level sub-images meet a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, yielding the 3D reconstruction result, including:

[0145] Calculate 2 M-2 The non-uniform sampling grid corresponding to each secondary sub-image;

[0146] The SDC method is used to perform spatial downconversion on the first-level sub-images of each subarray, resulting in 2 M-1 One compressed sub-image;

[0147] 2 M-1 Each compressed sub-image is interpolated into its corresponding non-uniform sampling grid, and then each compressed sub-image is multiplied by its conjugate space down-conversion function to obtain 2 M-1 One original spectrum sub-image;

[0148] 2 M-1 The original spectral sub-images are coherently superimposed pairwise to obtain 2 M-2 Sub-images;

[0149] Judgment 2 M-2 If a sub-image does not meet a preset condition, a new non-uniform sampling grid is iteratively calculated, and pairwise coherent superposition is performed until the preset condition is met, thus obtaining the 3D reconstruction result. The preset condition is 2. M-2 Whether each sub-image is the final imaging result.

[0150] In some embodiments, the formula for solving the compressed sub-image is:

[0151] f′ m,n (p)=f m,n (p)e -jΦ

[0152]

[0153] Where, f′ m,n (p) is the compressed sub-image of the current level sub-image of the subarray, where m is the current level, 2≤m≤M, and n=2. M-M The number of subarrays at the current level, p is the coordinate of the scattering point within the imaging region, and f is the number of subarrays at the current level. m,n(p) represents the current level sub-image of the subarray, e is the natural logarithm, j is the imaginary unit, Φ is the phase term, and k max k min Let x' be the lower and upper bounds of the millimeter-wave signal wavenumber, r be the sum of the distances between the spatial point p and the four vertices of the subarray, and x' be the upper and lower bounds of the wavenumber. min 、x′ max y′ min y′ max These are the upper and lower limits of the x′ and y′ coordinates of the inner array elements of the integrated aperture, respectively.

[0154] Specifically, such as Figure 4 As shown, first calculate 2 M-2 Two secondary sub-images f 2,n The corresponding non-uniform sampling grid G 2,n The SDC method is used to divide the first-level sub-image f of each subarray. 1,n Performing a spatial downconversion operation yields 2 M-1 A compressed sub-image, containing sub-image f 1,2n-1 and f 1,2n ; 2 M-1 Each compressed sub-image is interpolated to its corresponding non-uniform sampling grid G. 2,n Then, each compressed sub-image is multiplied by its conjugate space downconversion function to obtain 2. M-1 Two original spectral sub-images will be used to... M-1 The original spectral sub-images are coherently superimposed pairwise to form 2 M-2 From the sub-images, we obtain the second-level sub-image f. 2,n .

[0155] Similarly, the imaging result of level m is obtained by coherently superimposing the sub-images of level m-1 pairwise, which is achieved through interpolation. To avoid aliasing during the interpolation process, the sub-images of level m-1 are first subjected to spatial down-conversion using formulas (25) and (28) to convert the sub-images into a compressed form f′. m,2n-1 and f′ m,2n Interpolate these (m-1)th level sub-images with well-defined spectral support ranges to the m-th level sampling grid G. m,n Then, the sub-image is multiplied by its conjugate space downconversion function to restore the spectral position, ensuring the coherence of the sub-image superposition process:

[0156] f m,n (p)=f′ m,n (p)e jΦ (35)

[0157] Subsequently, the processed (m-1)th level sub-images are coherently superimposed using formula (36) to obtain the result in the sampled grid G. m,n The m-th sub-image of the upsampled image.

[0158] f m,n (p)=f m-1,2n-1 (p)+f m-1,2n (p), p∈G m,n (36)

[0159] Repeat this operation to obtain the Mth level imaging result, which is the 3D imaging result of the handheld SAR system.

[0160] Taking target level M=3 as an example, calculate the first-level sub-image f respectively. 1,1 and f 1,2 The sub-image pairs and f 1,3 f 1,4 The resulting sub-image pairs correspond to the non-uniform sampling grid G 2,1 G 2,2 Using the SDC method, each first-level sub-image f 1,1 f 1,2 f 1,3 f 1,4 Perform spatial downconversion to obtain each compressed sub-image f′. 1,1 f′ 1,2 f′ 1,3 f′ 1,4 ; for each compressed sub-image f′ 1,1 f′ 1,2 f′ 1,3 f′ 1,4 Interpolate to its corresponding non-uniform sampling grid G 2,1 G 2,2 Then, each compressed sub-image is multiplied by its conjugate space downconversion function to obtain the corresponding original spectrum sub-image f. 1,1 f 1,2 f 1,3 f 1,4 Then, they are coherently superimposed pairwise to form two sub-images f. 2,1 and f 2,2 ;

[0161] Furthermore, due to the two sub-images f 2,1 and f 2,2 If it is not the final imaging result, then calculate the secondary sub-image f. 2,1 and f 2,2 The resulting sub-image pairs correspond to the non-uniform sampling grid G 3,1 Using the SDC method, each secondary sub-image f 2,1 and f 2,2 Perform spatial downconversion to obtain each compressed sub-image f′. 2,1 f′ 2,2 ; for each compressed sub-image f′ 2,1 f′2,2 Interpolate to its corresponding non-uniform sampling grid G 3,1 Then, each compressed sub-image is multiplied by its conjugate space downconversion function to obtain the corresponding original spectrum sub-image f. 2,1 and f 2,2 Then, the sub-images f are coherently superimposed pairwise. 3,1 Because of sub-image f 3,1 For the final imaging result, the sub-image f 3,1 The result is a 3D reconstruction, which is then output.

[0162] Taking target level M=4 as an example, calculate the first-level sub-image f respectively. 1,1 f 1,2 The sub-image pairs formed, f 1,3 f 1,4 The sub-image pairs formed, f 1,5 f 1,6 The sub-image pairs formed, f 1,7 f 1,8 The resulting sub-image pairs correspond to the non-uniform sampling grid G 2,1 G 2,2 G 2,3 G 2,4 Using the SDC method, each first-level sub-image f 1,1 f 1,2 f 1,3 f 1,4 f 1,5 f 1,5 f 1,6 f 1,7 f 1,8 Perform spatial downconversion to obtain each compressed sub-image f′. 1,1 f′ 1,2 f′ 1,3 f′ 1,4 f′ 1,5 f′ 1,6 f′ 1,7 f′ 1,8 ;

[0163] Each compressed sub-image f′ 1,1 f′ 1,2 f′ 1,3 f′ 1,4 f′ 1,5 f′ 1,6 f′ 1,7 f′ 1,8 Interpolate to its corresponding non-uniform sampling grid G 2,1 G 2,2 G 2,3 G 2,4Then, each compressed sub-image is multiplied by its spatial downconversion function to obtain the corresponding original spectral sub-image f. 1,1 f 1,2 f 1,3 f 1,4 f 1,5 f 1,6 f 1,7 f 1,8 Then, they are coherently superimposed pairwise to form four sub-images f. 2,1 f 2,2 f 2,3 f 2,4 ;

[0164] Furthermore, due to the four sub-images f 2,1 f 2,2 f 2,3 f 2,4 If it is not the final imaging result, then calculate the secondary sub-images f separately. 2,1 f 2,2 The sub-image pairs and f 2,3 f 2,4 The resulting sub-image pairs correspond to the non-uniform sampling grid G 3,1 G 3,2 Using the SDC method, each secondary sub-image f 2,1 f 2,2 f 2,3 f 2,4 Perform spatial downconversion to obtain each compressed sub-image f′. 2,1 f′ 2,2 f′ 2,3 f′ 2,4 ; for each compressed sub-image f′ 2,1 f′ 2,2 f′ 2,3 f′ 2,4 Interpolate to its corresponding non-uniform sampling grid G 3,1 G 3,2 Then, each compressed sub-image is multiplied by its conjugate space downconversion function to obtain the corresponding original spectrum sub-image f. 2,1 f 2,2 f 2,3 f 2,4 Then, they are coherently superimposed pairwise to form two sub-images f. 3,1 and f 3,2 ;

[0165] Furthermore, due to the two sub-images f 3,1 and f 3,2 If it is not the final imaging result, then calculate the third-level sub-image f. 3,1 and f 3,2The resulting sub-image pairs correspond to the non-uniform sampling grid G 4,1 Using the SDC method, each third-level sub-image f 3,1 and f 3,2 Perform spatial downconversion to obtain each compressed sub-image f′. 3,1 f′ 3,2 ; for each compressed sub-image f′ 3,1 f′ 3,2 Interpolate to its corresponding non-uniform sampling grid G 4,1 Then, each compressed sub-image is multiplied by its conjugate space downconversion function to obtain the corresponding original spectrum sub-image f. 3,1 f 3,2 Then, they are coherently superimposed to obtain the sub-image f. 4,1 Because of sub-image f 4,1 For the final imaging result, the sub-image f 4,1 The process of generating and outputting the 3D reconstruction results is the same as when M=3.

[0166] Therefore, as Figure 5 As shown, to address the high computational complexity of traditional BPA time-domain imaging methods, the embodiments of this application achieve accurate and efficient near-field 3D imaging for handheld SAR systems by effectively compressing the spectrum through spatial down-conversion (SDC) and local linear transformation (LLT) of the handheld SAR system. The downsampled sub-images of each subarray are reconstructed using time-domain imaging methods, and subsequent sub-images are obtained through pairwise merging. This achieves accurate and efficient near-field 3D imaging for handheld SAR systems with arbitrary trajectories. Compared to traditional BPA time-domain imaging methods, the computational complexity of the benchmark BPA method is O(N). 5 The order of magnitude is O(N), while the embodiments of this application only require O(N) while ensuring imaging quality. 3 With a computational complexity of logN), it achieves accurate and efficient near-field 3D imaging for handheld SAR systems with arbitrary trajectories.

[0167] In summary, the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression proposed in this invention has the following beneficial effects:

[0168] (1) In view of the characteristics of the spectrum support range of close-range handheld SAR images, the embodiments of the present invention analyze the shape of the wavenumber domain spectrum at different locations in space, and analytically represent the spatial variation characteristics of the spectrum of close-range handheld SAR images, providing theoretical support for reducing the number of sampling points in the imaging interval and thus reducing the computational burden.

[0169] (2) In view of the problem that the spatial downconversion (SDC) and local linear transformation (LLT) operations implemented in spectrum compression require complex numerical pre-computation, the embodiments of the present invention derive analytical expressions for SDC and LLT suitable for handheld SAR imaging systems with constantly changing array structures, thereby avoiding the large amount of time required for numerical pre-computation, and applying them to the spectrum compression of sub-images.

[0170] (3) The embodiments of the present invention solve the problem of image quality degradation caused by approximation in existing handheld SAR imaging methods, and can achieve almost the same image quality as the back projection method (BPA) as the benchmark.

[0171] (4) The embodiments of the present invention solve the problem of high computational complexity of traditional time-domain imaging methods, requiring only O(N) computational complexity while ensuring imaging quality. 3 With a computational complexity of logN), accurate and efficient near-field three-dimensional imaging of a handheld SAR system with arbitrary trajectories can be achieved.

[0172] Next, referring to the accompanying drawings, a handheld SAR three-dimensional fast near-field imaging device based on local spectrum compression according to an embodiment of the present invention is described.

[0173] Figure 6 This is a block diagram of a handheld SAR three-dimensional fast near-field imaging device based on local spectrum compression, according to an embodiment of the present invention.

[0174] like Figure 6 As shown, the handheld SAR three-dimensional fast near-field imaging device 60 based on local spectrum compression includes: an acquisition module 601, a segmentation module 602, a network conversion module 603, a reconstruction module 604, and an iterative overlay module 605.

[0175] The acquisition module 601 is used to acquire the raw echo data of the preset handheld SAR transmitting system. The segmentation module 602 is used to segment the synthetic aperture of the preset handheld SAR system according to the target level, obtaining 2 M-1 The system uses a first-level subarray and extracts the echo signal of each first-level subarray from the original echo data, where M is the target level, M≥2. The network conversion module 603 uses LLT to calculate a uniform sampling grid with a uniform sampling rate for each first-level subarray in the (u,v,n) coordinate system, and converts the uniform sampling grid into a first-level non-uniform sampling grid in the (x,y,z) coordinate system. The reconstruction module 604 uses the echo signal of each first-level subarray in the first-level non-uniform sampling grid to reconstruct the 2... M-1 Reconstructing the first-order subarrays yields 2 M-1 One first-level sub-image. The iterative overlay module 605 is used for iterative overlay of 2... M-1Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether the at least two second-level sub-images meet the preset conditions. If not, the pairwise coherent superposition process is iteratively executed until the preset conditions are met, and the 3D reconstruction result is obtained.

[0176] In some embodiments, the formula for solving the uniform sampling grid is:

[0177]

[0178] Where (u,v,n) is the coordinate system after LLT coordinate transformation, k max k min x′ represents the lower and upper limits of the wavenumber of millimeter-wave signals, respectively. min,SA 、x′ max,SA y′ min,SA and y′ max,SA These represent the upper and lower limits of x′ and y′ for each first-order subarray, where x, y, and z are the coordinates of the scattering points within the imaging region, and x′... 0,SA y′ 0,SA Let x′ and y′ be the closest points to x and y in each first-order subarray, respectively, and r be the sum of the distances between the coordinates of the scattering point p in the imaging region and the four vertices of the subarray. min 、x′ max y′ min y′ max These are the upper and lower limits of the x′ and y′ coordinates of the inner array elements of the integrated aperture, respectively.

[0179] In some embodiments, 2 M-1 The formula for solving each first-order sub-image is:

[0180]

[0181] Among them, f 1,n (p) is the sub-image of each first-level subarray, s(p′,τ) is the echo signal of each first-level subarray, τ is the two-way propagation delay between the array element and the scattering point in the imaging region, p is the coordinate of the scattering point in the imaging region, p′ is the coordinate of the array element in the synthetic aperture, and c is the speed of light.

[0182] In some embodiments, the iterative overlay module includes:

[0183] 2 M-1 Each first-level sub-image is coherently superimposed pairwise to obtain at least two second-level sub-images. It is then determined whether these at least two second-level sub-images meet a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, yielding the 3D reconstruction result, including:

[0184] Calculate 2 M-2 The non-uniform sampling grid corresponding to each secondary sub-image;

[0185] The SDC method is used to perform spatial downconversion on the first-level sub-images of each subarray, resulting in 2 M-1 One compressed sub-image;

[0186] 2 M-1 Each compressed sub-image is interpolated into its corresponding non-uniform sampling grid, and then each compressed sub-image is multiplied by its conjugate space down-conversion function to obtain 2 M-1 One original spectrum sub-image;

[0187] 2 M-1 The original spectral sub-images are coherently superimposed pairwise to obtain 2 M-2 Sub-images;

[0188] Judgment 2 M-2 If a sub-image does not meet the preset conditions, then iteratively calculate 2. M-2 The non-uniform sampling grid corresponding to each sub-image is used, and pairwise coherent superposition is performed until the preset conditions are met to obtain the three-dimensional reconstruction result.

[0189] In some embodiments, the formula for solving the compressed sub-image is:

[0190] f′ m,n (p)=f m,n (p)e -jΦ

[0191]

[0192] Where, f′ m,n (p) is the compressed sub-image of the current level sub-image of the subarray, where m is the current level, 2≤m≤M, and n=2. M-m The number of subarrays at the current level, p is the coordinate of the scattering point within the imaging region, and f is the number of subarrays at the current level. m,n (p) represents the current level sub-image of the subarray, e is the natural logarithm, j is the imaginary unit, Φ is the phase term, and k max k min Let x' be the lower and upper bounds of the millimeter-wave signal wavenumber, r be the sum of the distances between the spatial point p and the four vertices of the subarray, and x' be the upper and lower bounds of the wavenumber. min 、x′ max y′ min y′ max These are the upper and lower limits of the x′ and y′ coordinates of the inner array elements of the integrated aperture, respectively.

[0193] In some embodiments, the preset condition is 2. M-2 Whether each sub-image is the final imaging result.

[0194] It should be noted that the foregoing explanation of the handheld SAR three-dimensional fast near-field imaging method based on local spectrum compression also applies to the handheld SAR three-dimensional fast near-field imaging device based on local spectrum compression in this embodiment, and will not be repeated here.

[0195] The handheld SAR three-dimensional fast near-field imaging device based on local spectrum compression proposed in the embodiments of the present invention has the following beneficial effects:

[0196] (1) In view of the characteristics of the spectrum support range of close-range handheld SAR images, the embodiments of the present invention analyze the shape of the wavenumber domain spectrum at different locations in space, and analytically represent the spatial variation characteristics of the spectrum of close-range handheld SAR images, providing theoretical support for reducing the number of sampling points in the imaging interval and thus reducing the computational burden.

[0197] (2) In view of the problem that the spatial downconversion (SDC) and local linear transformation (LLT) operations implemented in spectrum compression require complex numerical pre-computation, the embodiments of the present invention derive analytical expressions for SDC and LLT suitable for handheld SAR imaging systems with constantly changing array structures, thereby avoiding the large amount of time required for numerical pre-computation, and applying them to the spectrum compression of sub-images.

[0198] (3) The embodiments of the present invention solve the problem of image quality degradation caused by approximation in existing handheld SAR imaging methods, and can achieve almost the same image quality as the back projection method (BPA) as the benchmark.

[0199] (4) The embodiments of the present invention solve the problem of high computational complexity of traditional time-domain imaging methods, requiring only O(N) computational complexity while ensuring imaging quality. 3 With a computational complexity of logN), accurate and efficient near-field three-dimensional imaging of a handheld SAR system with arbitrary trajectories can be achieved.

[0200] Figure 7 This is a schematic diagram of an electronic device provided in an embodiment of the present invention. The electronic device may include:

[0201] The memory 701, the processor 702, and the computer program stored on the memory 701 and executable on the processor 702.

[0202] When the processor 702 executes the program, it implements the handheld SAR three-dimensional fast near-field imaging method based on local spectrum compression provided in the above embodiments.

[0203] Furthermore, electronic devices also include:

[0204] Communication interface 703 is used for communication between memory 701 and processor 702.

[0205] The memory 701 is used to store computer programs that can run on the processor 702.

[0206] The memory 701 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0207] If the memory 701, processor 702, and communication interface 703 are implemented independently, then the communication interface 703, memory 701, and processor 702 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0208] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, then the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0209] The processor 702 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0210] This invention also provides a computer program product, which, when executed by a processor, implements the above-described handheld SAR three-dimensional fast near-field imaging method based on local spectral compression.

[0211] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described handheld SAR three-dimensional fast near-field imaging method based on local spectral compression.

[0212] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0213] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0214] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of preferred embodiments of the invention includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as will be understood by those skilled in the art to which embodiments of the invention pertain.

[0215] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0216] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0217] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0218] Furthermore, the functional units in the various embodiments of the present invention can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0219] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A handheld SAR three-dimensional fast near-field imaging method based on local spectral compression, characterized in that, Includes the following steps: Acquire raw echo data transmitted by a pre-set handheld SAR system; The composite aperture of the preset handheld SAR system is divided according to the target level to obtain... Each first-level subarray is used, and the echo signal of each first-level subarray is extracted from the original echo data, wherein... For the target level, ≥2; LLT is used to compute each first-order subarray in A uniform sampling grid with a uniform sampling rate in the coordinate system is used to transform the uniform sampling grid into a coordinate system. A first-order non-uniform sampling grid in the coordinate system; Using the echo signal of each first-level subarray on the first-level non-uniform sampling grid, the Reconstructing each first-order subarray yields... Each first-level sub-image specifically includes: Based on the BPA imaging method, the echo signals of each first-level subarray are used to reconstruct the image on its first-level non-uniform sampling network. One first-level sub-image; Regarding the Each primary sub-image is coherently superimposed pairwise to obtain at least two secondary sub-images. It is then determined whether these at least two secondary sub-images satisfy a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, resulting in a 3D reconstruction result. The preset condition is... Whether each sub-image is the final imaging result.

2. The handheld SAR three-dimensional fast near-field imaging method based on local spectral compression according to claim 1, characterized in that, The solution formula for the uniform sampling grid is: in, The coordinate system after LLT coordinate transformation. , These are the lower and upper limits of the wavenumber for millimeter-wave signals, respectively. , , and For each first-level subarray Upper and lower limits, , , The coordinates of the scattering points within the imaging region, , The middle and middle of each first-level subarray are respectively and closest and , Coordinates of scattering points within the imaging region The sum of the distances between the subarray and the four vertices. , , , These are the coordinates of the inner elements of the synthetic aperture. Upper and lower limits.

3. The handheld SAR three-dimensional fast near-field imaging method based on local spectral compression according to claim 1, characterized in that, The The formula for solving each first-order sub-image is: in, For each first-level subarray, a sub-image. The echo signal for each first-level subarray, This is the two-way propagation delay between the array element and the scattering point in the imaging region. The coordinates of the scattering points within the imaging region, To synthesize the coordinates of the inner elements of the aperture, It is the speed of light.

4. The handheld SAR three-dimensional fast near-field imaging method based on local spectral compression according to claim 1, characterized in that, The above Each primary sub-image is coherently superimposed pairwise to obtain at least two secondary sub-images. It is then determined whether the at least two secondary sub-images meet a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, resulting in a 3D reconstruction result, including: calculate The non-uniform sampling grid corresponding to each secondary sub-image; The first-level sub-image of each sub-array is spatially down-converted using the SDC method to obtain... One compressed sub-image; The Each compressed sub-image is interpolated into its corresponding non-uniform sampling grid, and then each compressed sub-image is multiplied by its conjugate space down-conversion function to obtain... One original spectrum sub-image; The The original spectral sub-images are coherently superimposed pairwise to obtain Sub-images; Determine the If a sub-image does not meet the preset conditions, a new non-uniform sampling grid is iteratively calculated and pairwise coherent superposition is performed until the preset conditions are met, and the three-dimensional reconstruction result is obtained.

5. The handheld SAR three-dimensional fast near-field imaging method based on local spectral compression according to claim 4, characterized in that, The formula for solving the compressed sub-image is: in, This is a compressed sub-image of the current level sub-image of the sub-array. m For the current level, 2≤ m ≤ M , This represents the number of subarrays at the current level. The coordinates of the scattering points within the imaging region, For the current level sub-image of the sub-array, It is the natural logarithm. The imaginary unit, For phase terms, , These represent the lower and upper limits of the wavenumber for millimeter-wave signals. Let p be the sum of the distances between the spatial point p and the four vertices of the subarray. , , , These are the coordinates of the inner elements of the synthetic aperture. Upper and lower limits.

6. A handheld SAR three-dimensional fast near-field imaging device based on local spectral compression, characterized in that, include: The acquisition module is used to acquire the raw echo data transmitted by the preset handheld SAR system; The segmentation module is used to segment the composite aperture of the preset handheld SAR system according to the target level, and obtain... Each first-level subarray is used, and the echo signal of each first-level subarray is extracted from the original echo data, wherein... For the target level, ≥2; The network transformation module is used to compute the network structure of each first-level subarray using LLT. A uniform sampling grid with a uniform sampling rate in the coordinate system is used to transform the uniform sampling grid into a coordinate system. A first-order non-uniform sampling grid in the coordinate system; The reconstruction module is used to reconstruct the image using the echo signal from each of the first-level subarrays on the first-level non-uniform sampling grid. Reconstructing each first-order subarray yields... Each first-level sub-image specifically includes: Based on the BPA imaging method, the echo signals of each first-level subarray are used to reconstruct the image on its first-level non-uniform sampling network. One first-level sub-image; The iterative overlay module is used to iteratively overlay the... Each primary sub-image is coherently superimposed pairwise to obtain at least two secondary sub-images. It is then determined whether these at least two secondary sub-images satisfy a preset condition. If not, the pairwise coherent superposition process is iteratively executed until the preset condition is met, resulting in a 3D reconstruction result. The preset condition is... Whether each sub-image is the final imaging result.

7. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression as described in any one of claims 1-5.

8. A computer program product, characterized in that, When the computer program / instruction is executed by the processor, it implements the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression as described in any one of claims 1-5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the handheld SAR three-dimensional fast near-field imaging method based on local spectral compression as described in any one of claims 1-5.

Images