Sparse MIMO array near-field imaging method based on distance quantization coding acceleration

Abstract

The invention relates to the technical field of millimeter wave near-field imaging, in particular to a sparse MIMO array near-field imaging method based on distance quantization coding acceleration, and the method comprises the steps: obtaining original echo data; the pre-calculated pulse peak index and the phase compensation item are subjected to compression coding and fixed-point quantization and are pre-stored in a high-speed storage resource; performing oversampling and pulse compression on the echo data; according to the current pixel point and the antenna pair serial number, peak position data and a phase compensation item are extracted from the pre-calculation result for phase compensation, and a phase compensation result is obtained; and traversing all antenna pairs to obtain a reconstruction value of the current pixel point, and traversing all pixel points in the target imaging scene to obtain a final imaging result. Therefore, the problem that time consumption of non-linear calculation in a related BP method is large is solved, operation codes related to non-linear variables can be effectively quantized and pre-stored in high-speed storage resources, and the method is suitable for hardware acceleration.

Description

Technical Field

[0001] The present invention relates to the technical field of millimeter wave near-field imaging, and in particular to a sparse MIMO array near-field imaging method based on range quantization coding acceleration. Background Art

[0002] Millimeter-wave imaging technology based on near-field active arrays can achieve high spatial resolution when realizing large aperture and wide operating bandwidth. Coupled with the non-ionizing characteristics and penetrability of millimeter waves, it has been widely studied and applied in various fields, including security inspection, medical diagnosis, through-wall imaging and non-destructive testing.

[0003] Compared with the dense array of related technologies, MIMO imaging systems have lower system complexity and richer target information due to their sparseness and multi-angle observation advantages. Since multiple antenna radar signals are sent and received simultaneously, the throughput of radar data acquisition can be significantly improved, which has gradually become a hot topic in industry research.

[0004] Although the multi-station structure of the MIMO radar array can reduce the number of physical antenna units, in order to obtain the same imaging quality as the SISO radar array, the antenna channels need to be realized through the diversity of spatial distribution, so the total number of channels after permutation and combination has not been reduced. At the same time, in order to achieve high-resolution radar imaging and avoid aliasing of echo signal sampling, the radar system needs to meet the sampling criteria by increasing the number of sampling points, and the amount of data processed by imaging also increases. The huge amount of data generated by near-field MIMO radar makes the processing speed of digital signals a major difficulty in real-time imaging. Therefore, in order to facilitate engineering implementation, the MIMO system of related technologies usually adopts a uniform and sparse antenna array arrangement. In contrast, the sparse MIMO array has a more flexible and optimized antenna layout design, which can use fewer transmitting antennas and receiving antennas to achieve the same spatial resolution and lower sidelobe levels, and can further reduce the number of antennas required while ensuring imaging accuracy, thereby reducing system costs.

[0005] However, for near-field MIMO radar imaging systems, due to the substantial increase in data transmission throughput, how to achieve fast or even real-time imaging is also worth studying for application requirements such as mobility, integration, and low power consumption. Among them, FPGA (Field Programmable Gate Array) has the advantages of high parallel granularity, flexible hardware design, and ultra-high energy consumption ratio compared with the von Neumann architecture, CPU (Central Processing Unit) and GPU (Graphics Processing Unit) using SIMD (Single Instruction, Multiple Data) instructions. With the rapid development of semiconductor technology, the scale of FPGA logic gates, high-speed interfaces, and the transmission rate of storage media are constantly improving. In addition, the main computing architecture of FPGA is fixed-point data design, which is reflected in the fact that the decimal point position of the output value does not float with the dynamic range of the input value. Therefore, the multiplication and addition logic of fixed-point numbers is simpler, the operation efficiency is higher, and the logic power consumption is lower. It has become possible to use FPGA to implement high-speed digital signal processing. By adopting a sparse MIMO array and combining it with FPGA technology, it is possible to effectively reduce system cost and complexity while improving imaging efficiency and accuracy, providing a new solution for realizing efficient and low-power radar imaging systems. This combination can not only give full play to the advantages of FPGA in high-speed digital signal processing, but also realize fast imaging of sparse MIMO arrays to meet the needs of real-time imaging.

[0006] Among the existing traditional imaging methods, such as the back-projection method (BPA) and Kirchhoff migration method (KMA) in the time domain, there are no constraints on the array form and can be applied to various array topologies to achieve high-precision image reconstruction. However, their computational complexity is too high and it is difficult to achieve real-time imaging on devices such as CPUs. The spatial wavenumber domain imaging method uses FFT (Fast Fourier Transform) to accelerate the calculation of DFT (Discrete Fourier Transform), reducing the computational complexity. However, FFT has requirements for the distribution of sampling points. In the sparse array structure, ultra-wideband design and anti-aliasing processing are required. Therefore, the application scope of some frequency domain imaging methods is limited in near-field scenarios. Summary of the invention

[0007] The present invention provides a sparse MIMO array near-field imaging method based on distance quantization coding acceleration to solve the problems in the related art that the classical BP imaging method has large nonlinear calculation overhead, is not suitable for hardware acceleration, and cannot achieve real-time imaging.

[0008] The first aspect of the present invention provides a sparse MIMO array near-field imaging method based on distance quantization coding acceleration, comprising the following steps: obtaining original echo data of a sparse MIMO array; based on a target imaging scene, utilizing the characteristics of the sparse MIMO array, compressing and quantizing the pulse peak index and phase compensation item corresponding to each pixel in the target imaging scene, and storing the obtained pre-calculated results in a high-speed storage resource; filling the original echo data with zeros to obtain oversampled echo data, and performing pulse compression on the oversampled echo data through a fast inverse Fourier transform to obtain pulse compression data; based on a current pixel point sequence number and a current antenna pair sequence number, extracting a pulse peak position index and a phase compensation item from the pre-calculated results, and taking out the pulse peak position data from the pulse compression data based on the pulse peak index, multiplying the pulse peak position data and the phase compensation item for phase compensation, and obtaining a phase compensation result; based on the phase compensation result, traversing all antenna pairs to obtain a reconstruction value of the current pixel, and based on the reconstruction value, traversing all pixels to obtain the final imaging result of the sparse MIMO array.

[0009] Optionally, in one embodiment of the present invention, the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene are compressed and quantized, and the obtained pre-calculated results are stored in a high-speed storage resource, including: calculating the round-trip propagation distance based on the transmitting antenna coordinates, the receiving antenna coordinates and the grid point sequence number where each pixel point is located, wherein the expression of the round-trip propagation distance is:

[0010] r=||a TX -g||+||a RX -g||=t(tid,gid)+T(rid,gid)

[0011] Among them, g is the vector corresponding to the pixel coordinates of the target imaging area, a TX is the vector corresponding to the transmitting antenna coordinates, a RX is the vector corresponding to the receiving antenna coordinates, ||·|| is the vector amplitude, tid is the grid point number where the transmitting antenna is located, rid is the grid point number where the receiving antenna is located, gid is the grid point number where the imaging pixel is located, and T is the mapping function from the grid point number to the distance value;

[0012] By utilizing the characteristics of the sparse MIMO array and based on the two-way propagation distance, a compressed coded distance term is obtained, wherein the expression of the compressed coded distance term is:

[0013] r=norm(proj_coord-red_coord)

[0014] Among them, ref_coord is the reference point coordinate, proj_coord is the grid point coordinate to be encoded in the projection plane;

[0015] Based on the compressed coded distance term, a compressed coded pulse peak index and a compressed coded real part and an imaginary part of a phase compensation term are obtained.

[0016] Optionally, in one embodiment of the present invention, the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene are compressed and quantized, and the obtained pre-calculated results are stored in a high-speed storage resource, including: based on the dynamic range of the compressed and coded pulse peak index, removing the bias during encoding to obtain the quantization bit width of the compressed and coded pulse peak index after the bias is removed, wherein the calculation expression of the dynamic range is:

[0017]

[0018] Among them, r min is the minimum value of the quantized distance, r max To quantify the maximum value of the distance, R is the distance from the imaging plane to the antenna array plane, and L is the aperture size of the antenna array;

[0019] Based on the simulation results of the peak signal-to-noise ratio PSNR, the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term is determined to obtain the pre-calculated result by combining the quantization bit width of the compressed and encoded pulse peak index after the bias is removed and the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term.

[0020] Optionally, in one embodiment of the present invention, based on the phase compensation result, traversing all antenna pairs to obtain the reconstruction value of the current pixel point, and based on the reconstruction value, traversing all pixel points to obtain the final imaging result of the sparse MIMO array, including: traversing all antenna pairs for the current pixel point, extracting the corresponding pulse peak position index and the phase compensation item from the first-level cache storing the pre-calculated results, and taking out the pulse peak position data from the pulse compression data according to the peak position index, performing phase compensation on the pulse peak position data and the phase compensation item to obtain a phase compensation result; coherently superimposing the phase compensation results of all antenna pairs to obtain the reconstruction value of the current pixel point, and traversing all pixel points in the target imaging scene for iterative calculation until a preset iteration stop condition is reached to obtain the reconstruction value of each pixel point.

[0021] The second aspect of the present invention provides a sparse MIMO array near-field imaging device based on distance quantization coding acceleration, including: an acquisition module for acquiring original echo data of the sparse MIMO array; a pre-calculation module for compressing and quantizing the pulse peak index and phase compensation item corresponding to each pixel in the target imaging scene based on the characteristics of the sparse MIMO array, and storing the obtained pre-calculation results in a high-speed storage resource; a first processing module for padding the original echo data with zeros to obtain oversampled echo data, and performing fast inverse Fourier transform on the oversampled echo data. The data is pulse compressed to obtain pulse compressed data; a second processing module is used to extract the pulse peak position index and the phase compensation item from the pre-calculated result based on the current pixel point sequence number and the current antenna pair sequence number, and take out the pulse peak position data from the pulse compression data based on the pulse peak index, multiply the pulse peak position data and the phase compensation item for phase compensation to obtain a phase compensation result; a reconstruction module is used to traverse all antenna pairs based on the phase compensation result to obtain the reconstruction value of the current pixel point, and traverse all pixel points based on the reconstruction value to obtain the final imaging result of the sparse MIMO array.

[0022] Optionally, in one embodiment of the present invention, the pre-calculation module includes: a compression coding unit, which is used to calculate the round-trip propagation distance based on the transmitting antenna coordinates, the receiving antenna coordinates and the grid point number where each pixel point is located, and obtain the compressed coded distance term based on the round-trip propagation distance by utilizing the characteristics of the sparse MIMO array, and obtain the compressed coded pulse peak index and the compressed coded real part and imaginary part of the phase compensation term based on the compressed coded distance term, wherein the expression of the round-trip propagation distance is:

[0023] r=||a TX -g||+||a RX -g||=T(tid,gid)+T(rid,gid)

[0024] Among them, g is the vector corresponding to the pixel coordinates of the target imaging area, a TX is the vector corresponding to the transmitting antenna coordinates, a RX is the vector corresponding to the receiving antenna coordinates, ||·|| is the vector amplitude, tid is the grid point number where the transmitting antenna is located, rid is the grid point number where the receiving antenna is located, gid is the grid point number where the imaging pixel is located, and T is the mapping function from the grid point number to the distance value;

[0025] The expression of the distance term of the compression coding is:

[0026] r=norm(proj_coord-red_coord)

[0027] Among them, ref_coord is the reference point coordinate, and proj_coord is the grid point coordinate to be encoded in the projection plane.

[0028] Optionally, in one embodiment of the present invention, the pre-calculation module includes: a first fixed-point quantization unit, configured to remove the bias during encoding based on the dynamic range of the compressed encoded pulse peak index to obtain a quantization bit width of the compressed encoded pulse peak index after the bias is removed, wherein the calculation expression of the dynamic range is:

[0029]

[0030] Among them, r min is the minimum value of the quantized distance, r max To quantify the maximum value of the distance, R is the distance from the imaging plane to the antenna array plane, and L is the aperture size of the antenna array;

[0031] The second fixed-point quantization unit is used to determine the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term based on the simulation result of the peak signal-to-noise ratio PSNR, so as to combine the quantization bit width of the compressed and encoded pulse peak index after the bias is removed and the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term to obtain the pre-calculated result.

[0032] Optionally, in one embodiment of the present invention, the reconstruction module includes: an extraction unit, used to traverse all antenna pairs for the current pixel point, extract the corresponding pulse peak position index and the phase compensation item from the first-level cache storing the pre-calculated results, and take out the pulse peak position data from the pulse compression data according to the peak position index, perform phase compensation on the pulse peak position data and the phase compensation item, and obtain a phase compensation result; a calculation unit, used to coherently superimpose the phase compensation results of all antenna pairs to obtain the reconstruction value of the current pixel point, and traverse all pixel points in the target imaging scene for iterative calculation until a preset iteration stop condition is reached to obtain the reconstruction value of each pixel point.

[0033] A third aspect of the present invention provides an electronic device, comprising: 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 sparse MIMO array near-field imaging method based on distance quantization coding acceleration as described in the above embodiment.

[0034] A fourth aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned sparse MIMO array near-field imaging method based on distance quantization coding acceleration.

[0035] A fifth aspect of the present invention provides a computer program product, which, when executed, is used to implement the above-mentioned sparse MIMO array near-field imaging method based on range quantization coding acceleration.

[0036] The embodiment of the present invention compresses and codes the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene and performs fixed-point quantization to obtain a pre-calculation result, and then fills the acquired echo data with zeros to obtain oversampled echo data, performs pulse compression on the oversampled echo data through a fast inverse Fourier transform to obtain pulse compressed data, and extracts peak position data and phase compensation items from the pre-calculation result based on the current pixel point sequence number and the current antenna pair sequence number of the target imaging scene to perform phase compensation to obtain a phase compensation result, thereby traversing all antenna pairs to obtain the reconstruction value of the current pixel point in the target imaging scene, traversing all pixel points in the target imaging scene to obtain the final imaging result of the sparse MIMO array, which can effectively improve the imaging efficiency and accuracy while reducing the system cost and complexity, and realizing real-time imaging. Thus, the problem of large nonlinear calculation time consumption in the related BP method is solved, and the operation code related to the nonlinear variable can be effectively quantized and pre-stored in high-speed storage resources, which is suitable for hardware acceleration.

[0037] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0039] Figure 1 A schematic diagram of the geometric structure of a near-field sparse MIMO imaging system provided by an embodiment of the present invention;

[0040] Figure 2 A schematic diagram of grid quantization of an imaging plane and an array plane provided by an embodiment of the present invention;

[0041] Figure 3 A schematic diagram of distance coding of a projection plane provided by an embodiment of the present invention;

[0042] Figure 4A design flow chart of a sparse MIMO array near-field imaging method based on range quantization coding acceleration provided by an embodiment of the present invention;

[0043] Figure 5 A schematic diagram of the dynamic range of the quantization distance provided by an embodiment of the present invention;

[0044] Figure 6 A schematic diagram of PSNR after quantization of the phase compensation term provided in an embodiment of the present invention;

[0045] Figure 7 A flowchart of a sparse MIMO array near-field imaging method based on range quantization coding acceleration provided by an embodiment of the present invention;

[0046] Figure 8 A schematic diagram comparing the imaging time of the related technology provided by the embodiment of the present invention and the imaging time of the present invention;

[0047] Fig. 9 A schematic diagram of the imaging effect provided by an embodiment of the present invention;

[0048] Fig.10 A block diagram of a sparse MIMO array near-field imaging device based on range quantization coding acceleration provided by an embodiment of the present invention;

[0049] Fig.11 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0050] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0051] The following describes a sparse MIMO array near-field imaging method based on range quantization coding acceleration according to an embodiment of the present invention with reference to the accompanying drawings.

[0052] In order to facilitate those skilled in the art to understand the sparse MIMO array near-field imaging method based on range quantization coding acceleration of the present invention, before introducing the specific implementation methods of the embodiment, the following Table 1 first explains the formula parameters and symbols involved in the embodiment of the present invention, and gives the geometric structure of the near-field sparse MIMO imaging system to analyze the design principle of the BP imaging method of range quantization coding, providing theoretical support for subsequent implementation.

[0053] Specifically, as shown in Table 1, Table 1 is a parameter explanation table.

[0054] Table 1

[0055]

[0056]

[0057] Taking the near-field sparse periodic MIMO array as an example, the geometric structure of its imaging system is as follows: Figure 1 As shown in the figure, the sparse periodic MIMO array shown in the figure has the characteristics of symmetry and equal spacing in the XY plane. The center of the antenna aperture is located at the origin of the Cartesian coordinate system, the normal direction is the positive direction of the z-axis, and the antenna aperture size is L, a TX and a RX is the vector corresponding to the coordinates of the transmitting antenna and the receiving antenna. The imaging target is located in the boundary area of ​​the cube shape in three-dimensional space, and g is the vector corresponding to the coordinates of the pixel point in the target imaging area.

[0058] The millimeter wave signal propagation process satisfies the first-order Born approximation and ignores the propagation attenuation. The echo signal can be expressed as:

[0059]

[0060] Where s represents the echo signal; k = 2πf / c, c is the speed of light; a p It is defined as the combination of the transmitting and receiving antenna unit pairs in the MIMO array; ||·|| represents the magnitude of the vector, ||a TX -g||with||a RX -g|| respectively represent the distances between the transmitting array element and the receiving array element to the scattering point; f(·) represents the reflectivity distribution function of the target, which is also the objective function in the reconstruction process.

[0061] Among the existing image reconstruction methods, BPA is widely regarded as the gold standard due to its excellent imaging quality. By performing correlation operations on the echo signal to achieve the best estimate of the target scattering intensity, it will have the maximum output signal-to-noise ratio under additive white noise, and its formula is:

[0062]

[0063] Among them, the wave number k has an exponential multiplication relationship with the echo signal in the integral, so the transformation relationship between the spatial domain and the spatial wave number domain can be used to perform pulse compression on the spatial frequency domain signal, that is, the inverse Fourier transform. At the same time, in order to make the transformed spatial domain signal have a higher physical resolution in the digital signal, it is necessary to reduce the influence of the "fence effect" by increasing the number of FFT points, that is, to perform zero padding on the measured signal before performing the fast inverse Fourier transform to achieve oversampling in the spatial frequency domain. Therefore, the BPA formula for pulse compression is:

[0064] I(g)=∫S(a p,r)da p #(3)

[0065] Where S represents the variable after the echo signal s is pulse compressed, and r represents the distance of pulse compression. In formula (3), BPA reconstructs the imaging result pixel by pixel by performing spatial wavenumber integration at each position in the imaging area.

[0066] Furthermore, from the steps of the BPA imaging method, it can be seen that there are three nonlinear functions in its implementation process: the square root function is used to calculate the distance r between the coordinates of the transmitting and receiving antennas and the imaging pixel point; the trigonometric function is used to calculate the complex value of the phase compensation term exp(jk c r); the rounding function plays the role of neighboring interpolation and calculates the index of the pulse peak corresponding to r. However, whether in computers or digital circuits, nonlinear calculations are more time-consuming than arithmetic operations. The independent and dependent variables of the nonlinear function are the signal propagation distance r, so consider storing all calculation results related to r and then using the lookup table method to restore them. The distance r is:

[0067] r=||a TX -g||+||a RX -g||#(4)

[0068] Thus, we can calculate the space complexity required to store all r values ​​as N TX N RX N x N y N z , a large storage resource is required to store all the pre-calculated results. Since the computer's storage structure or the on-chip storage logic unit in digital chips such as ASIC / FPGA has a small capacity and low access latency, while the external storage device has a large capacity but high access latency, it is necessary to compress the distance calculation results.

[0069] The embodiment of the present invention maps the norm operation in formula (4) into a function of a finite precision independent variable by quantization. Figure 2 As shown, the three-dimensional space of the imaging scene is sliced ​​along the Z axis, and a two-dimensional XY plane is considered. After rasterization, the transmitting antenna coordinates of the array plane, the receiving antenna coordinates and the pixel coordinates of the imaging plane are placed on equally spaced grid points. At this time, the two-way propagation distance r of the signal can be expressed as a function of the corresponding grid number, as shown in formula (5):

[0070] r=||a TX -g||+||a RX -g||=T(tid,gid)+T(rid,gid)#(5)

[0071] Among them, tid is the number of the grid point where the transmitting antenna is located, rid is the number of the grid point where the receiving antenna is located, gid is the number of the grid point where the imaging pixel is located, and T is the mapping function from the grid point number to the distance value.

[0072] Furthermore, after grid quantization, the two-way propagation distance is mapped to a set of function values ​​of the independent variables tid, rid, and gid, but the total amount of data in the lookup table is not reduced. TX and r RX Since the sparse periodic MIMO array has the characteristics of symmetry and equal spacing in the XY plane, there are distances r of the same length TX and r RX Therefore, there is redundancy in the output set of the mapping function, and the total amount of data in the lookup table can be reduced by compression coding.

[0073] It is understandable that if Figure 3 As shown in , consider projecting the imaging plane onto the plane where the antenna aperture is located. Since the same grid is used for quantization, vectors with equal projected lengths are also equal in length before projection. According to the Pythagorean theorem, a right triangle with equal right-angled sides also has equal hypotenuse lengths. Therefore, the size of all propagation vector lengths can be represented by a two-element array consisting of the lengths of the two right-angled sides, Figure 3 The two propagation vectors in the projection plane are projected onto the X-axis and the Y-axis respectively, and the quantized and encoded binary arrays (2, 3) and (2, 1) are obtained according to the projection lengths of the X-axis and the Y-axis.

[0074] Furthermore, after the quantization and compression coding operations, all binary arrays form the index of the two-dimensional array lookup table. The storage depth of the lookup table variable is the total number N of quantized grids. x N y In three-dimensional imaging, an additional storage dimension N is required z . The space complexity of storing variables at this time is calculated to be N x N y N z , reducing the dimension N of the number of antenna pairs TX N RX , the amount of data is reduced by 2 orders of magnitude.

[0075] It is worth noting that Figure 4As can be seen from the flowchart, the present invention only changes the calculation method of the pulse peak index and phase compensation in the traditional BP imaging method in the related art, so it still has good imaging accuracy in theory. After pulse compression processing, the process of cyclic calculation of pulse peak and phase compensation is improved to the index value calculation of the lookup table, and the data stored in the lookup table is the pre-calculated result: pulse peak index rng_idx and phase compensation term cor_phi. Therefore, the nonlinear function of the BP imaging method in the related art is converted into a multiplication and addition index value calculation, which greatly reduces the calculation time.

[0076] In order to arrange the pre-calculated quantized code values ​​regularly and simplify the decoding logic, the vertices of the imaging plane boundary are selected as reference points, and it is assumed that the imaging area is consistent with the size of the antenna aperture, then:

[0077]

[0078] proj_coord=vec3(Gx,Gy,0)#(7)

[0079] r=norm(proj_coord-ref_coord)#(8)

[0080] Among them, R is the distance from the imaging plane to the antenna array plane; G is a two-dimensional array composed of quantized grid points, Gx, Gy are the two-dimensional coordinates of the quantized grid; r is the distance that needs to be quantized and encoded, also r tX and r RX Projection on the array plane; ref_coord is the reference point coordinate, proj_coord is the grid point coordinate to be encoded in the projection plane (array plane); the real and imaginary parts of the phase compensation term and the index of the peak after pulse compression can be expressed as:

[0081] cor_phi_i[Gx,Gy]=cos(k c *r)#(9)

[0082] cor_phi_q[Gx,Gy]=sin(k c *r)#(10)

[0083] rng_index[Gx,Gy]=round(r / dr)#(11)

[0084] Among them, k c is the wave number corresponding to the carrier frequency, and dr is the size of the distance unit after pulse compression.

[0085] In addition, different data types have different computing efficiencies on different hardware platforms. The main computing architecture of the GPU platform is single-precision floating-point design, while the application of fixed-point data types on the FPGA side is more extensive, mainly reflected in the fact that the decimal point position of the output value does not float with the dynamic range of the input value. Therefore, the multiplication and addition logic of fixed-point numbers is simpler, the operation efficiency is higher, and the logic power consumption is lower. However, the dynamic range of fixed-point data types is small and the precision is limited. Therefore, the quantization bit width of fixed-point data needs to be carefully designed to make the error of the output result within an acceptable range, while shortening the bit width of the data as much as possible. Among them, the specific fixed-point quantization design of the pulse peak index and phase compensation items is as follows:

[0086] For fixed-point quantization of the pulse peak index variable, we first need to determine the bit width of rng_index after fixed-point quantization. Since the size of rng_index is proportional to the quantization distance r, the dynamic range of rng_index can be expressed by the dynamic range of r. Figure 5 As shown, here taking 25 sampling grid points as an example, the minimum value of r is at G[0,0] and the maximum value is at G[4,4]. The expression of the dynamic range Δr of the variable r is shown in formula (12). When the imaging distance R is smaller, the dynamic range of the quantization distance r is larger. Therefore, when the imaging area distance term range is 0.3m to 1m, the maximum dynamic range of the variable rng_index appears at 0.3m closest to the antenna plane.

[0087]

[0088] Furthermore, combined with Figure 5 The diagram of the dynamic range of the quantization distance when R=0.3m is shown. The dynamic range of the pulse peak index variable rng_index obtained after precalculation under the 51*51 grid size is 24 to 36. There is a bias of 24. The actual required dynamic range of quantization is 12, and they are all positive integers, so unsigned fixed-point quantization is used. And the bias 24 is removed during encoding and then stored, and the bias value is restored during decoding. Using this method, the rng_index variable can be stored with a data bit width of 4 bits, while 6 bits are required before debiasing. However, this quantization method also has the problem of different bias values ​​in different imaging planes, so it is necessary to increase the step of updating the bias value when performing three-dimensional imaging.

[0089] For the quantization of the phase compensation variable cor_phi, since the error of the phase compensation term is mainly reflected in the phase of the accumulated value, and the phase error cannot be constrained by specific indicators, the simulation results are used for evaluation. The phase compensation terms with a width of 1 to 16 decimal places are traversed and simulated, and the control variable method is used to compare the impact of the phase compensation term on the imaging results. In order to quantitatively describe the quality of the imaging results, the peak signal-to-noise ratio (PSNR) is used as an evaluation indicator. If an 8-bit image is used to quantify the calculation results of the MIMO-BP method, the calculation formula of PSNR is as follows:

[0090]

[0091] Among them, MSE (Mean Squared Error) is the mean square error, mn is the number of pixels in the imaging area, I is the imaging result before quantization, and K is the imaging result after the phase compensation term is quantized.

[0092] In addition, combined Figure 6 The PSNR result after quantization of the phase compensation term obtained by the simulation test is shown in the figure. When the decimal width of the quantization is greater than 5 bits, the PSNR value remains basically unchanged and is close to the theoretical upper limit. Therefore, a 5-bit decimal width is selected to quantize the phase compensation term, which can minimize the data width while ensuring the quality of the imaging result.

[0093] Based on the above theoretical analysis, Figure 7 As shown, the sparse MIMO array near-field imaging method based on distance quantization coding acceleration proposed in an embodiment of the present invention specifically includes the following steps:

[0094] In step S701, original echo data of a sparse MIMO array is acquired.

[0095] Specifically, the original echo data of the sparse MIMO array is obtained, and the amplitude and delay inconsistencies between different channel signals are corrected, and the amplitude imbalance between different frequencies of each channel signal is corrected to obtain s(a p , k), wherein the millimeter wave signal propagation process satisfies the first-order Born approximation and the propagation attenuation is ignored. Optionally, in one embodiment of the present invention, the expression of the echo data is:

[0096]

[0097] Among them, a p is the combination of the transmitting and receiving antenna unit pairs in the MIMO array, s(·) is the echo signal, k = 2πf / c, c is the speed of light, f is the frequency, k is the wave number, f(·) is the target reflectivity distribution function, g is the vector corresponding to the pixel coordinates of the target imaging area, and aTX is the vector corresponding to the transmitting antenna coordinates, a RX is the vector corresponding to the receiving antenna coordinates, and ||·|| is the vector amplitude.

[0098] In step S702, based on the target imaging scene and utilizing the characteristics of the sparse MIMO array, the pulse peak index and phase compensation item corresponding to each pixel in the target imaging scene are compressed and quantized, and the pre-calculated results are stored in a high-speed storage resource.

[0099] Specifically, the embodiment of the present invention can pre-calculate the pulse peak index and the phase compensation term according to the imaging scene, including:

[0100] Compression coding: Based on the symmetric and equally spaced characteristics of the sparse periodic MIMO array in the XY plane, the redundancy of the mapping function output set is reduced. The imaging plane is projected onto the plane where the antenna aperture is located, and the distance term is compressed and encoded according to formulas (5) to (8). Then, the compression coding of the pulse peak index rng_index and the real part cor_phi_i and imaginary part cor_phi_q of the phase compensation term is obtained using formulas (9) and (11).

[0101] Fixed-point quantization: The dynamic range of the pulse peak index variable rng_index is obtained according to formula (12), and the bias is removed during encoding to reduce the data storage bit width. The simulation results of PSNR are used to determine the fractional quantization bit width of the real part cor_phi_i and the imaginary part cor_phi_q of the phase compensation term, so as to minimize the data bit width while ensuring the quality of the imaging result.

[0102] Through the above compression coding and fixed-point quantization operations, the results of the pulse peak index and the phase compensation item are pre-stored in the high-speed storage resources of the hardware device (such as the first-level cache BRAM of the FPGA) to reduce the delay of subsequent reading.

[0103] In step S703, zero padding is performed on the original echo data to obtain oversampled echo data, and pulse compression is performed on the oversampled echo data by fast inverse Fourier transform to obtain pulse compression data.

[0104] Specifically, in order to make the transformed spatial domain signal have a higher physical resolution in the digital signal, it is necessary to reduce the influence of the "fence effect" by increasing the number of IFFT points, that is, to perform zero padding on the measured signal before performing the fast inverse Fourier transform to achieve oversampling in the spatial frequency domain. p ,k) Perform fast inverse Fourier transform after zero padding:

[0105]

[0106] In step S704, based on the current pixel point number and the current antenna pair number, the pulse peak position index and the phase compensation item are extracted from the pre-calculated results, and based on the pulse peak index, the pulse peak position data is taken out from the pulse compression data, and the pulse peak position data and the phase compensation item are multiplied for phase compensation to obtain the phase compensation result.

[0107] Specifically, the embodiment of the present invention can retrieve the peak position data rng_index and the phase compensation item cor_phi (including the real part cor_phi_i and the imaginary part cor_phi_q) from the first-level cache BRAM storing the pre-calculated data according to the current pixel point sequence number (xid and yid of gid) and the current antenna pair sequence number (rid and tid), where:

[0108] cor_phi=exp(j*k c *r)=cor_phi_i+j*cor_phi_q#(15)

[0109] In step S705, based on the phase compensation result, all antenna pairs are traversed to obtain the reconstruction value of the current pixel point, and based on the reconstruction value, all pixel points are traversed to obtain the final imaging result of the sparse MIMO array.

[0110] Specifically, an embodiment of the present invention can traverse all antenna pairs for the current pixel point, extract the corresponding pulse peak position index and phase compensation item from the first-level cache storing the pre-calculated results, and take out the pulse peak position data from the pulse compression data according to the peak position index, perform phase compensation on the pulse peak position data and the phase compensation item to obtain a phase compensation result; coherently superimpose the phase compensation results of all antenna pairs to obtain the reconstructed value of the current pixel point, and traverse all pixel points in the target imaging scene for iterative calculation until a preset iteration stop condition is reached to obtain the reconstructed value of each pixel point.

[0111] It can be understood that the preset iteration stop condition may be whether the last pixel point is traversed.

[0112] Therefore, in order to solve the problem of large nonlinear calculation time consumption in related BP methods, the embodiment of the present invention innovatively changes the calculation method of phase compensation term and pulse peak index in the traditional BP method by compressing and coding the variables related to nonlinear operations and quantizing them, and pre-storing them in the high-speed storage resources (such as the first-level cache BRAM, Block Random Access Memory) on the FPGA side of the device. And the BP graphics reconstruction of distance quantization coding is completed on the FPGA side, and the parallel processing capability and high-speed interface of FPGA are utilized, so that the imaging speed of this method is much higher than that of traditional BP under the premise of small image distortion and high energy efficiency ratio, which effectively solves the problem of large nonlinear calculation overhead in the classic BP method, and has important significance for the design and implementation of future miniaturized low-power near-field radar imaging systems.

[0113] In summary, the sparse MIMO array near-field imaging method based on distance quantization coding acceleration proposed in an embodiment of the present invention has the following beneficial effects:

[0114] (1) In order to solve the problem that the calculation of three nonlinear functions in related BP imaging methods is time-consuming, the present invention innovatively encodes and quantizes the independent variables and dependent variables of the nonlinear functions, and pre-stores them in the high-speed storage resources (such as the first-level cache BRAM) on the FPGA side, thereby changing the calculation method of the phase compensation term and the pulse peak index in the traditional BP method without introducing additional approximations. Therefore, it has high computational efficiency while still maintaining good imaging accuracy.

[0115] (2) Due to the symmetric and equally spaced characteristics of the sparse periodic MIMO array in the two-dimensional imaging plane, the distance coding of the mapping function is redundant. Through compression coding and fixed-point quantization, the storage resource occupancy is reduced, making it more suitable for hardware acceleration and further reducing the calculation time.

[0116] (3) Aiming at the real-time imaging requirements in near-field MIMO radar imaging systems, the present invention innovatively completes the BP image reconstruction of distance quantization coding on the FPGA side. By utilizing the parallel processing capability and high-speed interface of FPGA, the imaging speed is significantly improved, and the problem of large nonlinear calculation overhead in related BP methods is solved under the premise of small image distortion and high energy efficiency ratio, which is of great significance to the design and implementation of future miniaturized, low-power near-field radar imaging systems.

[0117] Furthermore, combined with Figure 8 and Fig. 9 As shown, the embodiment of the present invention can obtain a comparison of MATLAB simulation results of different algorithms, as shown in Table 2, which is a PSNR table of imaging time comparison of different algorithms and imaging results at different pixel points.

[0118] Table 2

[0119]

[0120] According to the sparse MIMO array near-field imaging method based on distance quantization coding acceleration proposed in an embodiment of the present invention, the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene are compressed and quantized at fixed points to obtain a pre-calculated result, and then the acquired echo data is zero-filled to obtain oversampled echo data, and pulse compression is performed on the oversampled echo data through a fast inverse Fourier transform to obtain pulse compressed data, and based on the current pixel point sequence number and the current antenna pair sequence number of the target imaging scene, the peak position data and the phase compensation item are extracted from the pre-calculated result to perform phase compensation to obtain a phase compensation result, so as to traverse all antenna pairs to obtain the reconstruction value of the current pixel point in the target imaging scene, and traverse all pixel points in the target imaging scene to obtain the final imaging result of the sparse MIMO array, which can effectively improve the imaging efficiency and accuracy while effectively reducing the system cost and complexity, and realizing real-time imaging. Therefore, the problem of large nonlinear calculation time consumption in the related BP method is solved, and the operation code related to the nonlinear variable can be effectively quantized and pre-stored in high-speed storage resources, which is suitable for hardware acceleration.

[0121] Next, a sparse MIMO array near-field imaging device based on range quantization coding acceleration proposed in an embodiment of the present invention will be described with reference to the accompanying drawings.

[0122] Fig.10 It is a block diagram of a sparse MIMO array near-field imaging device based on range quantization coding acceleration according to an embodiment of the present invention.

[0123] like Fig.10 As shown, the sparse MIMO array near-field imaging device 10 based on range quantization coding acceleration includes: an acquisition module 100 , a pre-calculation module 200 , a first processing module 300 , a second processing module 400 and a reconstruction module 500 .

[0124] Specifically, the acquisition module 100 is used to acquire the original echo data of the sparse MIMO array;

[0125] The pre-calculation module 200 is used to compress and quantize the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene based on the target imaging scene and using the characteristics of the sparse MIMO array, and store the obtained pre-calculation results in a high-speed storage resource;

[0126] The first processing module 300 is used to fill zeros on the original echo data to obtain oversampled echo data, and perform pulse compression on the oversampled echo data by fast inverse Fourier transform to obtain pulse compression data;

[0127] The second processing module 400 is used to extract the pulse peak position index and the phase compensation item from the pre-calculated result based on the current pixel point sequence number and the current antenna pair sequence number, and take out the pulse peak position data from the pulse compression data based on the pulse peak index, multiply the pulse peak position data and the phase compensation item to perform phase compensation, and obtain a phase compensation result;

[0128] The reconstruction module 500 is used to traverse all antenna pairs to obtain the reconstruction value of the current pixel based on the phase compensation result, and traverse all pixel points to obtain the final imaging result of the sparse MIMO array based on the reconstruction value.

[0129] Optionally, in one embodiment of the present invention, the pre-calculation module 200 includes: a compression encoding unit.

[0130] The compression coding unit is used to calculate the round-trip propagation distance based on the transmitting antenna coordinates, the receiving antenna coordinates and the grid point number where each pixel point is located, and obtain the compression coded distance term based on the round-trip propagation distance by utilizing the characteristics of the sparse MIMO array, and obtain the compression coded pulse peak index and the real part and imaginary part of the compression coded phase compensation term based on the compression coded distance term, wherein the expression of the round-trip propagation distance is:

[0131] r=||a TX -g||+||a RX -g||=T(tid,gid)+T(rid,gid)

[0132] Among them, g is the vector corresponding to the pixel coordinates of the target imaging area, a TX is the vector corresponding to the transmitting antenna coordinates, a RX is the vector corresponding to the receiving antenna coordinates, ||·|| is the vector amplitude, tid is the grid point number where the transmitting antenna is located, rid is the grid point number where the receiving antenna is located, gid is the grid point number where the imaging pixel is located, and T is the mapping function from the grid point number to the distance value;

[0133] The expression of the distance term of the compressed encoding is:

[0134] r=norm(proj_coord-ref_coord)

[0135] Among them, ref_coord is the reference point coordinate, and proj_coord is the grid point coordinate to be encoded in the projection plane.

[0136] Optionally, in one embodiment of the present invention, the pre-calculation module 200 includes: a first fixed-point quantization unit and a second fixed-point quantization unit.

[0137] The first fixed-point quantization unit is used to remove the bias during encoding based on the dynamic range of the pulse peak index after compression coding to obtain the quantization bit width of the pulse peak index after compression coding after the bias is removed, wherein the calculation expression of the dynamic range is:

[0138]

[0139] Among them, r min is the minimum value of the quantized distance, r max To quantify the maximum value of the distance, R is the distance from the imaging plane to the antenna array plane, and L is the aperture size of the antenna array;

[0140] The second fixed-point quantization unit is used to determine the fractional quantization bit width of the real and imaginary parts of the phase compensation term after compression coding based on the simulation result of the peak signal-to-noise ratio PSNR, so as to obtain the pre-calculated result by combining the quantization bit width of the pulse peak index after compression coding after removing the bias and the fractional quantization bit width of the real and imaginary parts of the phase compensation term after compression coding.

[0141] Optionally, in one embodiment of the present invention, the reconstruction module 500 includes: an extraction unit and a calculation unit.

[0142] Among them, the extraction unit is used to traverse all antenna pairs for the current pixel point, extract the corresponding pulse peak position index and phase compensation item from the first-level cache storing the pre-calculated results, and take out the pulse peak position data from the pulse compression data according to the peak position index, perform phase compensation on the pulse peak position data and the phase compensation item, and obtain the phase compensation result;

[0143] The calculation unit is used to coherently superimpose the phase compensation results of all antenna pairs to obtain the reconstruction value of the current pixel point, and traverse all pixel points in the target imaging scene for iterative calculation until a preset iteration stop condition is reached to obtain the reconstruction value of each pixel point.

[0144] It should be noted that the aforementioned explanation of the embodiment of the sparse MIMO array near-field imaging method based on distance quantization coding acceleration is also applicable to the sparse MIMO array near-field imaging device based on distance quantization coding acceleration of this embodiment, and will not be repeated here.

[0145] According to the sparse MIMO array near-field imaging device based on distance quantization coding acceleration proposed in an embodiment of the present invention, the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene are compressed and quantized at fixed points to obtain a pre-calculated result, and then the acquired echo data is padded with zeros to obtain oversampled echo data, and the oversampled echo data is pulse compressed by fast inverse Fourier transform to obtain pulse compressed data, and based on the current pixel point sequence number and the current antenna pair sequence number of the target imaging scene, the peak position data and the phase compensation item are extracted from the pre-calculated result to perform phase compensation to obtain a phase compensation result, so as to traverse all antenna pairs to obtain the reconstruction value of the current pixel point in the target imaging scene, and traverse all pixel points in the target imaging scene to obtain the final imaging result of the sparse MIMO array, which can effectively improve the imaging efficiency and accuracy while effectively reducing the system cost and complexity, and realizing real-time imaging. Therefore, the problem of large nonlinear calculation time consumption in the related BP method is solved, and the operation code related to the nonlinear variable can be effectively quantized and pre-stored in high-speed storage resources, which is suitable for hardware acceleration.

[0146] Fig.11 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. The electronic device may include:

[0147] A memory 1101 , a processor 1102 , and a computer program stored in the memory 1101 and executable on the processor 1102 .

[0148] When the processor 1102 executes the program, the sparse MIMO array near-field imaging method based on range quantization coding acceleration provided in the above embodiment is implemented.

[0149] Furthermore, the electronic device further comprises:

[0150] The communication interface 1103 is used for communication between the memory 1101 and the processor 1102 .

[0151] The memory 1101 is used to store computer programs that can be executed on the processor 1102 .

[0152] The memory 1101 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0153] If the memory 1101, the processor 1102 and the communication interface 1103 are implemented independently, the communication interface 1103, the memory 1101 and the processor 1102 can be connected to each other through a bus and communicate with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Fig.11 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0154] Optionally, in a specific implementation, if the memory 1101, the processor 1102 and the communication interface 1103 are integrated on a chip, the memory 1101, the processor 1102 and the communication interface 1103 can communicate with each other through an internal interface.

[0155] The processor 1102 may be a Field Programmable Gate Array (FPGA), or an Application Specific Integrated Circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention.

[0156] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned sparse MIMO array near-field imaging method based on range quantization coding acceleration.

[0157] An embodiment of the present invention further provides a computer program product, which can run computer instructions. When the computer instructions are executed by a processor, the above-mentioned sparse MIMO array near-field imaging method based on range quantization coding acceleration is implemented.

[0158] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0159] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0160] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

[0161] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or N wirings (electronic devices), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways as necessary and then storing it in a computer memory.

[0162] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiment, the N steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented by hardware, as in another embodiment, it can be implemented by any one or a combination of multiple of the following technologies known in the art: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0163] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0164] In addition, each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0165] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A sparse MIMO array near-field imaging method based on range quantization coding acceleration, characterized in that: The method comprises the following steps: Get the raw echo data of the sparse MIMO array; Based on the target imaging scene, utilizing the characteristics of the sparse MIMO array, compressing and quantizing the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene, and storing the obtained pre-calculated results in a high-speed storage resource; Filling the original echo data with zeros to obtain oversampled echo data, and performing pulse compression on the oversampled echo data by fast inverse Fourier transform to obtain pulse compression data; Based on the current pixel point sequence number and the current antenna pair sequence number, a pulse peak position index and a phase compensation item are extracted from the pre-calculated result, and based on the pulse peak index, pulse peak position data is taken out from the pulse compression data, and the pulse peak position data and the phase compensation item are multiplied to perform phase compensation to obtain a phase compensation result; Based on the phase compensation result, all antenna pairs are traversed to obtain the reconstruction value of the current pixel point, and based on the reconstruction value, all pixel points are traversed to obtain the final imaging result of the sparse MIMO array.

2. The method according to claim 1, characterized in that The method of compressing and quantizing the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene, and storing the obtained pre-calculated results in a high-speed storage resource, includes: Based on the transmitting antenna coordinates, the receiving antenna coordinates and the grid point number where each pixel point is located, the two-way propagation distance is calculated, wherein the expression of the two-way propagation distance is: r=||a TX -g||+||a RX -g||=T(tid,gid)+T(rid,gid) Among them, g is the vector corresponding to the pixel coordinates of the target imaging area, a TX is the vector corresponding to the transmitting antenna coordinates, a RX is the vector corresponding to the receiving antenna coordinates, ||·|| is the vector amplitude, tid is the grid point number where the transmitting antenna is located, rid is the grid point number where the receiving antenna is located, gid is the grid point number where the imaging pixel is located, and T is the mapping function from the grid point number to the distance value; By utilizing the characteristics of the sparse MIMO array and based on the two-way propagation distance, a compressed coded distance term is obtained, wherein the expression of the compressed coded distance term is: r=norm(proj_coord-ref_coord) Among them, ref_coord is the reference point coordinate, proj_coord is the grid point coordinate to be encoded in the projection plane; Based on the compressed coded distance term, a compressed coded pulse peak index and a compressed coded real part and an imaginary part of a phase compensation term are obtained.

3. The method according to claim 1, characterized in that The pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene are compressed and quantized, and the obtained pre-calculated results are stored in a high-speed storage resource, including: Based on the dynamic range of the pulse peak index after compression coding, the bias is removed during coding to obtain the quantization bit width of the pulse peak index after compression coding after the bias is removed, wherein the calculation expression of the dynamic range is: Among them, r min is the minimum value of the quantized distance, r max To quantify the maximum value of the distance, R is the distance from the imaging plane to the antenna array plane, and L is the aperture size of the antenna array; Based on the simulation results of the peak signal-to-noise ratio PSNR, the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term is determined to obtain the pre-calculated result by combining the quantization bit width of the compressed and encoded pulse peak index after the bias is removed and the fractional quantization bit width of the real and imaginary parts of the compressed and encoded phase compensation term.

4. The method according to claim 1, characterized in that: The step of traversing all antenna pairs to obtain a reconstruction value of the current pixel based on the phase compensation result, and traversing all pixel points to obtain a final imaging result of the sparse MIMO array based on the reconstruction value, includes: Traversing all antenna pairs for the current pixel point, extracting the corresponding pulse peak position index and the phase compensation item from the first-level cache storing the pre-calculated result, taking out the pulse peak position data from the pulse compression data according to the peak position index, performing phase compensation on the pulse peak position data and the phase compensation item, and obtaining a phase compensation result; The phase compensation results of all the antenna pairs are coherently superimposed to obtain the reconstruction value of the current pixel point, and all the pixel points in the target imaging scene are traversed for iterative calculation until a preset iteration stop condition is reached to obtain the reconstruction value of each pixel point.

5. A sparse MIMO array near-field imaging device based on range quantization coding acceleration, characterized in that: The device comprises: An acquisition module, used for acquiring original echo data of a sparse MIMO array; A pre-calculation module, for compressing and quantizing the pulse peak index and phase compensation item corresponding to each pixel point in the target imaging scene based on the target imaging scene and utilizing the characteristics of the sparse MIMO array, and storing the obtained pre-calculation results in a high-speed storage resource; A first processing module, used for padding the original echo data with zeros to obtain oversampled echo data, and performing pulse compression on the oversampled echo data by fast inverse Fourier transform to obtain pulse compression data; A second processing module is used to extract a pulse peak position index and a phase compensation item from the pre-calculated result based on the current pixel point sequence number and the current antenna pair sequence number, and take out the pulse peak position data from the pulse compression data based on the pulse peak index, multiply the pulse peak position data and the phase compensation item to perform phase compensation, and obtain a phase compensation result; A reconstruction module is used to traverse all antenna pairs to obtain a reconstruction value of the current pixel based on the phase compensation result, and to traverse all pixel points to obtain a final imaging result of the sparse MIMO array based on the reconstruction value.

6. An electronic device, characterized in that: include: 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 sparse MIMO array near-field imaging method based on range quantization coding acceleration as described in any one of claims 1 to 4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the sparse MIMO array near-field imaging method based on range quantization coding acceleration as described in any one of claims 1 to 4.

8. A computer program product, comprising a computer program, characterized in that The computer program is executed to implement the sparse MIMO array near-field imaging method based on range quantization coding acceleration as described in any one of claims 1 to 4.

Images