A SAR baseband echo fast generation method based on bilinear interpolation surface target template

Abstract

The application discloses a kind of based on bilinear interpolation face target template SAR baseband echo fast generation method, belong to radar technical field.The application includes: by dividing face target into N×M rectangular grid, each rectangle represents a point target, according to the various parameters of airborne strip SAR radar to determine the number of rectangular grid in radar irradiation scene and the target template of radar irradiation scene, and establish the slant range difference storage table between radar irradiation scene center and other point targets, according to target template using bilinear interpolation algorithm adjustment face target's backscattering coefficient matrix format, finally on the basis of one-dimensional fast fourier transform echo generation method to the time delay caused by target slant range is approximately handled and the echo phase produced by slant range is accurately calculated, obtain SAR baseband echo.Compared with existing concentric circle algorithm, the present application method greatly reduces the SAR echo generation calculation amount, and the calculation efficiency is improved by more than 55%.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar, and particularly relates to a SAR baseband echo fast generation method based on a bilinear interpolation surface target template. BACKGROUND

[0002] Synthetic Aperture Radar (SAR) is a kind of high-resolution microwave imaging radar, which is not affected by clouds, night and extreme weather conditions, can work all day and all year round, and has certain penetration capability, can identify disguised equipment and buildings, and is currently widely used in military information reconnaissance, target detection and other fields, and is also widely used in geographical terrain mapping, resource exploration and other fields in civil aspects. In the SAR imaging algorithm research, system scheme design and verification test, a large amount of SAR raw echo data under specific conditions is needed, however, the echo data collected on site by the SAR system cannot meet the research and experimental needs under specific conditions, and it is also difficult to improve the reliability and stability of the system, and the cost is high, therefore, generating a large amount of echo data through SAR echo simulation is an important development direction of SAR radar.

[0003] SAR echo simulation as an important part of the SAR system faces some problems, when a large scene with large amplitude and high resolution needs to be simulated, the calculation amount is usually very large, and the time cost also increases dramatically, and the operation efficiency is the main problem to be solved. At present, there are two main methods for fast SAR echo simulation, one is to optimize the algorithm from the algorithm point of view, and improve the efficiency of the algorithm, and the other is to use a high-efficiency hardware processing platform for parallel acceleration processing. When the scene is very large, the number of SAR echo data points will be very large, and the calculation amount will increase dramatically, and the use of a hardware platform for acceleration processing can improve the calculation speed to a certain extent, but due to the large amount of calculation data, it is still difficult to complete the calculation quickly, therefore, the algorithm needs to be optimized to improve the calculation speed. SUMMARY

[0004] In view of the problems in the prior art, the purpose of the present application is to provide a SAR baseband echo fast generation method based on a bilinear interpolation surface target template, which does not need to calculate the slant range of each target in real time, only needs to calculate the slant range of the scene center in real time, and then calculates the slant range of other targets in combination with the slant range difference lookup table, thereby greatly reducing the calculation amount.

[0005] In order to achieve the above purpose, the following technical scheme is adopted, a SAR baseband echo fast generation method based on a bilinear interpolation surface target template, comprising the following steps:

[0006] Step 1, establish the airborne SAR radar stop geometry mode, set the SAR radar transmitting signal as a linear frequency modulation signal, and set the SAR radar platform parameters such as working frequency, signal bandwidth, signal pulse width, pulse repetition frequency, airborne platform height, platform speed, antenna size, target scene distance and azimuth size, minimum distance of target scene to radar platform, etc.

[0007] Step 2, determine the target grid template in the radar illuminated scene according to the SAR radar platform parameters, the azimuth length of the target grid is the distance of the pulse repetition period time of the integer multiple Nslow_N of the flight of the airborne SAR; calculate the slant range difference of the center target and other targets in the radar illuminated scene in Nslow_N pulse repetition periods, and store it in the slant range difference lookup table.

[0008] Step 3, the backscattering coefficient matrix format of the surface target is preprocessed by using the bilinear interpolation algorithm, and the target represented by the processed backscattering coefficient matrix is matched with the target grid template in step 2.

[0009] Step 4, on the basis of the improved fast Fourier transform echo generation method, the superposition system response function of all targets in a distance unit in the scene at a certain azimuth time is calculated.

[0010] Step 5, the SAR radar transmitting signal at each azimuth time is convolved with the scene target superposition system response function calculated in step 4 to obtain the target baseband echo signal, and all azimuth times are traversed to obtain the total target baseband echo matrix of the airborne SAR;

[0011] The beneficial effects of the present application are:

[0012] In view of the two problems that the slant range of each target needs to be calculated in real time in the simulation process of the SAR radar signal echo, and the approximate processing is adopted in the echo phase calculation generated by the slant range in the traditional time domain superposition echo simulation and the most commonly used concentric circle algorithm simulation, the present application establishes a scene target grid template, and the grid template can be adaptively adjusted according to the SAR radar parameters, the slant range difference of the center target and other targets in the radar illuminated scene is calculated in advance according to the established grid template and a storage table is established, the slant range of each target does not need to be calculated in real time in the subsequent echo simulation process, and the approximate processing is not used in the echo phase calculation generated by the slant range, which greatly improves the echo simulation efficiency, greatly shortens the simulation time, is more efficient and flexible when different scenes need to be simulated and analyzed, and the simulation scene can also be more complex and detailed. BRIEF DESCRIPTION OF DRAWINGS

[0013] Figure 1 It is a flowchart of the fast echo generation method of the present application;

[0014] Figure 2 SAR space geometry schematic diagram;

[0015] Figure 3 Radar irradiation scene slant range plane target template geometric relationship schematic diagram;

[0016] Figure 4 Radar irradiation scene slant range plane geometric relationship schematic diagram at different azimuth moments;

[0017] Figure 5 Radar irradiation scene slant range plane geometric relationship schematic diagram of the same distance unit;

[0018] Figure 6 Bilinear interpolation algorithm schematic diagram;

[0019] Figure 7 Real SAR image needing to be simulated for echo simulation; (a) is a simulation scene of 512m*512m, (b) is a simulation scene of 640m*640m, and (c) is a simulation scene of 1024m*1024m;

[0020] Figure 8 Simulation image using a concentric circle algorithm to generate and image echoes; (a) is a simulation scene of 512m*512m, (b) is a simulation scene of 640m*640m, and (c) is a simulation scene of 1024m*1024m;

[0021] Figure 9 Simulation image using the method of the present application to generate and image echoes; (a) is a simulation scene of 512m*512m, (b) is a simulation scene of 640m*640m, and (c) is a simulation scene of 1024m*1024m. DETAILED DESCRIPTION

[0022] In order to facilitate those skilled in the art to understand the technical content of the application, the content of the application is further explained below in combination with the drawings and specific embodiments.

[0023] Reference Figure 1 The present application provides a SAR baseband echo fast generation method based on a bilinear interpolation surface target template, including the following steps:

[0024] Step 1, establish an airborne SAR radar walk-stop geometry mode, and set the SAR radar platform related parameters:

[0025] Reference Figure 2 Establish an airborne SAR radar walk-stop geometry model, and set the radar parameters such as pulse repetition period PRI, airborne SAR flight speed V, and antenna azimuth size D, etc.

[0026] Step 2: Determine the target grid template in the radar illumination scene using SAR radar platform parameters, calculate the slant range difference between the central target and other targets in the radar illumination scene, and establish a slant range difference storage table.

[0027] Based on the SAR radar platform parameters, the following is established: Figure 3 Target mesh template:

[0028] (1a) Calculate the length dx of the target grid in the azimuth direction based on the pulse repetition period PRI, the airborne SAR flight speed V, and the antenna azimuth dimension D:

[0029]

[0030] dx = dxx·Nslow_N (16)

[0031] Where ceil(·) represents rounding up, c is the speed of light, Nslow_N represents the azimuth length of one grid required for the airborne SAR radar to fly, and dxx is the airborne SAR flight distance within each pulse repetition period.

[0032] (1b) Number of target grids M in the azimuth direction of the area illuminated by the computer-borne SAR radar at each azimuth time:

[0033]

[0034] Where Lsar is the synthetic aperture length of the airborne SAR radar, and floor(·) indicates rounding down.

[0035] (1c) Calculate the length dy of the target grid in the range direction and the number N of the target grid in the range direction of the area illuminated by the airborne SAR radar at each azimuth time:

[0036]

[0037] Where c is the speed of light, bw is the bandwidth of the transmitted signal, and L is the distance length of the simulated scene.

[0038] (1d) According to the appendix Figure 4 As shown in the radar illumination area at different PRI times, the slant range difference matrix between the airborne SAR radar and the point target in the radar illumination area has a cycle period of Nslow_N. That is, the slant range difference matrix between the airborne SAR radar and the target in the illumination area at the nth azimuth time is the same as the slant range difference matrix of the target at the (n+Nslow_N)th azimuth time. Therefore, the slant range difference lookup table only needs to store the slant range difference matrix between the target in the SAR radar illumination area and the central target within the Nslow_Nth azimuth time.

[0039] (1e) Let ta For a certain moment of Nslow_N azimuth moments, calculate the slant range difference between the scene center target and other targets at this moment:

[0040]

[0041] Where H is the flight height of the airborne SAR radar, R min represents the minimum plane distance from the simulation scene to the radar platform, represents the upward rounding operation, R(p,q) represents the t a slant range between the target with grid coordinates (p,q) irradiated by the radar at the moment and the radar platform, dR(p,q) represents the t a slant range difference between the scene center target and the target with grid coordinates (p,q) irradiated by the radar at the moment, p=1,2,...,N, q=1,2,...,M.

[0042] (1f) Traverse Nslow_N azimuth moments, calculate the slant range difference between the scene center target and other targets at each azimuth moment, and store the slant range difference in the slant range difference lookup table in order.

[0043] Step 3, using bilinear interpolation algorithm to preprocess the backscattering coefficient matrix format of the surface target, matching the target represented by the processed backscattering coefficient matrix with the target grid template.

[0044] (2a) Since the target grid size is adjusted adaptively according to the SAR radar parameters, the target size represented by the backscattering coefficient matrix of the echo simulation scene may not match the target grid size, so the bilinear interpolation algorithm is used to preprocess the backscattering coefficient matrix of the simulation scene.

[0045] (2b) Bilinear interpolation method is the most commonly used algorithm for image scaling at present, as shown in the attached Figure 5 Bilinear interpolation algorithm is to calculate one point in new image from four points in original image. This algorithm calculates a total of three times of single linear interpolation in two directions. First, find the corresponding point coordinates T(x,y) in the original image according to a point in the new image and the scaling factor:

[0046]

[0047] Where dstX represents the x coordinate of the pixel point in the target image, dstY represents the y coordinate of the pixel point in the target image, srcwidth represents the width of the original image, dstwidth represents the width of the target image, srcheight represents the height of the original image, and dstheight represents the height of the target image.

[0048] Then the coordinates of the four nearest points corresponding to the points in the original image are:

[0049] Q 11 (x1,y1), Q 12 (x1,y2), Q 21 (x2,y1), Q 22 (x2,y2), wherein represents a down rounding operation, x2=x1+1, y2=y1+1.

[0050] For the coordinates of the four adjacent points, first perform quadratic single linear interpolation in the x direction to obtain two temporary point pixel values R1(x, y1) and R2(x, y2), and then perform linear single interpolation in the y direction to obtain the pixel value of the point T(x, y):

[0051]

[0052] wherein f(·) represents the pixel value represented by the point, and the pixel value f(T) of the point T(x, y) to be solved is obtained by integrating the above three formulas:

[0053]

[0054] (2c) According to the target size represented by the backscattering coefficient matrix of the scene, the size of the backscattering coefficient matrix may be adjusted after the ratio of the target grid size, and the pre-processed backscattering coefficient matrix of the scene is calculated by traversing each scene target.

[0055] Step 4, improve the fast Fourier transform echo generation method, calculate the superposition system response function of all targets in a certain distance unit in the scene at a certain azimuth moment, that is, an expression containing target echo delay, amplitude and phase information:

[0056] (3a) According to the fast Fourier transform echo generation method, the signal received by the radar is the result of convolution of the transmitted signal with an impulse response function having a certain amplitude, phase, and delay:

[0057]

[0058] In the formula, t r is the distance fast time, t m is the azimuth slow time, R pq (t m ) represents the distance from the (p, q)th grid target to the radar platform at t m , a(t r )exp(jπk r t 2 ) is the transmitted baseband signal, represents convolution of t r , and j represents the imaginary unit.

[0059] (3b) Set up the radar receiver to sample at frequency F s Discrete sampling is performed, and the corresponding sampling interval is calculated as c represents the speed of light;

[0060] (3c) Since the echo signal is discretized, the time delay position of any point target is not always exactly at the sampling point in the calculation of formula (28), so an approximation of the time delay caused by the slant range is needed. Let the time delay of the target at position (p, q) be τ (p, q), and the slant range be R (p, q), and the time delay of the target at position (p, q) can be expressed as m The imaging scene is divided into different distance units with Δr as the interval between adjacent distance units, as shown in FIG. 2. Figure 6 As shown in FIG. 2, the targets located in the same equidistance ring will be divided into the same distance unit, and the slant range of the radar at the current sampling time position and the center target of the illuminated scene is calculated; the slant range of other targets in the illuminated scene is calculated according to the target slant range difference lookup table calculated in step 2, that is, the slant range matrix R (p, q) of all targets in the illuminated scene at the current time is obtained, and the targets are divided into distance units, and the energy of all targets in the same distance unit is finally added together;

[0061] (3d) According to the target slant range matrix R (p, q) calculated in (3c), the gray value of the target corresponding to the real SAR image point is read as the target backscattering coefficient, and the total echo information of P k points in the kth distance unit is expressed as:

[0062]

[0063] In the formula, t r is the range fast time, R (a p,q ) represents the distance from the a th point target in the k th distance unit (i.e., the point target located at the coordinate (p, q) ) to the phase center of the radar, represents the backscattering coefficient value of the a th point target after preprocessing the scene target backscattering coefficient matrix, λ is the carrier wavelength, and δ (·) is the impulse function, represents the echo time delay information of the point target; represents the echo phase caused by the slant range, and j represents the imaginary unit.

[0064] Step 5: Convolve the SAR radar transmit baseband signal at a certain azimuth time with the total target superposition system response function of the SAR radar illuminated scene at the azimuth time to obtain the target baseband echo signal, and traverse all azimuth times to obtain the total target baseband echo matrix of the airborne SAR;

[0065] (4a) The total target baseband echo matrix of the airborne SAR is convolved with the total target superposition system response function of the SAR radar illuminated scene at the azimuth time t mThe target echo information of all range cells in the scene is superimposed to obtain the total target echo information S(t) at that azimuth and time. r );

[0066]

[0067] In the formula, t r For range-time intervals, ne represents the number of range cells in the SAR radar illumination scene at that azimuth time, and r... max This represents the maximum slant range between the SAR radar illumination scene and the radar platform at that azimuth time, r. min This represents the minimum slant range between the SAR radar illumination scene and the radar platform at that azimuth time, where Δr is the interval between adjacent range cells.

[0068] (4b) Perform Fast Fourier Transform (FFT) on the total target echo information and the transmitted baseband linear frequency modulated signal, respectively. Then multiply the two transform results and perform an inverse Fast Fourier Transform on the product to obtain the airborne SAR radar signal at t. m The total target baseband echo signal S at azimuth time r (t r ):

[0069] S r (t r ) = IFFT{FFT(s(t) r ))*FFT(S t (t r ))} (32)

[0070] In the formula, S t (t r ) represents the baseband linear frequency modulated signal transmitted by the SAR radar.

[0071] (4c) By iterating through all azimuth sampling times and arranging the calculated target baseband echo signals in azimuth-time order, the total target baseband echo matrix S of the airborne SAR can be obtained. r (t r ,t m ).

[0072] Simulation Experiment

[0073] To demonstrate the effectiveness of this invention, the following simulation comparison experiments are used for further illustration.

[0074] (1) Simulation conditions:

[0075] This invention is a pure software simulation, and the experimental platform is Matlab R2022b.

[0076] The transmitted signal uses a linear frequency modulated (LFM) signal. The transmitted signal parameters and geometric model parameters are shown in Table 1 below:

[0077] Table 1 Simulation Parameters

[0078]

[0079] (2) Simulation results and analysis:

[0080] The attached image uses pre-prepared real SAR images. Figure 7 As imaging simulation scenarios, Figures (a), (b), and (c) correspond to simulation scenarios of different sizes. For ease of comparison, firstly, using the parameters in Table 1, the concentric circle algorithm is used to simulate the overall SAR process under interference-free conditions. Then, under the same conditions, the method of this invention is used to simulate the overall SAR process. The results of SAR echo generation using the concentric circle algorithm and imaging using the RD algorithm are attached. Figure 8 As shown in the attached figure, the results of SAR echo generation and imaging using the RD algorithm according to the method of the present invention are as follows. Figure 9 As shown, both images can basically characterize the scattering characteristics of the actual ground scene, and the similarity between the two images obtained by the corr2 function in Matlab is shown in Table 2:

[0081] Table 2. Similarity of echo imaging images generated by the two methods.

[0082] Simulation scene size Similarity of echo imaging images generated by two methods (%) 512m*512m 97.1％ 640m*640m 98.2％ 1024m*1024m 97.6％

[0083] Therefore, it can be determined that the images generated by the two methods after imaging are basically the same, verifying the correctness of the method of the present invention. In order to verify the improvement of the SAR echo simulation calculation efficiency of the method of the present invention, five tests were conducted on the concentric circle algorithm and the method of the present invention respectively, and the time was statistically analyzed, as shown in Tables 3, 4 and 5.

[0084] Table 3 Comparison of simulation times for the two methods when the simulation scenario is 512m*512m.

[0085]

[0086] Table 4 Comparison of simulation times for the two methods when the simulation scenario is 640m*640m

[0087]

[0088] Table 5 Comparison of simulation times for the two methods when the simulation scenario is 1024m*1024m

[0089]

[0090] In summary, the above simulation experiment proves the reliability and effectiveness of the application. The echo error of the fast echo generation provided by the application is small, and good imaging can be completed. Meanwhile, the target grid template established according to the application can calculate part of the parameters of the target slant range in advance and store them in a lookup table, thereby reducing the calculation redundancy in subsequent calculation and greatly reducing the calculation amount. The application is especially suitable for echo generation of large scene surface targets and overcomes the difficulty of large calculation amount of the time domain superposition algorithm in traditional echo generation.

Claims

1. A SAR baseband echo fast generation method based on a bilinear interpolation plane target template, characterized in that, The method comprises the following steps: Step 1, dividing a surface target into N*M rectangular grids according to SAR radar platform parameters, setting the rectangular grids as a target grid template of a radar irradiation scene, each rectangle representing a point target, and determining the size of the grids in the target grid template, the number of the rectangular grids in the radar irradiation scene, and establishing a slant range difference storage table of a center of the radar irradiation scene and other point targets; The SAR radar platform parameters comprise a working frequency, a signal bandwidth, a signal pulse width, a pulse repetition frequency, an airborne platform height, a platform speed, an antenna size, a target scene distance and azimuth size, and a minimum distance of the target scene to the radar platform; Step 2, pre-processing a backscattering coefficient matrix format of the target by using a bilinear interpolation algorithm, so that the target size represented by a backscattering coefficient matrix of a simulation scene of the echo is matched with the target grid size; Step 3, approximately processing a time delay caused by a distance between the target and the radar by using an improved method of a fast Fourier transform echo generation method, and accurately calculating an echo phase caused by the distance, to obtain a system response function of all targets in a distance unit in the scene at a certain azimuth time; (3a) setting up a radar receiver to sample at a frequency F s Discrete samples are taken, with the corresponding sampling interval being c denotes the speed of light; (3b) Set a certain orientation sampling time t m The imaging scene is divided into different distance units with distance unit interval of Δr, and the slant range of the radar and the center target of the illuminated scene is calculated; the slant ranges of other targets in the illuminated scene are calculated according to the target slant range difference look-up table, and the targets are divided into corresponding distance units. (3c) According to the slant range matrix R(p, q) of all targets in the illuminated scene calculated in (3b), read the gray value of the corresponding real SAR image point of the target as the backscattering coefficient of the target, and superimpose the echo information of P k point targets in the kth range cell as: where t r is the fast time, (p, q) denotes the target position coordinates, p = 1, 2,..., N, q = 1, 2,..., M, R(a p,q ) denotes the slant range from the kth range cell and ath point target to the radar, denotes the floor operation, denotes the backscattering coefficient value of the ath point target after the preprocessing of the scene target backscattering coefficient matrix, λ is the carrier wavelength, and δ(·) is the impulse function, denotes the point target echo time delay information; denotes the echo phase generated by the slant range, and j denotes the imaginary unit. Step 4, convolving a SAR radar baseband signal at each azimuth time with the system response function of the targets in all distance units in the scene to obtain a target baseband echo signal, and traversing all azimuth times to obtain a total target baseband echo matrix of the airborne SAR.

2. The SAR baseband echo fast generation method based on bilinear interpolation surface target template according to claim 1, characterized in that, The specific setting method of the target grid template in step 1 is as follows: (1a) calculating a flight distance of each pulse repetition period time of the SAR platform as dxx=V·PRI, wherein V represents a flight speed of the airborne SAR, and PRI represents a pulse repetition period of a radar transmission signal; calculating an azimuth length of a grid in the target grid template according to the SAR radar parameter antenna azimuth size D and the flight distance dxx: dx=dxx·Nslow_N (4) wherein ceil(·) represents upward rounding, c is a light speed, bw is a transmission signal bandwidth, Nslow_N represents a multiple of the pulse repetition period time required for the airborne SAR radar to fly an azimuth length of a grid, dx is the azimuth length of the grid in the target grid template, dy is a distance length of the grid, N represents a target grid number of the target grid in a distance of an irradiation area of the airborne SAR radar at each azimuth time, and L is a simulation scene distance length; (1b) calculating a target grid number M of the irradiation area of the airborne SAR radar in the azimuth at each azimuth time: wherein Lsar is a synthetic aperture length of the airborne SAR radar, and floor(·) represents downward rounding; (1c) the slant range difference matrix of the point target of the radar irradiation area is taken as a loop period Nslow_N, and the slant range difference matrix of the target of the SAR radar irradiation area and the center target in Nslow_N azimuth times is stored as a slant range difference lookup table.

3. The SAR baseband echo fast generation method based on bilinear interpolation surface target template according to claim 1, characterized in that, The specific method of the pre-processing is as follows: First, finding a corresponding point coordinate T(x, y) in an original image according to a certain point of a target image and a scaling multiple: x = (dstX + 0.5) * (srcwidth / dstwidth) - 0.5 (9) y = (dstY + 0.5) * (srcheight / dstheight) - 0.5 (10) Wherein, dstX represents the pixel point x coordinate in the target image, dstY represents the pixel point y coordinate in the target image, srcwidth represents the original image width, dstwidth represents the target image width, srcheight represents the original image height, dstheight represents the target image height; The coordinates of the four points closest to the corresponding point of the original image are: Q 11 (x1,y1), Q 12 (x1,y2), Q 21 (x2,y1), Q 22 (x2,y2), wherein denotes a floor operation, x2=x1+1, y2=y1+1; Using the coordinates of the four adjacent points, quadratic single linear interpolation is first performed in the x direction to obtain R1(x, y1) and R2(x, y2), two temporary point pixel values, and then one single linear interpolation is performed in the y direction to obtain the pixel value of the T(x, y) point: Wherein f(·) represents the pixel value represented by the point, and the pixel value f(T) of the point T(x, y) is obtained by integrating the above three formulas:

4. The SAR baseband echo fast generation method based on bilinear interpolation surface target template according to claim 2, characterized in that, The calculation method of the slant distance difference in step 1 is: Let t a For a certain moment of Nslow_N azimuthal moments, calculate the slant range difference between the scene center target and other targets: where H is the flight height of the airborne SAR radar, R min denotes the minimum planar distance from the radar platform to the simulation scene, denotes the upward rounding operation, R(p,q) denotes the t a denotes the slant range of the target with grid coordinates (p,q) in the scene irradiated by the radar at time t, dR(p,q) denotes the t a denotes the slant range difference between the center target of the scene irradiated by the radar and the target with grid coordinates (p,q) at time t, p=1,2,...,N, q=1,2,...,M.

Images