A ground moving target refocusing method without parameter search

Abstract

This invention provides a ground moving target refocusing method without parameter search. Based on the moving target echo data received by radar, a range-frequency difference function is constructed in the range-frequency domain that requires no parameter estimation. By multiplying the echo data by the conjugate of this range-frequency difference function, the coupling between the moving target's range frequency and slow time can be eliminated, achieving the goal of refocusing the moving target's energy. Since the range-frequency difference reference function is derived entirely from the actual echo and requires no additional parameters, it effectively solves the complex range envelope migration and Doppler defocusing problems caused by higher-order phase coupling. Even with platform motion errors, moving target Doppler center blurring, and Doppler spectrum splitting, this invention can directly achieve robust correction of moving target range migration. Furthermore, this ground moving target refocusing method without parameter search achieves moving target refocusing without parameter estimation and search processing, resulting in high computational efficiency.

Description

Technical Field

[0001] This invention belongs to the field of moving target detection technology, specifically relating to a ground moving target refocusing method that does not require parameter search. Background Technology

[0002] Synthetic Aperture Radar (SAR), with its all-weather observation capabilities, has been widely used in various fields such as topographic mapping, maritime surveillance, environmental and disaster monitoring, and battlefield reconnaissance. Moving target detection, as one of the core tasks of radar detection, plays a vital role in battlefield situational awareness and traffic control. Therefore, the ability to obtain real-time motion information of moving targets is important in both military and civilian applications.

[0003] In radar look-down detection missions, the radar receives echo signals from weak moving targets superimposed with background clutter. Due to platform movement, the Doppler spectrum of the background clutter becomes significantly wider. After transforming the received data to the range-Doppler domain, the moving target echo signal will overlap with the clutter, making it impossible to directly detect the moving target.

[0004] While methods to suppress background clutter can preserve moving target information and improve the signal-to-noise ratio, making moving targets easier to detect, SAR radar imaging requires a long synthetic aperture time to achieve higher two-dimensional resolution. During this time, the complex relative motion between the moving target and the platform leads to range migration and Doppler spectral spread. Furthermore, when imaging with fixed scene parameters, the moving target inevitably defocuses due to mismatched compensation functions, resulting in a significant decrease in the target's peak energy and hindering its detection.

[0005] To address the energy focusing problem of moving targets in moving target detection, range migration (RCM) correction is necessary. Most correction methods typically use signal envelope autocorrelation characteristics and minimum entropy criteria, which perform well only under high signal-to-noise ratio conditions and are difficult to apply in practice.

[0006] Under low signal-to-noise ratio (SNR) conditions, the Keystone transform (KT) method can be used to achieve first-order range migration correction, but it cannot correct higher-order motion. While a second-order Keystone transform method can be used to correct second-order moving target motion, both methods require fuzzy search operations when the moving target experiences Doppler center blurring or Doppler spectrum splitting, leading to a significant increase in computational complexity. Furthermore, these methods cannot correct range migration caused by higher-order motion, resulting in residual range migration that is detrimental to target refocusing.

[0007] After appropriate correction for range migration, in order to accurately refocus the moving target, it is necessary to estimate the motion parameters of the moving target. Methods for estimating moving target parameters are mainly divided into two categories: linear time-frequency transformation methods and nonlinear time-frequency transformation methods.

[0008] The main linear time-frequency transformation methods include maximum likelihood transform (ML), fractional Fourier transform (FRFT), and discrete Fourier transform (DCFT), but these methods all require full-dimensional search and have high computational complexity.

[0009] Nonlinear time-frequency transformation methods mainly include Higher-Order Ambiguity Function (HAF), Discrete Polynomial Phase Transform (DPT), Wigner-Ville Distribution (WVD), and Scaled Fourier Transform (SCFT). These methods encounter the cross terms of polynomial moving target signals, thus affecting the estimation of moving target parameters in multi-moving-target scenarios. Under low signal-to-noise ratio conditions, these methods struggle to obtain moving target parameters.

[0010] As can be seen from the above, obtaining accurate parameters of a moving target in practical applications is quite difficult, and the methods used to obtain these parameters are prone to significant errors, which can lead to a decline in the performance of moving target focusing methods that rely on parameter estimation. Therefore, there are also methods that can achieve moving target focusing without knowing the moving target parameters. These methods include two-dimensional frequency domain matched filter algorithms; DerampKT (DKT) algorithms; and Time Reversal Transform (TRT).

[0011] Two-dimensional frequency domain matched filter algorithms construct specific matched filters in the two-dimensional frequency domain to refocus moving targets without estimating motion parameters. However, this fixed-phase method suffers from significant errors when the time-bandwidth product (TBP) of the moving target is small, limiting its applicability to signals with large TBPs. Furthermore, significant errors in the filter function can retain a large amount of residual RCM, hindering moving target refocusing. The DerampKT algorithm constructs an azimuth deramp function in the range-frequency domain using static scene information to compensate for range curvature and Doppler modulation, and uses the KT algorithm to correct the RCM before focusing the moving target in the azimuth frequency domain. Both of these methods neglect the azimuth motion of the moving target, thus failing to compensate for higher-order phase effects caused by complex moving target motion, inevitably leading to defocusing. To address these issues, the TRT method directly constructs a compensation function in the time-reversed slow time domain to compensate for the second-order Doppler phase of the moving target. Although the TRT method considers the azimuth velocity of the moving target, Doppler shift still occurs when the moving target has higher-order motion quantities such as acceleration. Furthermore, since both the DerampKT algorithm and the TRT method utilize KT for preprocessing to correct the moving target's motion, the spread of the range envelope becomes more pronounced in high-resolution radar when the platform has motion errors or the moving target exhibits higher-order motion, making effective correction of range migration impossible.

[0012] Existing techniques for distance migration correction only perform well under high signal-to-noise ratio (SNR) conditions and fail under low SNR conditions. Although some methods, such as the Keystone transform (KT), can achieve correction under low SNR conditions, they can only handle first-order distance migration correction and cannot correct distance migration caused by higher-order motion. Algorithms for higher-order distance migration correction all require fuzzy search operations, which significantly increases computational complexity. Furthermore, these methods cannot correct distance migration caused by higher-order motion and may produce distance migration residuals that are detrimental to focusing on moving targets.

[0013] After range migration correction, motion parameters of the moving target need to be estimated to achieve moving target refocusing. Existing techniques for estimating motion parameters of moving targets include algorithms such as linear time-frequency transform methods, which require full-dimensional search and have high computational complexity. Nonlinear time-frequency transform methods encounter the cross terms of polynomial moving target signals, making them unable to handle motion parameter estimation in scenarios with multiple moving targets. Furthermore, both methods struggle to obtain motion target parameters under low signal-to-noise ratio conditions.

[0014] There are also moving target refocusing methods that do not require parameter estimation, such as the two-dimensional frequency domain matched filter algorithm, the DerampKT algorithm, and the TRT algorithm. The two-dimensional frequency domain matched filter algorithm is only suitable for signals with a large TBP and requires a highly accurate filter function; otherwise, it will produce a large amount of residual RCM. The DerampKT algorithm cannot compensate for the high-order phase effects caused by the complex motion of the moving target, inevitably resulting in defocusing of the moving target. When using the TRT method, when the moving target has high-order motion quantities such as acceleration, Doppler frequency shift will occur, making it impossible to achieve moving target refocusing. Summary of the Invention

[0015] To address the aforementioned problems in the existing technology, this invention provides a ground moving target refocusing method that does not require parameter search. The technical problem to be solved by this invention is achieved through the following technical solution:

[0016] This invention provides a ground moving target refocusing method without parameter search, comprising:

[0017] S100: Receives echo signals from moving targets via a radar platform and performs range-direction pulse compression on the echo signals to obtain a pulse-compressed signal.

[0018] S200, the pulse-compressed signal is converted to the distance-frequency domain to obtain the distance-frequency domain compressed signal;

[0019] S300, construct a distance-frequency difference function based on the distance compression signal;

[0020] S400, based on the distance-frequency difference function and the distance compression signal in the distance-frequency domain, the signal after moving target correction is obtained;

[0021] S500 performs an inverse Fourier transform on the signal after correction of the moving target in the range direction, and then performs a Fourier transform on the azimuth direction to obtain the refocused image data of the moving target.

[0022] Beneficial effects:

[0023] 1. The ground moving target refocusing method proposed in this invention, which requires no parameter search, constructs a range-frequency difference function in the range-frequency domain based on the moving target echo data received by radar. By multiplying the echo data by the conjugate of the range-frequency difference function, the coupling between the moving target's range frequency and slow time can be eliminated, achieving the purpose of refocusing the moving target's energy. Since the range-frequency difference reference function is entirely derived from the actual echo and requires no additional parameters, it can effectively solve the complex range envelope migration and Doppler defocusing problems caused by higher-order phase coupling.

[0024] 2. Unlike traditional kt-based moving target focusing methods, the ground moving target refocusing method proposed in this invention, which does not require parameter search, can directly achieve robust correction of moving target range migration even in the presence of platform motion errors, moving target Doppler center blurring, and Doppler spectrum splitting.

[0025] 3. Compared with traditional methods, the ground moving target refocusing method of the present invention, which does not require parameter search, can achieve moving target refocusing without parameter estimation and search processing, and has high computational efficiency.

[0026] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of a ground moving target refocusing method without parameter search provided by the present invention;

[0028] Figure 2 This is a schematic flowchart of the radar-electro-optical integrated method for detecting and tracking slow-moving, small targets provided by the present invention.

[0029] Figure 3 This is a simulation result diagram of detecting a slow-moving target provided by the present invention;

[0030] Figure 4 This is a simulation result diagram of detecting fast-moving targets provided by the present invention;

[0031] Figure 5 These are refocusing effect diagrams of three different moving targets provided by this invention;

[0032] Figure 6 These are simulation results of various algorithms for the slow-moving target 1 provided by this invention;

[0033] Figure 7 These are simulation results of various algorithms provided by this invention for moving targets 2 with large speeds at a distance;

[0034] Figure 8 These are simulation results of the algorithms for the fast-moving target 3 provided by this invention; Detailed Implementation

[0035] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0036] This invention relates to a single-channel synthetic aperture radar (SAR-GMTI) ground moving target detection system. It performs range-direction pulse compression on the received moving target echo signal, converts the compressed signal to the range-frequency domain, and constructs specific range-frequency difference functions for slow-moving and fast-moving targets based on the range-frequency domain data. These range-frequency difference functions are then multiplied by the moving target echo to achieve range migration correction. Simultaneously, by setting an appropriate frequency offset, the azimuth signal can be effectively represented as a small linear frequency modulated (LFM) signal function, simplifying the analysis process. Finally, the processed data undergoes an inverse Fourier transform in the range direction and a Fourier transform in the azimuth direction to convert it to the range-Doppler domain, enabling focusing on the moving target.

[0037] like Figure 1 As shown, this invention provides a ground moving target refocusing method without parameter search, comprising:

[0038] S100: Receives echo signals from moving targets via a radar platform and performs range-direction pulse compression on the echo signals to obtain a pulse-compressed signal.

[0039] This step utilizes the geometric relationship between the radar platform and the moving target to establish an expression for the echo signal of the moving target observed by the zero-oblique-view side-looking radar platform, expressed as:

[0040]

[0041] In the formula, s(t,t) m ) represents the echo signal of a moving target, σ s Let w(t) be the complex scattering coefficient of the moving target. m Let ) be the azimuth window function, t be the fast time, and R be the azimuth window function. s (t m T is the instantaneous slant distance between the platform and the moving target. p K is the pulse duration. r Here, λ is the pulse modulation frequency, c is the speed of light, and λ is the carrier wavelength;

[0042] Instantaneous slant distance R s (t m ) is represented as:

[0043]

[0044] In the formula, v represents the platform speed, and t m It's slow time, v a Indicates the velocity in the azimuth direction, a a R represents the acceleration in the azimuth direction. B v represents the initial slant distance between the platform and the moving target.c a represents the velocity in the distance direction. c This represents acceleration in the distance direction.

[0045] The pulse compression signal in the absence of noise is represented as follows:

[0046]

[0047] In the formula, G is the distance compression gain, and B is the bandwidth of the transmitted signal.

[0048] Considering Gaussian white noise, the pulse compression signal with noise is represented as follows:

[0049]

[0050] In the formula, n(t,t) m ) represents Gaussian white noise in the distance-azimuth time domain.

[0051] S200, the pulse-compressed signal is converted to the distance-frequency domain to obtain the distance-frequency domain compressed signal;

[0052] This step converts the pulse-compressed signal to the range-frequency domain to obtain the range-compressed signal in the range-frequency domain with and without noise.

[0053] If noise is disregarded, the noise-free range compression signal in the range-frequency domain is represented as:

[0054]

[0055] In the formula, A t Let P(f) be the amplitude of the echo signal of the moving target in the frequency domain. r ) = rect(f r / B) is the Fourier transform of the envelope of a linear frequency modulated signal that approximately satisfies a rectangular envelope, f r For the range frequency, f c Let x be the carrier frequency, and rect(x) be a rectangular window function, expressed as follows: x represents the input data;

[0056] If we consider the case with noise (Gaussian white noise), the range compression signal converted to the range-frequency domain when there is noise is represented as follows:

[0057]

[0058] In the formula, n(f r ,t m ) represents Gaussian white noise in the range frequency domain and azimuth time domain.

[0059] S300, construct a distance-frequency difference function based on the distance compression signal;

[0060] This step constructs range-frequency difference functions in the range-frequency domain and azimuth time domain for both noise-free and noisy conditions based on the range compression signal.

[0061] In the noise-free range-frequency domain and azimuth time domain, the range-frequency difference function is expressed as follows:

[0062]

[0063] The range-frequency difference function in the range-frequency domain and azimuth time domain with noise is expressed as follows:

[0064]

[0065] In the formula, Δf represents the distance-frequency interval n(f) r -Δf,t m ) represents Gaussian white noise in the range frequency domain and azimuth time domain with frequency difference.

[0066] In the formula, Δf represents the distance-frequency interval, which satisfies the following conditions:

[0067]

[0068] To avoid range spectrum aliasing Δf, the following formula must also be satisfied:

[0069] |Δf|<F s -B;

[0070] Δf≠0;

[0071] In the formula, F s / N is the range frequency sampling interval, N is the number of sampling points for the range gate, and T a It is the time to synthesize the aperture, F s Where B is the sampling frequency and B is the signal bandwidth.

[0072] Meanwhile, to avoid coherent noise accumulation in the range-time domain, Δf cannot be zero. Furthermore, because the radar signal undergoes discrete sampling after reception, to reduce frequency shift error and ensure that the reference signal and the original moving target signal do not completely overlap in the range-frequency domain support region, Δf should be the range frequency sampling interval F. s An integer multiple of / N, i.e., Δf = DF s / N; N is the number of sampling points for the distance gate, D represents a positive integer, and T a It is the time to synthesize the aperture, F s Where B is the sampling frequency and B is the signal bandwidth.

[0073] It is worth noting that the distance-frequency difference function is easy to construct in practice because it does not require prior knowledge of the moving target and platform motion parameters.

[0074] S400, based on the distance-frequency difference function and the distance compression signal in the distance-frequency domain, the corrected data of the moving target is obtained;

[0075] This step involves taking the conjugate function of the range-frequency difference function and multiplying the conjugate function by the range compression signal in the range-frequency domain of S200 to obtain the corrected echo signal of the moving target, expressed as:

[0076]

[0077] In the formula, A t Let w(t) be the amplitude of the moving target signal in the frequency domain. m P is the azimuth windowing function. * This indicates taking the conjugate function.

[0078] S500 performs an inverse Fourier transform on the corrected signal of the moving target in the range direction, and then a Fourier transform on the azimuth direction to obtain the refocused image of the moving target, represented as:

[0079]

[0080] In the formula, A t ′ represents the amplitude of the echo signal of the moving target in the range-time domain. f represents the convolution operation. a The azimuth frequency is in the range of -PRF / 2 ≤ f a ≤PRF / 2, where PRF is the pulse repetition frequency transmitted by the radar. E(f a ) represents a small linear frequency modulated signal The Fourier transform of the moving target. Compared with the original echo signal of the moving target, the azimuth modulation frequency of the moving target is significantly reduced by a factor Δf / f. c This means that the Doppler support area of ​​the moving target has been greatly reduced.

[0081] The effectiveness of the present invention will be verified through simulation experiments below.

[0082] The simulation parameters for slow-moving targets are shown in Table 1. A range-frequency difference function s(f) is constructed in the azimuth time domain. r -Δf,t m ), selecting the corresponding frequency range Δf as 5MHz, after applying the processing method proposed in this invention, the signal azimuth angle signal is approximately equivalent to a linear frequency modulated signal with a small modulation frequency. For slow-moving targets, the specific calculated value of the small modulation frequency γ is... new =2Δf((vv)a ) 2 -a c R B ) / cR B =0.095Hz / s 2 , and the original modulation frequency: γ=2f0((vv a ) 2 -a c R B ) / cR B = 95.23Hz / s 2 Since Δf is much smaller than f0, the frequency modulation after processing has been greatly reduced, which means that the Doppler diffusion of the moving target has been greatly reduced after RFD processing.

[0083] Table 1 Simulation parameters of radar system and moving target

[0084]

[0085] Figure 3 Figure (a) shows the trajectory of the moving target after distance compression, with a lateral speed of 9 m / s and a longitudinal speed of 10 m / s. Figure 3 As can be seen, the trajectory of the moving target is not a straight line; the energy of the moving target spans several distance units. After processing using the method proposed in this invention, the following result is obtained: Figure 3 The moving target image shown in Figure (b) shows that the moving target is well focused.

[0086] For single-channel synthetic aperture radar (SAR) detection of fast-moving targets on the ground, the simulation parameters are the same as for slow-moving targets. A specific range-frequency difference function is constructed to achieve range migration correction.

[0087] For fast-moving targets, the Doppler frequency of the moving target will exceed the radar pulse repetition frequency (PRF). Then, the Doppler center of the moving target becomes blurred. At this point, the spectrum of the moving target sampled along the azimuth direction has two different cases: one is the complete spectrum, and the other is the spectrum divided into two different ambiguity numbers. A1~A2 represents the processing procedure for the complete spectrum, and B1~B2 represents the processing procedure for the two spectra with different ambiguity numbers.

[0088] A1, for the complete spectrum, assuming the moving target spectrum is not divided into two parts, and the moving target is located within the m-th PRF of the azimuth frequency axis, the range-frequency domain and azimuth-time domain functions of the moving target signal received by the radar can be rewritten as:

[0089]

[0090] In the formula, M is a fuzzy number, which can specifically take the values ​​M = ..., -2, -1, 1, 2...

[0091] A2: Construct the range-frequency difference function s for the moving target signal received by the radar in S3. fast1 (f r -Δf,t m And multiply the moving target signal by the conjugate function of the range-frequency difference function. The corrected signal is obtained, and the specific expression of the distance-frequency domain function of the signal is as follows:

[0092]

[0093] As shown in the above equation, the coupling effect between range frequency and azimuth time (including linear and higher-order components) has been completely eliminated. This indicates that even if the moving target has Doppler center ambiguity, the proposed method still only needs to construct an RFD function to achieve range envelope migration correction for the moving target, without the need for a search operation.

[0094] B1: The spectrum of a moving target with different ambiguity numbers needs to be divided into two parts. The range-frequency domain and azimuth-time domain functions of the moving target signal received by the radar can be expressed as:

[0095]

[0096] Among them, w1(t) m ) and w2(t m ) represent the azimuth envelopes of the m-th PRF and the (m+1)-th PRF portions of the moving target signal on the azimuth frequency axis, respectively.

[0097] B2: Based on the time-frequency relationship, the supporting envelope w1(t) of the moving target signal can also be separated in the azimuth time domain. m ) and w2(t m Construct the range-frequency difference function for each azimuth pulse individually, and then combine the radar received signal in B1 with the complex conjugate of the constructed range-frequency difference function. The processed signal can be obtained by multiplying:

[0098]

[0099] Where w(t) m ) is w1(t m ) and w2(t m () combination.

[0100] Simulation results for detecting fast-moving targets are as follows: Figure 4 As shown, Figure 4 The simulation results for two different types of fast-moving targets are shown in the figure. Figure 4Figure (a) shows the range-compressed moving target with a lateral trajectory velocity of 30 m / s (Doppler center ambiguity number of 1) and a longitudinal trajectory velocity of 10 m / s. As can be seen from the figure, the signal energy traverses a large number of range cells within the synthetic aperture time. After applying the moving target energy refocusing algorithm of this invention, the following result is obtained: Figure 4 The moving target image shown in Figure (b) shows that the fast-moving target is well focused. Figure 4 Figure (c) shows the trajectory of a moving target with a lateral velocity of 13 m / s after range pulse compression. Figure 4 As shown in Figure (d), the Doppler spectrum of the moving target is divided into two parts, each corresponding to a different ambiguity number. After focusing using the refocusing method of this invention, the following result is obtained: Figure 4 The focused moving target is shown in Figure (e).

[0101] Real-world simulation examples:

[0102] The airborne radar has three channels along the range direction. The first channel transmits signals, and all three channels receive echoes. The radar is configured to operate in X-band broadband mode, and the relevant system parameters are shown in Table 2.

[0103] Table 2 Measured Data and Simulation Parameters

[0104] Parameters / Units Value Carrier frequency / GHz 8.85 Bandwidth / MHz 40 Pulse repetition frequency / Hz 1000 Slope distance / m 9000 Platform speed (m / s) 120 Sampling frequency / MHz 72 Number of channels s 3 Synthetic pore size / m 0.56 Pulse count 3072 Distance unit number 500

[0105] Figure 5 The middle (a) figure is an image of the original echo data converted to the range-Doppler domain, from... Figure 5 As can be seen, the moving target is obscured by strong background clutter. Figure 5 Figure (b) shows the clutter suppression results, which effectively suppressed strong background clutter while preserving the moving target. From... Figure 5 Three moving targets with different motion characteristics can be obtained, such as Figure 5 As shown in the yellow ellipse in Figure (b).

[0106] Assuming the moving target's speed does not exceed 120 km / h and its acceleration is less than 8 m / s², 2 Substitute the system and motion target parameters into the formula

[0107]

[0108] The distance frequency interval for obtaining a focused image of a moving target should be less than 16.7 MHz, and the calculated range frequency sampling interval F s / N is 0.2MHz, and the selectable Δf range is between 0.2MHz and 16.7MHz. The larger the Δf, the lower the accumulated energy of the moving target. In order to minimize the loss of energy accumulation of the moving target, Δf can be selected as 0.4MHz.

[0109] Simulation results of various algorithms for slow-moving target 1 are as follows: Figure 6 As shown in (a)-(k); the trajectory of the slow-moving target 1 after pulse compression is as follows. Figure 6 As shown in Figure (a). Figure 6 Chinese (b) map and Figure 6 Figure (c) shows the result of direct azimuth FFT processing after KT processing. It can be seen that the energy of the moving target diffuses in both the range and azimuth dimensions. Figure 6 Chinese (d) diagram and Figure 6 Figure (e) shows the processing result after the two-dimensional frequency matching algorithm. It can be seen that the refocusing effect of the moving target is improved to a certain extent. However, due to the inability to compensate for higher-order phase modulation and residual range envelope migration, the diffusion of the moving target's energy still exists. Figure 6 (f) diagram and Figure 6 Figure (g) shows the result of DKT processing on the moving target. Due to the mismatch of higher-order terms in the azimuth direction, the moving target appears out of focus. Figure 6 Chinese (h) diagram and Figure 6 Figure (i) shows the processing result of the TRT method. Although this method alleviates the mismatch caused by the azimuth velocity of the moving target, the higher-order phase error caused by non-ideal motion still leads to azimuth defocus. The processing result of the algorithm proposed in this invention is shown in Figure (j) of 5. Figure 6 As shown in Figure (k), it can be seen from the figure that the algorithm proposed in this invention can effectively compensate for higher-order phase errors and correct non-ideal range migration, and the moving target 1 has a good refocusing effect in the range-Doppler domain.

[0110] Simulation results of various algorithms for moving targets with high speeds at varying distances are as follows: Figure 7 As shown in (a)-7(k), the trajectory of the moving target 2 with a larger velocity after distance compression is shown in the figure. Figure 7 As shown in Figure (a), due to the large range velocity of this moving target, a more pronounced range migration was observed. Figure 7 Chinese (b) map and Figure 7 The middle (c) figure shows the result of orientation FFT processing. It can be seen from the figure that the energy of the moving target has obvious diffusion in two dimensions, which is not conducive to the detection of moving targets. Figure 7 Chinese (d) diagram and Figure 7 Figure (e) shows the results of the two-dimensional frequency matched filtering algorithm. Figure 7 (f) diagram and Figure 7 Figure (g) shows the results of the DKT method. It can be seen that due to the rapid movement of the moving target, the mismatch in constructing the matched filter is more severe than that of the slow-moving target 1. The results of the TRT method are as follows... Figure 7 Chinese (h) diagram and Figure 7As shown in Figure (i), the TRT method suffers from azimuth defocusing because it cannot handle third-order and higher phase errors; the algorithm proposed in this invention processes the problem as follows. Figure 7 Chinese (j) diagram and Figure 7 As shown in Figure (k), the processed image exhibits a better energy focusing effect on moving targets.

[0111] Simulation results of various algorithms for fast-moving target 3 are as follows: Figure 8 As shown in (a)-8(k), the trajectory of the fast-moving target 3 is as follows. Figure 8 As shown in Figure (a), the Doppler blur of moving target 3 is -1, indicating that the moving target has obvious distance migration, which causes the energy of the moving target to be unable to be refocused. Figure 8 Figures (b) to (c) show the results of direct orientation FFT processing for unfocused energy of moving targets; Figure 8 Figures (d) to (e) show the results of the two-dimensional frequency matched filtering algorithm. The results show that the SAR image of the moving target is blurred due to the high speed of the moving target and insufficient range migration correction. Figure 8 Figures (f) to (i) illustrate the processing results of the DKT and TRT methods. From these figures, we can see that although first-order motion can be corrected by combining Doppler fuzzy search and KT processing, significant inaccuracies still exist when compensating for higher-order phase errors in the azimuth angle. Compared to moving target 2, this difference leads to greater Doppler energy diffusion for the moving target. Furthermore, in reality, since the velocity of non-joint moving targets is unknown, both the DKT and TRT methods require searching for Doppler fuzzy numbers to effectively accumulate the moving target's energy. Therefore, KT processing and Doppler fuzzy search operations significantly increase the computational complexity of these two algorithms. Figure 8 Figures (j) and (k) show the processing results of the method of the present invention, which are consistent with the simulation results. The method can achieve accurate correction and focusing of moving target motion even in the case of Doppler center blur without the need for search operation.

[0112] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A ground moving target refocusing method without parameter search, characterized in that, include: S100: Receives echo signals from moving targets via a radar platform and performs range-direction pulse compression on the echo signals to obtain a pulse-compressed signal. S200, the pulse-compressed signal is converted to the distance-frequency domain to obtain the distance-frequency domain compressed signal; S300, construct a distance-frequency difference function based on the distance compression signal; S400, based on the distance-frequency difference function and the distance compression signal in the distance-frequency domain, the signal after moving target correction is obtained; S500 performs an inverse Fourier transform on the signal after correction of the moving target in the range direction, and then performs a Fourier transform on the azimuth direction to obtain the refocused image data of the moving target. The S300 includes: Based on the aforementioned range compression signal, range-frequency difference functions in the range-frequency domain and azimuth time domain are constructed for both noise-free and noisy conditions; In the noise-free range-frequency domain and azimuth time domain, the range-frequency difference function is expressed as follows: In the formula, A t f represents the amplitude of the echo signal from the moving target in the frequency domain. r For the range frequency, w(t) m ) is the azimuth window function, c is the speed of light, Δf represents the distance-frequency interval, and f c R is the carrier frequency. s (t m ) is the instantaneous slant distance between the platform and the moving target, t m It's slow time; The range-frequency difference function in the range-frequency domain and azimuth time domain with noise is expressed as follows: In the formula, rect(x) is a rectangular window function, expressed as: x is the input data, n(f) r -Δf,t m () represents Gaussian white noise in the range frequency domain and azimuth time domain with frequency difference; the range-frequency interval Δf satisfies the following conditions: |Δf|<F s -B; Δf≠0; Δf=DF s / N; In the formula, R B The initial slant distance between the platform and the moving target is represented by v, and the platform velocity is represented by v. a Indicates the velocity in the azimuth direction, a c F represents acceleration in the distance direction. s / N is the range frequency sampling interval, D represents a positive integer, N is the number of range gate sampling points, and T a It is the time to synthesize the aperture, F s Where B is the sampling frequency and B is the signal bandwidth.

2. The ground moving target refocusing method without parameter search according to claim 1, characterized in that, S100 includes: S110, utilizing the geometric relationship between the radar platform and the moving target, establishes an expression for the echo signal of the moving target observed by the zero-oblique-view side-looking radar platform, expressed as: In the formula, s(t,t) m ) represents the echo signal of a moving target, σ s Let T be the complex scattering coefficient of the moving target, t be the fast time, and T be the scattering coefficient of the moving target. p K is the pulse duration. r λ is the pulse modulation frequency, and λ is the carrier wavelength; Instantaneous slant distance R s (t m ) is represented as: In the formula, a a The acceleration in the azimuth direction, v c Indicates velocity in the direction of distance; S120, the echo signal is subjected to range pulse compression to obtain pulse-compressed signals with and without noise.

3. The ground moving target refocusing method without parameter search according to claim 2, characterized in that, The pulse compression signal in S120 without noise is represented as follows: In the formula, G is the distance compression gain, and B is the bandwidth of the transmitted signal; The pulse compression signal in the presence of noise is represented as follows: In the formula, n(t,t) m ) represents Gaussian white noise in the distance-azimuth time domain.

4. The ground moving target refocusing method without parameter search according to claim 3, characterized in that, S200 includes: The pulse-compressed signal is converted to the range-frequency domain to obtain the range-compressed signals in the range-frequency domain with and without noise. The range compression signal converted to the range-frequency domain in the absence of noise is represented as follows: In the formula, P(f r ) = rect(f r / B) is the Fourier transform of the envelope of a linear frequency modulated signal that approximately satisfies a rectangular envelope; The range compression signal converted to the range-frequency domain when there is noise is represented as follows: In the formula, n(f r ,t m ) represents Gaussian white noise in the range frequency domain and azimuth time domain.

5. The ground moving target refocusing method without parameter search according to claim 4, characterized in that, The S400 includes: Taking the conjugate function of the range-frequency difference function in the converted domain, and multiplying the conjugate function by the corresponding range compression signal in the range-frequency domain of S200, yields the corrected echo signal of the moving target, expressed as: In the formula, A t Let w(t) be the amplitude of the moving target signal in the frequency domain. m P is the azimuth windowing function. * This indicates taking the conjugate function.

6. The ground moving target refocusing method without parameter search according to claim 5, characterized in that, The refocused image data of a moving target in S500 is represented as follows: In the formula, A t ′ represents the amplitude of the echo signal of the moving target in the range-time domain. E(f) represents the convolution operation. a ) represents a small linear frequency modulated signal Fourier transform.

Images