A multi-channel radar forward-looking super-resolution imaging method

Abstract

The present invention discloses a multi-channel radar forward-looking super-resolution imaging method. The method first acquires echo data from the area to be imaged, performs range-direction pulse compression on the acquired data, detects abnormal channels, and obtains the location of the abnormal channels. The method then considers the influence of the abnormal channels and platform motion errors, obtains the platform's yaw and roll using an inertial navigation device, reconstructs a steering matrix, performs super-resolution processing, and reconstructs the target scattering coefficient to achieve super-resolution imaging, thereby obtaining a super-resolution imaging result. The method overcomes the problems of fuzzy forward-looking imaging and low imaging resolution in single-base radars, and facilitates super-resolution imaging of multi-channel radars while taking into account motion errors and channel loss. By utilizing the correlation of pulse compression data from different channels, faulty channels can be quickly detected and their influence on super-resolution performance can be overcome. By modifying the steering matrix to account for the influence of motion errors, super-resolution imaging without geometric distortion can be achieved.

Description

Technical Field

[0001] The present invention belongs to the field of radar imaging technology, and in particular relates to a multi-channel radar forward-looking super-resolution imaging method. Background Art

[0002] Forward-looking radar imaging has important applications in autonomous landing, autonomous navigation, and forward-looking reconnaissance and guidance. However, conventional single-channel SAR or Doppler beam sharpening techniques suffer from blind spots in forward-looking imaging due to Doppler symmetry ambiguity and minimal Doppler variation in the forward-looking region. Bistatic SAR, by separating the transmitter and receiver, can image the forward-looking region of the receiving platform. However, the need for external radiation sources limits imaging autonomy. Furthermore, this separation introduces complex synchronization and motion compensation issues.

[0003] The literature "Hou Haiping, Qu Changwen, Xiang Yingchun, et al. Research on airborne SAR forward imaging based on LFMIGW. Journal of Circuits and Systems, 2011, 16(3):7" proposes that by using one or more transmitting channels and multiple receiving channels on a single platform, an aperture can be formed in the azimuth direction, which has the potential for forward imaging. However, due to the limitation of the platform size, its azimuth resolution is usually very low. The literature "Zhang Jie, Wu Di, Zhu Daiyin. A forward super-resolution imaging algorithm for airborne / missile array radar. Radar Science and Technology, 2018, 16(2):6" proposes the MUSIC algorithm for forward super-resolution imaging. Since the algorithm needs to know the number of sources, it usually requires multiple snapshots of data to achieve good performance. The literature "Wang Jian, Zong Zhulin. Compressed sensing imaging algorithm for forward SAR. Radar Science and Technology, 2012, 10(1):27-31" proposes that the compressed sensing method is used for forward super-resolution imaging, but this method is sensitive to the signal-to-noise ratio (SNR). In addition, these algorithms do not consider the effects of abnormal channels and motion errors.

[0004] During data acquisition, multi-channel radar inevitably encounters abnormal channels. Furthermore, due to atmospheric disturbances, the flight platform is always unstable, exhibiting phenomena such as yaw and roll. Whether these abnormal channels or motion errors are not accounted for during the imaging process, super-resolution results will inevitably be affected.

[0005] Regarding the motion errors of different channels, the paper "Hou Haiping, Qu Changwen, Ding Can, et al. Research on the Influence and Compensation of Array Antenna Micro-motion on Forward-Looking SAR Imaging. Journal of Electronics and Information Technology, 2011, 33(4): 7" uses traditional imaging algorithms to compensate for the errors. However, due to the different imaging mechanisms, how to embed motion compensation into super-resolution algorithms remains a problem. Summary of the Invention

[0006] To solve the above technical problems, the present invention proposes a multi-channel radar forward-looking super-resolution imaging method, which aims to overcome the problems of single-base radar forward-looking imaging blur and low imaging resolution, realize multi-channel radar forward-looking super-resolution imaging, and at the same time overcome the influence of multi-channel motion error and channel loss on imaging.

[0007] The technical solution of the present invention is: a multi-channel radar forward-looking super-resolution imaging method, the specific steps are as follows:

[0008] A. Acquire echo data of the area to be imaged;

[0009] Multi-channel radar forward imaging adopts a one-transmit-multiple-receive channel configuration, in which multiple receiving channels receive sequentially; assuming that the transmitted signal is a linear frequency modulated pulse, the echo signal S received by multiple channels is echo (y k ,t r ; r0) can be expressed as:

[0010]

[0011] Among them, A0 represents a preset constant, K r represents the distance modulation frequency, c represents the speed of light, λ represents the wavelength of the transmitted signal, t r represents distance time, y k represents the azimuth coordinate of the kth receiving antenna, r0 represents the shortest slant distance from the platform to the scene, then the distance history R from any point P(x0,y0) to the transmitting antenna tx and the distance history R to different receiving antennas rx It can be expressed as:

[0012]

[0013]

[0014] Where h1 represents the height difference between the transmitting antenna and the receiving antenna, and h represents the flight altitude of the platform.

[0015] B. Perform range pulse compression on the acquired data;

[0016] Set the matching function of pulse compression to S ref (t r )=exp(-jπK r t r 2 ), then the pulse compressed signal S compress (y k ,t r ) can be expressed as:

[0017]

[0018] Here, IFFT represents an inverse Fourier transform operator, and FFT represents a Fourier transform operator.

[0019] C. Perform abnormal channel detection to obtain the abnormal channel location;

[0020] First, the range pulse compression results of each channel are non-coherently accumulated to obtain the benchmark E sum as follows:

[0021]

[0022] Among them, M represents the number of channels, X k Represents the pulse compression data of the kth channel.

[0023] Then, E sum The normalized result is defined as Y b , the normalized signal of each channel is defined as Y k , define the correlation coefficient ρ between channels k for:

[0024]

[0025] Where E(·) represents the mathematical expectation.

[0026] When the normalized correlation coefficient is less than the empirical value (ε), the channel is considered to be faulty. After determining and deleting L abnormal channels, the echo S of one range unit after range compression c (y k ) can be rewritten as:

[0027]

[0028] Where k represents the number of receiving antennas.

[0029] According to the knowledge of array signal processing, if the antenna array consists of ML channels and N narrowband signals are incident on the spatial array, the received signal z of the k channel is k It can be expressed as:

[0030]

[0031] Among them, β i represents the scattering coefficient of the i-th target, b k represents the noise of the kth channel, τ ki It represents the delay of the i-th signal reaching the k-th channel relative to the reference channel, that is, the path difference delay. represents the phase difference of the wave path, and f0 represents the carrier frequency of the transmitted signal. Then the echo of a distance unit in different channels can be written as:

[0032]

[0033] Then formula (9) can be written as:

[0034] Z (M-L)×1 =A (M-L)×N β N×1 +B (M-L)×1 (10)

[0035] Among them, Z (M-L)×1 Represents the echo in one range unit, B (M-L)×1 represents the noise vector, A (M-L)×N represents the steering matrix, β N×1 represents the scattering coefficient vector.

[0036] D. Obtain the platform's yaw and roll from the inertial navigation device and reconstruct the guidance matrix;

[0037] Obtain the platform yaw angle from the inertial navigation device And the roll angle θ, calculate the motion error (Δx, Δy, Δz) of each channel:

[0038]

[0039] Among them, x, y, and z represent the coordinates of each channel during the platform's data acquisition process.

[0040] Then the transmitting antenna T x and receiving antenna R x The coordinates are as follows:

[0041]

[0042] Among them, t ak Indicates the sampling time of the kth channel, Δx k , Δy k , Δz k represents the motion error fed back by the inertial navigation device at the sampling time of the kth channel, v represents the forward flight speed of the platform, and v s Indicates the switching speed of each array element.

[0043] Next, to obtain t ak , grid the imaging area as:

[0044]

[0045] Among them, (X j , Y i ,0) represents the coordinates of the i-th row and j-th column in the divided grid, L a Represents the azimuth length of the imaging scene, Lr Represents the distance length of the image scene, N a Indicates the number of channels, N r Indicates the distance from the sampling point, d r Indicates the projected length from the platform to the center of the scene.

[0046] Therefore, when sampling the kth channel, the coordinates P1 and P2 of the transmitting and receiving antennas are as shown in equation (12), and the coordinates of the i-th row and j-th column in the scene are R(X j , Y i , 0), then the delay time τ of the i-th signal ki It can be expressed as:

[0047]

[0048] In calculating τ ki Then, substitute into formula (9) to construct the steering matrix A (M-L)×N .

[0049] E. Perform super-resolution processing to obtain super-resolution imaging results;

[0050] According to formula (10), the weighted least squares cost function is defined as The form is as follows:

[0051]

[0052] in, represents the least squares weight coefficient, a k represents the kth column of the steering matrix A, R represents the covariance matrix, and R = A·P·A H , P represents a diagonal matrix, the diagonal element p k =|β k | 2 , k=1,2,...,N,β k represents the target scattering coefficient, (·) H Indicates the transpose operation on the matrix.

[0053] For formula (15) Taking the partial derivative and setting the result equal to zero, we get:

[0054]

[0055] According to the matrix inversion lemma, we can get Substituting it into formula (16), we can get:

[0056]

[0057] Because βk It is related to the goal, so it needs to be solved iteratively. The iterative steps are as follows:

[0058] a. Calculate P = [|β1| 2 ,|β2| 2 ,...,|β N | 2 ] H ;

[0059] b. Calculate the covariance matrix R = A (M-L)×N ·diag(P)·(A (M-L)×N ) H ;

[0060] c. Update

[0061] d. Return to step a.

[0062] Finally, the target scattering coefficient β k The reconstruction estimation result is the super-resolution imaging result.

[0063] Beneficial effects of the present invention: The method of the present invention first acquires echo data of the area to be imaged, performs range-direction pulse compression on the acquired data, performs abnormal channel detection, obtains the position of the abnormal channel, then considers the influence of the abnormal channel and platform motion error, obtains the platform yaw and roll by the inertial navigation device, reconstructs the steering matrix, performs super-resolution processing, reconstructs the target scattering coefficient to achieve super-resolution imaging, and obtains super-resolution imaging results. The method of the present invention overcomes the problems of fuzzy forward imaging and low imaging resolution of single-base radar, and helps to achieve super-resolution imaging of multi-channel radar under the consideration of motion error and channel loss. By utilizing the correlation of pulse compression data of different channels, it can quickly detect faulty channels and overcome their influence on super-resolution performance. By modifying the steering matrix to consider the influence of motion error, it can achieve super-resolution imaging without geometric distortion. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 2 is a geometric model diagram of multi-channel radar forward imaging in an embodiment of the present invention.

[0065] Figure 2 The present invention provides a flow chart of a multi-channel radar forward-looking super-resolution imaging method.

[0066] Figure 3 Schematic diagram of an observation scene in an embodiment of the present invention.

[0067] Figure 4 Schematic diagram of the echo of the observation scene in an embodiment of the present invention.

[0068] Figure 5This is a diagram showing the result of pulse compression of the echo of the observation scene in an embodiment of the present invention.

[0069] Figure 6 This is a result diagram of an echo with channel loss when SNR=0 dB after pulse compression in an embodiment of the present invention.

[0070] Figure 7 It is a reference image accumulated from each channel in the embodiment of the present invention.

[0071] Figure 8 This is a data diagram of a normal channel with SNR=0dB in an embodiment of the present invention.

[0072] Figure 9 Graph showing correlation coefficients between different channel data and the benchmark in an embodiment of the present invention.

[0073] Figure 10 In the embodiment of the present invention Motion error diagram.

[0074] Figure 11 This is a diagram of super-resolution imaging results taking into account motion compensation and abnormal channel detection in an embodiment of the present invention.

[0075] Figure 12 This is an imaging result diagram of the traditional algorithm (BP) in an embodiment of the present invention.

[0076] Figure 13 This is a super-resolution imaging result diagram without motion compensation and abnormal channel detection in an embodiment of the present invention. DETAILED DESCRIPTION

[0077] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0078] In this embodiment, the geometric configuration of the forward-looking multi-channel radar is as follows: Figure 1 The parameters of the forward-looking multi-channel radar are shown in Table 1:

[0079] Table 1

[0080]

[0081] like Figure 1 As shown, in an xyz space coordinate system, O represents the origin of the coordinate system, and the forward-looking multi-channel radar is moving at a speed v r =20m / s, height h e =1050mFlying in a straight line at a constant speed along the x-axis, the center of the target scene is X c =1050m, Y c =0m, Z c = 0, that is, the angle between the carrier and the center of the scene is 45°.x Located at h1 = 0.1m directly below the central receiving antenna, during the flight, the transmitting antenna T x The signal is transmitted at a high pulse repetition frequency (PRF = 15000 Hz), and each receiving antenna switches along the azimuth direction at a switching speed v s =PRF=L e / 256*PRF=703.125m / s sequentially switches to receive echo, which is equivalent to a receiving antenna R x In the y-axis direction, the same speed v s Movement. If the number of receiving channels is N a =64, then the total length in the azimuth direction is equivalent to L e =3.00m synthetic aperture, the array element spacing is d=L e / 64≈0.047m. In this simulation, it is assumed that the wavelength of the radar signal is λ=0.0315m and the pulse width is T r = 1μs, bandwidth B = 60MHz LFM signal. Number of sampling points N in the range direction r is 256, the number of sampling points in the azimuth direction is N a is 64, and the echo signal-to-noise ratio after pulse compression is 0dB.

[0082] like Figure 2 As shown in FIG, a flow chart of a multi-channel radar forward-looking super-resolution imaging method of the present invention, the specific steps are as follows:

[0083] A. Acquire echo data of the area to be imaged;

[0084] like Figure 3 As shown, in this embodiment, the azimuth length of the scene is 716.8m, and the distance length is 224m. According to the azimuth and distance sampling points N a and N r , assuming that the scene is evenly divided into 64×256 grids, using formula (13), the coordinate points of each target in the scene can be obtained.

[0085] In addition, assume that the yaw angle and roll angle of the carrier aircraft are Combined with the aircraft flight speed v r , using formula (15) we can get the position of each channel of the carrier.

[0086] Finally, the launch distance R can be calculated using formula (2) and (3) tx and receiving distance R rx , and then use formula (1) to calculate the echo signal, such as Figure 4 shown.

[0087] B. Perform range pulse compression on the acquired data;

[0088] Assume that the center position of the target scene is X c =0, Y c = 0m, Zc = 0, use formula (2) (3) to calculate the reference transmission distance and receiving distance, and then get the distance time t r and the matching function S of pulse compression ref (t r )=exp(-jπK r t r 2 ), considering the echo data generated in step A of this embodiment, the result after echo pulse compression can be obtained using formula (4), as shown in Figure 5 shown.

[0089] Based on step B of this embodiment, abnormal channel simulation is performed. For the results of pulse compression, 10 channels are randomly selected so that the data of these channels are 0. On this basis, Gaussian white noise is randomly introduced according to the total energy of all channels so that the overall signal-to-noise ratio (SNR) of all channels is 0dB. Figure 6 As shown in FIG, the numbers of the abnormal channels are 13, 18, 21, 24, 39, 45, 48, 55, 56, and 62, respectively, and only noise signals exist in these channels.

[0090] C. Perform abnormal channel detection to obtain the abnormal channel location;

[0091] Typically, a faulty channel contains only noise and no valid signal. In theory, the channel's health can be determined by comparing the amplitude of the compressed echo data with a fixed threshold. However, determining an appropriate threshold is difficult in practice. This approach uses data from multiple channels to determine channel health by calculating the correlation coefficient between the data from different channels and a baseline formed by combining the multi-channel data.

[0092] Considering the echo data of the abnormal channel simulated in step B of this embodiment, the data of each channel are accumulated using formula (5) to obtain the reference Y b ,like Figure 7 As shown, the data of the normal channel at this time is also given, such as Figure 8 As shown. On this basis, use formula (6) to calculate the correlation coefficient ρ of each channel k , when the normalized correlation coefficient of a channel is smaller than the empirical threshold, the channel is considered to be an abnormal channel, such as Figure 9 As shown, here the threshold is taken as 0.5, it can be obtained that by combining the normalized correlation coefficient with Figure 9 By comparing the thresholds represented by the straight line in the figure, all fault channels can be clearly identified. After obtaining the location of the abnormal channel, the new distance pulse pressure result is obtained according to formula (7).

[0093] D. Obtain the platform's yaw and roll from the inertial navigation device and reconstruct the guidance matrix;

[0094] Typically, the steering matrix is ​​designed based on the distance from each channel to the observation scene. Therefore, it is very important to determine the actual position of each channel of the platform. In the airborne forward-looking model discussed in step A, the position of each channel is usually affected by the yaw and roll of the platform. If this motion error is not taken into account, the steering matrix A will result in (M-L)×N Therefore, the steering matrix designed in this embodiment takes the motion error of the platform into consideration.

[0095] First, consider the motion error of the carrier aircraft. Through the feedback of the inertial navigation device, we can know that the yaw angle and roll angle of the carrier aircraft during the sampling period are According to formula (11), the motion error of each channel (Δx, Δy, Δz) can be calculated as follows: Figure 10 shown.

[0096] Further, according to formula (12), the coordinates P1 and P2 of the transmitting and receiving antennas of the carrier aircraft during the sampling period can be obtained. Then, the scene to be observed is divided by formula (13) to obtain the coordinates of each grid. Finally, combined with the position of the abnormal channel, using formulas (14) and (9), a steering matrix A of a distance unit can be constructed. (M-L)×N .

[0097] E. Perform super-resolution processing to obtain super-resolution imaging results;

[0098] Combined with the steering matrix A in step D of this embodiment (M-L)×N , according to the iterative steps in step E in the invention, it does not need to be iterated more than 10 to 15 times to obtain a better super-resolution imaging result, and the final imaging result can be shown as Figure 11 The result of introducing motion error using the traditional imaging algorithm BP is shown as follows. Figure 12 As shown in FIG, it is obvious that the multi-channel radar forward-looking super-resolution imaging method proposed in this embodiment achieves azimuth super-resolution. However, the super-resolution results without considering abnormal channels and motion errors are shown in FIG. Figure 13 As shown, it can be seen that if the motion error and abnormal channels are not considered, the super-resolution performance will degrade and the imaging results will be offset, which is caused by the mismatch between the steering matrix and the echo data.

[0099] The multi-channel radar forward-looking super-resolution imaging method described in this embodiment facilitates super-resolution imaging in multi-channel radars while accounting for motion errors and missing channels. Specifically, by leveraging the correlation of pulse compression data from different channels, the multi-channel radar forward-looking super-resolution imaging method of the present invention can rapidly detect faulty channels and mitigate their impact on super-resolution performance. Furthermore, by modifying the steering matrix to account for the effects of motion errors, the method of the present invention can achieve super-resolution imaging without geometric distortion.

Claims

1. A multi-channel radar forward-looking super-resolution imaging method, the specific steps are as follows: A. Acquire echo data of the area to be imaged; Multi-channel radar forward imaging adopts a one-transmit-multiple-receive channel configuration, in which: Multiple receiving channels receive sequentially; assuming that the transmitted signal is a linear frequency modulation pulse, the echo signal S received by multiple channels echo (y k ,t r ; r0) can be expressed as: Among them, A0 represents a preset constant, K r represents the distance modulation frequency, c represents the speed of light, λ represents the wavelength of the transmitted signal, t r represents distance time, y k represents the azimuth coordinate of the kth receiving antenna, r0 represents the shortest slant distance from the platform to the scene, then the distance history R from any point P(x0,y0) to the transmitting antenna tx and the distance history R to different receiving antennas rx It can be expressed as: Where h1 represents the height difference between the transmitting antenna and the receiving antenna, and h represents the flight altitude of the platform; B. Perform range pulse compression on the acquired data; Set the matching function of pulse compression to S ref (t r )=exp(-jπK r t r 2 ), then the pulse compressed signal S compress (y k ,t r ) can be expressed as: Wherein, IFFT represents the inverse Fourier transform operator, and FFT represents the Fourier transform operator; C. Perform abnormal channel detection to obtain the abnormal channel location; First, the range pulse compression results of each channel are non-coherently accumulated to obtain the benchmark E sum as follows: Among them, M represents the number of channels, X k represents the pulse compression data of the kth channel; Then, E sum The normalized result is defined as Y b , the normalized signal of each channel is defined as Y k , define the correlation coefficient ρ between channels k for: Where E(·) represents the mathematical expectation; When the normalized correlation coefficient is less than the empirical value (ε), it is considered that the channel is faulty. After determining and deleting L abnormal channels, the echo S of one range unit after range compression is c (y k ) can be rewritten as: Where k represents the number of receiving antennas; According to the knowledge of array signal processing, if the antenna array consists of ML channels and N narrowband signals are incident on the spatial array, the received signal z of the k channel is k It can be expressed as: Among them, β i represents the scattering coefficient of the i-th target, b k represents the noise of the kth channel, τ ki It represents the delay of the i-th signal reaching the k-th channel relative to the reference channel, that is, the path difference delay. represents the phase difference of the wave path, and f0 represents the carrier frequency of the transmitted signal; then the echo of a distance unit in different channels can be written as: Then formula (9) can be written as: Z (M-L)×1 =A (M-L)×N ·β N×1 +B (M-L)×1 (10) Among them, Z (M-L)×1 Represents the echo in one range unit, B (M-L)×1 represents the noise vector, A (M-L)×N represents the steering matrix, β N×1 represents the scattering coefficient vector; D. Obtain the platform's yaw and roll from the inertial navigation device and reconstruct the guidance matrix; Obtain the platform yaw angle from the inertial navigation device And the roll angle θ, calculate the motion error (Δx, Δy, Δz) of each channel: Among them, x, y, and z represent the coordinates of each channel during the platform's data acquisition process; Then the transmitting antenna T x and receiving antenna R x The coordinates are as follows: Among them, t ak Indicates the sampling time of the kth channel, Δx k ,Δy k ,Δz k represents the motion error fed back by the inertial navigation device at the sampling time of the kth channel, v represents the forward flight speed of the platform, and v s Indicates the switching speed of each array element; Next, to obtain t ak , grid the imaging area as: Among them, (X j , Y i ,0) represents the coordinates of the i-th row and j-th column in the divided grid, L a Represents the azimuth length of the imaging scene, L r Represents the distance length of the image scene, N a Indicates the number of channels, N r Indicates the distance from the sampling point, d r Represents the projection length from the platform to the center of the scene; Therefore, when sampling the kth channel, the coordinates P1 and P2 of the transmitting and receiving antennas are as shown in equation (12), and the coordinates of the i-th row and j-th column in the scene are R(X j ,Y i ,0), then the delay time τ of the i-th signal ki It can be expressed as: In calculating τ ki Then, substitute into formula (9) to construct the steering matrix A (M-L)×N ; E. Perform super-resolution processing to obtain super-resolution imaging results; According to formula (10), the weighted least squares cost function is defined as The form is as follows: in, represents the least squares weight coefficient, a k represents the kth column of the steering matrix A, R represents the covariance matrix, and R = A·P·A H , P represents a diagonal matrix, the diagonal element p k =|β k | 2 ,k=1,2,…,N,β k represents the target scattering coefficient, (·) H Indicates the transpose operation of the matrix; For formula (15) Taking the partial derivative and setting the result equal to zero, we get: According to the matrix inversion lemma, we can get Substituting it into formula (16), we can get: Because β k It is related to the goal, so it needs to be solved iteratively. The iterative steps are as follows: a、calculationP=,|β1| 2 ,|β2| 2 ,…,|b N | 2 ] H ; b. Calculate the covariance matrix R = A (M-L)×N ·diag(P)·(A (M-L)×N ) H ; c. Update d. Return to step a; Finally, the target scattering coefficient β k The reconstruction estimation result is the super-resolution imaging result.

Images