SAR three-dimensional super-resolution imaging method based on sparse constraint and back projection cancellation
Through sparse constraints and reverse projection cancellation processing, the problem of strong target side lobes covering weak targets in synthetic aperture radar imaging is solved, the distance upward resolution is improved, and high-quality three-dimensional super-resolution imaging is achieved.
Patent Information
- Application Number
- CN202410799790.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-20
- Publication Date
- 2024-11-08
AI Technical Summary
In synthetic aperture radar imaging, due to the limited bandwidth of the transmit signal, the strong target distance side lobe covers the weak target, affecting the imaging results. The existing windowing method will cause the image resolution to decrease when weakening the side lobe, making it difficult to obtain high-quality Distance resolution image.
Using SAR three-dimensional super-resolution imaging method based on sparse constraints and reverse projection cancellation, echo data is processed through RMA algorithm, and imaging results with different distances are obtained using sparse constraints, and sidelobe effects of excess energy are eliminated through reverse projection. , improve the resolution of the distance upward.
It effectively eliminates the influence of excess energy in the object, improves the resolution of distance upwards, obtains high-quality distance resolution images, and verifies the feasibility and effectiveness of the method.
Smart Images

Figure CN118915062A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a SAR three-dimensional super-resolution imaging processing method based on sparse constraints and back-projection cancellation, and belongs to the technical field of radar image processing. Background Art
[0002] Synthetic Aperture Radar (SAR) uses the large bandwidth of the transmitted signal to obtain high range resolution, and uses the relative motion between the radar and the target to obtain a larger synthetic aperture, thereby obtaining higher azimuth resolution.
[0003] In synthetic aperture radar imaging, the inverse fast Fourier transform is usually used to obtain the radar range information of the echo signal. Due to the limited bandwidth of the transmitted signal, there will be strong target range sidelobes covering weak targets during the imaging process, affecting the imaging results. In order to solve the sidelobe problem, the common method is to solve it by adding windows. However, the windowing method will also widen the main lobe while weakening the sidelobes, resulting in a decrease in image resolution. Therefore, in order to obtain high-quality range resolution images, it is very important to eliminate the energy of redundant objects without affecting the main lobe content. Summary of the invention
[0004] The purpose of the present invention is to address the defects and shortcomings of the above-mentioned prior art and propose a SAR three-dimensional super-resolution imaging method based on sparse constraints and back projection cancellation. The method processes the echo data of the original planar array through the RMA algorithm, and then uses sparse constraints to obtain imaging at different distances upward, and then uses back projection to eliminate the sidelobe effect of excess object energy brought about by imaging. Finally, experimental simulation is carried out to verify the feasibility of the method.
[0005] The technical solution adopted by the present invention to solve the technical problem is: a SAR three-dimensional super-resolution imaging method based on sparse constraints and back-projection cancellation, the method comprising the following steps:
[0006] Step S1: construct a SAR signal model of a planar scanning array to obtain an echo signal model after the radar plane is scanned.
[0007] Step S2: Construct a matched filter to convolve with the echo data to reconstruct the target image.
[0008] Step S3: Use the RMA algorithm to analyze the range migration of the radar echo signal (ie, the position change of the target during the exposure of the radar wave) to perform imaging.
[0009] Step S4: Sparsely constrain the imaging results in different distance directions to obtain output images with different target energies and side lobes.
[0010] Step S5: Perform back-projection cancellation on the sparse data to eliminate the influence of redundant energy and improve the upward resolution of the distance.
[0011] Step S6: Experimental simulation verifies the effectiveness of the method of the present invention for improving the imaging resolution of the target in the distance direction by using joint sparse and back-projection cancellation.
[0012] In the above technical solution, step S1 constructs a SAR planar scanning array to obtain the echo signal model as follows:
[0013]
[0014] Where (x0, y0, 0) is the radar coordinate, the target reflection point coordinate is (x, y, z0), z0 represents the height of the target plane from the radar plane, and f(x, y) is the scattering rate function of the two-dimensional target.
[0015] Step S2: Obtain wave numbers by spatial spectrum theory A matched filter is constructed based on the scattering function and the point spread function, and the filter is convolved with the echo data to reconstruct the target image. The echo data can be expressed as follows:
[0016]
[0017] in:
[0018]
[0019] Step S3 performs precise phase correction on each range unit through the range migration information of the target during radar illumination, transforms the echo signal in the range-azimuth domain, and separates the signals of different range units.
[0020] Step S4 performs sparse constraints on different distance directions. The specific steps are: Gram-Schmidt orthogonalization is performed on the atoms selected in each iteration, and the residual vector after each iteration is orthogonalized with the previous orthogonal atoms, thereby ensuring that the residual is orthogonal to the selected orthogonal atoms.
[0021] After completing the sparse constrained imaging, step S5 takes the imaging results at different distances upward, performs back-projection cancellation processing in the image domain and the data domain respectively, eliminates the influence of excess energy, and obtains a higher quality distance resolution output image.
[0022] Beneficial effects:
[0023] 1. The present invention performs sparse processing on the imaging results of the RMA algorithm, first obtains output images at different distances, then analyzes these images, and finally uses backprojection cancellation processing to obtain high-range resolution images.
[0024] 2. The present invention eliminates the influence of redundant object energy and improves the upward resolution of distance. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 The present invention discloses a SAR three-dimensional super-resolution imaging flow chart based on sparse constraints and back-projection cancellation.
[0026] Figure 2 It is the SAR signal model of the surface scanning array in the present invention.
[0027] Figure 3a It is a real front view of the imaging target selected in the present invention.
[0028] Figure 3b It is a physical side view of the imaging target selected in the present invention.
[0029] Figure 4a This is the RMA imaging diagram of the target at a distance of 16 in the present invention.
[0030] Figure 4b This is the RMA imaging diagram of the target at a distance of 19 in the present invention.
[0031] Figure 5a This is the result of the present invention after thinning the RMA imaging of the target at a distance of 16.
[0032] Figure 5b This is the result of the present invention after thinning the RMA imaging of the target at a distance of 19.
[0033] Figure 6a This is the result of the back-projection cancellation process of the sparse imaging of the target at a distance of 16 in the present invention. Figure 6b This is the result of the back-projection cancellation process of the sparse imaging of the target at a distance of 19 in the present invention. Figure 7a It is a spectrum diagram of the original imaging in the upward direction of the distance in the present invention.
[0034] Figure 7b It is the spectrum diagram of the sparse imaging in the distance direction after back-projection cancellation in the present invention.
[0035] Figure 8a The present invention performs three-dimensional imaging on the sparse results after the back-projection cancellation processing. DETAILED DESCRIPTION
[0036] In order to deepen the understanding of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. The embodiments are only used to explain the present invention and do not limit the protection scope of the present invention.
[0037] like Figure 1 As shown, the present invention provides that the method specifically comprises the following steps:
[0038] Step S1, constructing a SAR signal model of a plane scanning array, and obtaining an echo signal model after radar plane scanning;
[0039] Step S2, constructing a matched filter to convolve with the echo data to reconstruct the target image;
[0040] Step S3, using the RMA algorithm to analyze the range migration of the radar echo signal (i.e., the position change of the target during the radar wave irradiation) to perform imaging;
[0041] Step S4, sparsely constraining the imaging results in different distance directions to obtain output images with different target energies and side lobes;
[0042] Step S5, performing back-projection cancellation processing on the sparse data to eliminate the influence of redundant energy side lobes and improve the upward resolution of the range;
[0043] Step S6: Experimental simulation verifies the effectiveness of the method of the present invention for improving the imaging resolution of the target in the distance direction by using joint sparse and back-projection cancellation.
[0044] like Figure 2 As shown, the specific process of the above step S1 of the present invention is: use the radar platform to perform line-by-line scanning on a two-dimensional plane to form a synthetic aperture. Since the transmitting and receiving antennas of the selected millimeter-wave radar are very close to each other, a full-duplex antenna can be used to approximately represent the radar position, and the position of the radar is the center position of the equivalent transmitting and receiving antenna. Assuming that the radar coordinates are (x0, y0, 0), the coordinates of the target reflection point are (x, y, z0), and z0 represents the height of the plane where the target is located from the plane where the radar is located. When imaging, only the slice data at z0 needs to be processed to obtain the imaging of the target. Assuming that f(x, y) is the scattering rate function of the two-dimensional target, the reflected echo data can be expressed as
[0045]
[0046] The specific process of step S2 is as follows: The wave number can be obtained from the spatial spectrum theory The process of radar backscatter signal from the target to the scanning plane can be equivalent to passing through a bandpass filter, and the scattering function f(x,y) is derived:
[0047]
[0048] In order to reconstruct the target image, it is necessary to construct a matched filter h(x,y) and convolve it with the echo data. h(x,y) can be obtained by the point spread function h0(x,y)
[0049]
[0050] Formula (1) can be modified through (2) and (3) to:
[0051]
[0052] From (3) and (4), the final imaging result can be deduced as shown below:
[0053]
[0054] like Figure 3a , Figure 3b As shown in the figure, the selected data comes from the high-resolution three-dimensional millimeter-wave radar data set of the University of Electronic Science and Technology of China. The acquisition mode of this data set is that the millimeter-wave sensor scans along a two-dimensional plane for near-field imaging. The scanning size is 0.4m×0.4m, the radar center frequency is 78.8GHz, and the transmission signal bandwidth is 3.6GHz. The number of signal sampling points of the millimeter-wave sensor in the distance, horizontal and vertical directions are 256 points, 407 points and 200 points respectively, and the dimension of the single-channel original echo data formed is 256×407×200.
[0055] The specific process of the above step S3 of the present invention is: RMA includes SAR point target echo signals in range and azimuth directions as follows:
[0056] s τ (τ,η)=A0rect(τ-2R t (η) / c)w a (η-η c )×exp[-j4πf0R t (η) / c]
[0057] ×exp[jπK r (τ-2R t (η) / c) 2 ](6)
[0058] Where A0 is the amplitude complex constant, τ and η represent the fast time and azimuth time respectively, and η c represents the beam center deviation time, w a represents the gain pattern of the antenna in azimuth, f0 is the radar center frequency, Kr is the range chirp modulation frequency, R t () is the slant distance, and the data is transformed into the frequency domain using a two-dimensional Fourier transform:
[0059] S2df (f τ ,f η )=AW r (f τ )W a (f η -f ηc )×exp[jθ 2df (f τ ,f η )] (7)
[0060] Assuming the range pulse is an LFM signal with a hyperbolic equation, the phase is given by:
[0061]
[0062] To set the range and radar effective speed to their median or reference values, a reference function multiplier (RFM) filter is introduced with the phase:
[0063]
[0064] After processing with the above RFM filter, the phase becomes:
[0065]
[0066] The decoupling is then achieved by Stolt interpolation, which modifies the scale of the distance to frequency axis and converts θ RFM The square roots in are replaced by shifted and scaled variables, that is:
[0067]
[0068] When processing specific images, f τ The axis is mapped to f τ ′. After Stolt interpolation, the phase function is given by:
[0069]
[0070] Finally, the two-dimensional inverse Fourier transform is used to convert the data back to the time domain, that is, to generate a target image with compressed distance and azimuth:
[0071]
[0072] The specific process of step S4 of the present invention is as follows:
[0073] (1) Initialize the residual r0 = y, index set α=0, number of iterations t=1.
[0074] (2) Calculation of λ i =argmax|<rt-1 ,θ j >|, where θ j is the jth column of the sensor matrix.
[0075] (3) Update the index set: Ω t =Ω t-1 ∪λ.
[0076] (4) Calculation based on the least squares method
[0077] (5) Update the residual vector
[0078] When t≥k, t=t+1, continue the loop, otherwise terminate the loop and output
[0079] The specific process of the above step S5 is: Assume that at a certain distance to K a In the imaging, there are two objects a and b. The energy of a is much greater than that of b. In order to eliminate the sidelobe effect brought by b, another distance K is required. b (The energy of object b is much greater than that of object a, and object a brings about the side lobe effect) imaging. From equations (5) and (13), it can be deduced that:
[0080]
[0081]
[0082] in The distance to K a The sparse processing of the original image, K a The echo data at K a System function at (matched filter). Indicates the distance to K b Sparse processing of RMA imaging, K b The two-dimensional Fourier transform of the echo data at K b The phase factor at K is directly subtracted in the image domain to eliminate the phase factor at the distance. a The side lobe effect caused by the object b:
[0083]
[0084] Algorithm principle: Echo data and system functions at different distances are obtained by equations (14) and (15). Perform two-dimensional Fourier transform on equation (15):
[0085]
[0086] Then combine the above formula with K b System Function The conjugate multiplication of the two-dimensional Fourier transform obtains the representation of the original echo data in the frequency domain and projects the data onto K a System function at On the two-dimensional Fourier transform of :
[0087]
[0088] Perform a two-dimensional inverse Fourier transform on equation (18) to return the projection The imaging on this result indicates that K b The imaging at the system function The projection on is equivalent to K a Side lobe influence at the location:
[0089]
[0090] The final imaging result is:
[0091]
[0092] The specific steps of the above step S6 of the present invention are: experimental simulation verifies the effectiveness of the method of the present invention for joint sparse and backprojection cancellation to improve the imaging resolution of the target in the distance direction:
[0093] Simulation Results Figure 4a - Figure 6b As shown, Figure 4a and Figure 4b These are the RMA images at distances 16 and 19 respectively. Figure 5a and Figure 5b These are the results of sparse RMA imaging when the target is at distances 16 and 19, respectively. Figure 6a and Figure 6b They are the results of backprojection cancellation of the sparse imaging of the target at distances of 16 and 19 respectively. Figure 7a and Figure 7b They are respectively the spectrum diagram of the original imaging in the upward direction and the spectrum diagram of the sparse imaging in the upward direction after back-projection cancellation. Figure 8a This is the final 3D imaging result. It can be seen that the sparse result after back-projection cancellation has obvious peaks at 16 and 19 in the distance direction, which means that objects can be clearly distinguished in these two distance directions, improving the distance resolution and achieving high-resolution imaging in the distance direction.
[0094] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.
Claims
1. A SAR three-dimensional super-resolution imaging method based on sparse constraints and back-projection cancellation, characterized in that: The following steps are involved: Step S1: construct a SAR signal model of a plane scanning array, obtain the echo signal model after the radar plane scanning, the FMCW radar transmits a linear frequency modulated pulse signal (Linear Frequency Modulation, LFM), also known as a chirp signal, and then construct a SAR plane scanning array, and obtain the echo signal expression as follows: Where (x0, y0, 0) is the radar coordinate, the target reflection point coordinate is (x, y, z0), z0 represents the height of the target plane from the radar plane, and f(x, y) is the scattering rate function of the two-dimensional target; Step S2: Convolve a matched filter with the echo data to reconstruct the target image; Step S3: using the RMA algorithm to analyze the range migration of the radar echo signal (i.e., the position change of the target during the radar wave irradiation) to perform imaging; Step S4: Sparsely constrain the imaging results in different distance directions to obtain output images with different target energies and side lobes; Step S5: performing back-projection cancellation on the sparse data to improve the upward resolution of the distance; Step S6: Experimental simulation verifies the effectiveness of the method of combining sparse constraints and back-projection cancellation in the present invention to improve the imaging resolution of the target in the distance direction.
2. The SAR three-dimensional super-resolution imaging method based on sparse constraint and back-projection cancellation according to claim 1, characterized in that: The step S2 obtains the wave number by using the spatial spectrum theory A matched filter is constructed based on the scattering function and the point spread function, and the filter is convolved with the echo data to reconstruct the target image. The echo data can be expressed as follows: in:
3. The SAR three-dimensional super-resolution imaging method based on sparse constraint and back-projection cancellation according to claim 1, characterized in that: The step S3 performs accurate phase correction on each range unit according to the range migration information of the target during radar illumination, transforms the echo signal in the range-azimuth domain, and separates the signals of different range units.
4. The SAR three-dimensional super-resolution imaging method based on sparse constraint and back-projection cancellation according to claim 1, characterized in that: The step S4 performs sparse constraints on different distance directions, and the specific steps are: Gram-Schmidt orthogonalization is performed on the atoms selected in each iteration, and at the same time, the residual vector after each iteration is orthogonalized with the previous orthogonal atoms, ultimately ensuring that the residual is orthogonal to the selected orthogonal atoms.
5. The SAR three-dimensional super-resolution imaging method based on sparse constraint and back-projection cancellation according to claim 1, characterized in that: After completing the sparse constrained imaging, step S5 takes imaging results at different distances upward, performs back-projection cancellation in the image domain and the data domain respectively, eliminates the energy influence of redundant objects, and obtains a higher quality distance resolution output image.
