Vehicle-mounted SAR real-time imaging method based on improved distance Doppler algorithm

Abstract

The invention discloses a vehicle-mounted SAR (Synthetic Aperture Radar) real-time imaging method based on an improved distance Doppler algorithm, and relates to a close-range and large-squint-angle imaging algorithm in a vehicle-mounted scene. The method comprises the following steps: dividing a full aperture into sub-apertures; compensating a residual video phase term; doppler frequency shift items caused by fast time are eliminated; carrying out high-order Taylor expansion on the distance and azimuth coupling terms of the signal, constructing a distance migration filter based on the second term of Taylor expansion to carry out distance migration correction, and constructing a secondary distance compression filter based on the third term and higher-order terms of Taylor expansion to carry out secondary distance compression; azimuth compression processing is carried out after distance dimension focusing, and two-dimensional focusing is completed; and after all the sub-apertures are focused, splicing the sub-aperture images to obtain a full-aperture image. According to the invention, high-order Taylor expansion is carried out on the distance and azimuth coupling term, so that the coupling term is accurately compensated, and high-precision near-field imaging can be completed while the advantage of the classical RDA imaging speed is kept.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle-mounted SAR imaging, and in particular to a real-time vehicle-mounted imaging method with an improved RDA. Background Art

[0002] As a core component in the autonomous driving field, automotive millimeter-wave radar has become a research focus due to its advantages, including all-weather operation, miniaturization, high integration, and key sensing capabilities. 77GHz automotive millimeter-wave radar is widely used to measure the radial range, velocity, and angular position of targets in the vehicle environment. However, existing millimeter-wave radar technology struggles to meet the resolution requirements of autonomous driving. Numerous studies have attempted to utilize synthetic aperture radar (SAR) technology to improve azimuth resolution and enhance the accuracy of vehicle environmental perception.

[0003] Synthetic aperture radar (SAR) is a technology that uses the relative motion between the target and the radar to form a virtual aperture, thereby acquiring high-resolution images in the azimuth dimension. In recent years, civilian SAR, particularly automotive SAR, has experienced significant growth with the rise of autonomous driving. Compared to traditional SAR platforms, automotive platforms feature greater motion randomness, strict real-time requirements, and a wide beamwidth. As the core technology of automotive SAR systems, automotive SAR imaging algorithms must consider these requirements. Furthermore, the imaging speed of automotive SAR imaging algorithms is crucial. Faster imaging speed means a more robust vehicle's perception of external information and a quicker response to environmental changes.

[0004] Existing SAR imaging algorithms can be roughly divided into three categories: time domain imaging algorithms represented by the back projection algorithm (BPA), frequency domain imaging algorithms represented by the range Doppler algorithm (RDA), and interpolation imaging algorithms represented by the range migration algorithm (RMA).

[0005] BPA is an accurate algorithm for imaging by coherently summing radar echoes. The advantage of BPA is that it can image echo signals of any trajectory and theoretically has no imaging errors. The disadvantages are high computational complexity, low imaging efficiency, and difficulty meeting the real-time imaging requirements. The advantage of RMA is absolute accuracy in theory. The disadvantage is that a large number of interpolation operations lead to low imaging efficiency and poor adaptability to different scenarios. RDA is a representative of frequency domain algorithms. Its outstanding advantage is extremely fast imaging speed, but its imaging accuracy is limited and it is not suitable for situations with large squint and wide beams. In summary, RDA is the only algorithm that can meet the real-time imaging requirements of vehicle-mounted scenarios. However, because traditional RDA cannot perform uniform compensation under vehicle-mounted imaging conditions with large squint and wide beams, it will produce large imaging errors. Therefore, it is necessary to improve traditional RDA to make it suitable for vehicle-mounted scenarios.

[0006] The core assumption of classic RDA is that the echo signal can be represented as the product of two one-dimensional functions. This is the basis for classic RDA's rapid imaging capabilities, but this also leads to strict application requirements and low imaging accuracy. In near-field vehicular scenarios, antenna beamwidths are often large, slant ranges are small, and the range and azimuth dimensions are significantly coupled, making classic RDA unsuitable for SAR imaging in vehicular settings. Summary of the Invention

[0007] Purpose of the invention: The purpose of the present invention is to provide a vehicle-mounted SAR real-time imaging method based on an improved range Doppler algorithm, which can achieve high-precision near-field imaging while retaining the imaging speed advantage of classic RDA.

[0008] Technical solution: In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0009] A vehicle-borne SAR real-time imaging method based on an improved range Doppler algorithm comprises:

[0010] Divide the full aperture into subapertures;

[0011] Constructing a residual video phase filter to compensate for the residual video phase term;

[0012] Construct a Doppler frequency shift filter to compensate for the Doppler frequency shift caused by fast time;

[0013] Perform high-order Taylor expansion on the range-azimuth coupling term of the signal, construct a range migration filter based on the second term of the Taylor expansion to perform range migration correction, and construct a quadratic range compression filter based on the third term and higher-order terms of the Taylor expansion to perform quadratic range compression;

[0014] After focusing in the range dimension, an azimuth compression filter is constructed based on the first term of Taylor expansion to perform azimuth compression processing and complete two-dimensional focusing.

[0015] After all sub-apertures are focused, the sub-aperture images are stitched together to obtain the full-aperture image.

[0016] Preferably, in full aperture segmentation, based on the difference between the maximum irradiation distance and the minimum irradiation distance, the constraints of the antenna beam width are fully considered, and sub-aperture division is performed and zeros are padded at both ends of the aperture.

[0017] As a preferred method, in the compensation of the residual video phase, the signal is compensated in the range frequency domain, and the residual video phase filter is constructed as follows:

[0018]

[0019] Where γ represents the modulation frequency, f r Indicates distance frequency.

[0020] As a preferred method, in the compensation of Doppler frequency shift caused by fast time, the signal is converted into the range, time and azimuth frequency domains for compensation, and the Doppler frequency shift filter is constructed as

[0021] H RDPC (τ,f a )=exp(-j2πτf a )

[0022] Where τ represents fast time, f a Indicates the azimuth frequency.

[0023] As a preferred method, in the range migration correction and the secondary range compression, the range-azimuth coupling term of the signal is Taylor expanded, and the range-azimuth coupling term of the signal is expressed as

[0024]

[0025] Where R0 represents the reference distance, represents the oblique viewing angle, v represents the moving speed of the vehicle platform, and f c represents the carrier frequency, c represents the speed of light, and γ represents the modulation frequency;

[0026] Expand the coupling term to get

[0027]

[0028] where k i represents the coefficient of Taylor expansion, and the expansion phase is

[0029]

[0030] Where λ represents the carrier wavelength, f aM is the maximum Doppler shift.

[0031] As a preferred method, the distance migration filter constructed is

[0032]

[0033] The constructed quadratic distance compression filter is

[0034]

[0035] As a preference, after focusing in the range dimension, azimuth compression processing is performed to obtain a two-dimensional focused image, wherein the signal after focusing in the range dimension is

[0036]

[0037] where w(f a ) represents the signal azimuth envelope, A represents the signal amplitude, T p Indicates the signal time width, x0 indicates the position of the radar at this time,

[0038] Construct the azimuth compression filter as

[0039]

[0040] After azimuth compression, the two-dimensional focusing signal obtained by azimuth Fourier transform is:

[0041]

[0042] Among them B d represents the Doppler bandwidth, t a Indicates the slow time of radar movement.

[0043] Preferably, in splicing the sub-aperture images to obtain the full-aperture image, redundant information of each sub-image is removed according to the overlap rate set when dividing the sub-apertures and the size of the zero padding for the sub-apertures to obtain the full-aperture image.

[0044] A computer system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm are implemented.

[0045] A computer program product includes a computer program. When the computer program is executed by a processor, the computer program implements the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm.

[0046] Beneficial effects: Existing SAR imaging algorithms cannot achieve fast imaging while ensuring imaging accuracy in vehicle-mounted scenarios, which severely limits the application of SAR imaging technology in the field of autonomous driving. Although existing vehicle-mounted algorithms such as BPA and RMA can ensure imaging accuracy, the imaging speed is slow due to the need for interpolation during the imaging process, and the imaging time is unacceptable for autonomous driving systems. Traditional RDA can achieve fast imaging but cannot be applied to vehicle-mounted scenarios with severe range and azimuth coupling. Compared with the existing technology, the present invention provides a vehicle-mounted SAR real-time imaging method based on improved RDA, which improves the traditional RDA, accurately compensates for the high-order terms of the Taylor expansion of the range and azimuth coupling terms, and obtains better imaging effects, so that it can accurately image vehicle-mounted scenes while retaining the imaging speed.

[0047] In summary, the present invention improves the traditional RDA algorithm and can achieve high-precision imaging of vehicle-mounted scenes while ensuring imaging accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 This is the geometric relationship diagram of vehicle-mounted millimeter-wave radar SAR imaging;

[0050] Figure 2 is an algorithm flow chart of an embodiment of the present invention;

[0051] Figure 3 Schematic diagram of sub-aperture division;

[0052] Figure 4 Schematic diagram of sub-image stitching;

[0053] Figure 5 The simulation imaging and contour map results of the proposed algorithm are shown in Figure 2.

[0054] Figure 6 The optical image of the proposed algorithm in a simple building scene and the imaging results of each algorithm;

[0055] Figure 7 The optical image of the proposed algorithm in a complex parking lot scene and the imaging results of each algorithm. DETAILED DESCRIPTION

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0057] Figure 1 The figure shows the schematic diagram of SAR imaging in a vehicle-mounted scenario. The green area is the imaging area, the light blue area is the radar illumination range, the X-axis is the azimuth dimension, which is also the vehicle's driving direction, and the Y-axis is the distance dimension. Point P(x0,y0,z0) is the corresponding point of the vehicle at a certain moment in the reference distance, and Q(x0+x n ,y0,z0) is a certain imaging target point, R is the instantaneous slant range from the radar to the target at that moment, and R0 is the reference distance. The radar is fixed at height H.

[0058] The embodiment of the present invention discloses a vehicle-mounted SAR real-time imaging method based on an improved range Doppler algorithm. By performing a high-order Taylor expansion on the range-azimuth coupling term and then accurately compensating the coupling term, high-precision near-field imaging can be achieved while retaining the imaging speed advantage of the classic RDA. The specific process of this method is as follows: Figure 2 As shown, it specifically includes the following steps:

[0059] Step 1: To reduce cumulative error and improve computational efficiency, the full aperture is divided into n sub-apertures of equal length. The specific process includes:

[0060] Set the reference distance and calculate the maximum length of the sub-aperture based on the inherent parameters of the system such as the beam width, and round it up to get the appropriate sub-aperture length. Set the overlap rate α, and divide the full aperture with the appropriate sub-aperture length, dividing the full aperture into n sub-apertures with overlapping lengths of αL. sub The sub-aperture segmentation diagram is shown in the figure below: Figure 3 As shown in the figure, L a Indicates the azimuth dimension length of the data, L sub Indicates the actual length of the subaperture, L zp Indicates the length of padding at both ends of the original data.

[0061] Step 2: Compensate the residual video phase term for each original echo data, specifically including:

[0062] (2.a) The original echo data after de-chirp can be expressed as

[0063]

[0064] Where A represents the signal amplitude, t a Indicates slow time, T p Indicates the signal time width, f c represents the carrier frequency, γ represents the modulation frequency, rect(·) represents the rectangular window function, g(·) represents the azimuth dimension envelope, T a Indicates the signal azimuth time width, t n represents the transmission time of the nth signal, Δτ represents the round-trip delay of the signal, and τ represents the fast time.

[0065] (2.b) The RVP compensation filter is constructed as follows

[0066]

[0067] where f r represents the distance frequency;

[0068] After compensation in the range frequency domain, transform back to the two-dimensional time domain to obtain

[0069]

[0070] Step 3, compensating for the Doppler frequency shift caused by fast time;

[0071] (3.a) The azimuth frequency domain signal can be obtained by the stationary phase method (POSP) as

[0072]

[0073] Where G(·) represents the azimuth frequency domain envelope, R0 represents the reference distance, represents the oblique viewing angle, v represents the moving speed of the vehicle platform, and f c represents the carrier frequency, c represents the speed of light, f a represents the azimuth frequency, and x0 represents the position of the radar at this time.

[0074] (3.b) The RDP compensation filter is

[0075] H RDPC (τ,f a )=exp(-j2πτf a )

[0076] After compensating the RDP term, the signal can be expressed as

[0077]

[0078] Step 4: Taylor expand and compensate the signal range-azimuth coupling term. The phase of the signal range-azimuth coupling term after RDP compensation is:

[0079]

[0080] (4.a) In the vehicle-mounted scenario, γτ<<f c Therefore, the Taylor expansion of the coupling term at γτ = 0 is

[0081]

[0082] in

[0083]

[0084] Where λ represents the carrier wavelength, f aM is the maximum Doppler shift;

[0085] That is, the unfolded phase is

[0086]

[0087] The first term in the above formula is the azimuth frequency modulation term, which is compensated during azimuth compression; the second term is the RCM term, which needs to be compensated to obtain range information; the third and fourth terms are high-order coupling terms of range-Doppler, which also need to be compensated.

[0088] (4.b) In order to compensate for the RCM term, the range migration correction filter is constructed as

[0089]

[0090] In order to compensate for the high-order coupling terms, the quadratic range compression filter is constructed as

[0091]

[0092] (4.c) After compensating the range migration correction filter and the secondary range compression filter, the signal becomes

[0093]

[0094] Performing distance-dimensional Fourier transform on the signal yields

[0095]

[0096] At this point, the signal has completed distance-dimensional focusing.

[0097] Step 5: Perform azimuth compression processing on the signal after range-dimensional focusing, and complete sub-aperture two-dimensional focusing.

[0098] Specifically include:

[0099] (5.a) Construct the azimuth compression filter as

[0100]

[0101] (5.b) Compensate the azimuth compression filter and perform azimuth-dimensional Fourier transform to obtain

[0102]

[0103] Among them B d Indicates the Doppler bandwidth.

[0104] Step 6: After all sub-apertures are focused in two dimensions, the redundant parts of each sub-aperture are removed according to the overlap ratio set in step 1, and the effective images in each sub-aperture are stitched together to obtain a full-aperture SAR image. Figure 4 As shown, where L sti Indicates the actual splicing length of the sub-apertures.

[0105] The effectiveness of the present invention can be further illustrated by the following experiments.

[0106] (1) SAR simulation parameters and results

[0107] The key parameters of the simulation experiment are listed in Table 1. The two-dimensional coordinates of the three target points are (-0.5, R0+2), (0, R0) and (0.5, R0-2), and the oblique angle is 15°. The azimuth coordinate and distance coordinate of the central target point are 0 and R0 respectively. Figure 5 As shown in the simulation results, each imaging point is focused at the correct location, and the contour plot exhibits a standard two-dimensional sinc function shape, demonstrating the effectiveness of the proposed algorithm. The peak sidelobe ratio (PSLR) and integrated sidelobe ratio (ISLR) of the three target points in the azimuth dimension are shown in Table 2. The PSLR of the standard sinc function is -13.26 dB, and the ISLR is -9.39 dB. As shown in Table 2, the PSLR and ISLR of the three target points are similar to those of the standard sinc function, demonstrating the imaging capability of the proposed algorithm.

[0108] Table 1 Vehicle SAR simulation parameters

[0109]

[0110] Table 2 Azimuth performance analysis of selected targets

[0111]

[0112] It can be seen that the performance parameters of all targets processed by the proposed algorithm are close to the theoretical values ​​of the sinc function. In addition, the contour map shows that each target point presents a standard two-dimensional sinc shape, which proves the imaging capability of the proposed algorithm.

[0113] (2) Measured data processing

[0114] The measured radar system operates at 77 GHz in equivalent spotlight mode. To highlight the advantages of the proposed algorithm in imaging time and quality, the RMA and BPA algorithms were compared in terms of imaging quality and time. The system parameters for the measured data are shown in Table 3. The oblique viewing angle is approximately 12°.

[0115] Table 3 Measured parameters of vehicle-mounted SAR

[0116]

[0117] The experimental scenarios used were a simple building scene and a parking lot, two representative scenes in automotive radar applications. In this experiment, the subaperture overlap ratio was set to α = 0.5 and the reference distance R0 = 1m. Based on this, the simple building scene and the parking lot scene were divided into 7 subapertures.

[0118] The optical images and imaging results of the experiment are as follows Figure 6 、 Figure 7 As shown in the experimental results, Figure (a) is the actual optical scene image, and Figures (b), (c), and (d) are the imaging results of the corresponding optical scenes using the improved range Doppler algorithm (Improved RDA), the range migration algorithm (RMA), and the back projection algorithm (BPA), respectively. From the imaging results, it can be found that the imaging results of the three algorithms can well represent the position and reflection coefficient information of the object. The imaging time of the three algorithms in these two scenarios is shown in Table 4. The imaging time of Improved RDA is much shorter than that of BPA and RMA. Improved RDA can complete real-time high-definition imaging of vehicle-mounted scenes while ensuring imaging quality.

[0119] Table 4 Algorithm imaging time

[0120]

[0121] In summary, the present invention designs an improved range Doppler algorithm that can be applied to real-time SAR imaging in vehicle-mounted scenarios, significantly improving the possibility of applying SAR imaging in the field of autonomous driving and providing technical support for the development of autonomous driving.

[0122] An embodiment of the present invention also discloses a computer system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm are implemented.

[0123] An embodiment of the present invention further discloses a computer program product, including a computer program. When the computer program is executed by a processor, the computer program implements the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm.

[0124] Anything not described in detail in the present invention is well known to those skilled in the art.

[0125] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A vehicle-mounted SAR real-time imaging method based on an improved range Doppler algorithm, characterized in that: The method includes: Divide the full aperture into subapertures; Constructing a residual video phase filter to compensate for the residual video phase term; Construct a Doppler frequency shift filter to compensate for the Doppler frequency shift caused by fast time; Perform high-order Taylor expansion on the range-azimuth coupling term of the signal, construct a range migration filter based on the second term of the Taylor expansion to perform range migration correction, and construct a quadratic range compression filter based on the third term and higher-order terms of the Taylor expansion to perform quadratic range compression; After focusing in the range dimension, an azimuth compression filter is constructed based on the first term of Taylor expansion to perform azimuth compression processing and complete two-dimensional focusing. After all sub-apertures are focused, the sub-aperture images are stitched together to obtain the full-aperture image.

2. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: In full aperture segmentation, based on the difference between the maximum illumination distance and the minimum illumination distance, and taking full account of the constraints of the antenna beam width, sub-aperture division is performed and zeros are padded at both ends of the aperture.

3. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: In the compensation of the residual video phase, the signal is compensated in the distance frequency domain and the residual video phase filter is constructed as Where γ represents the modulation frequency, f r Indicates distance frequency.

4. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: In the compensation of Doppler frequency shift caused by fast time, the signal is transformed into the range, time and azimuth frequency domains for compensation, and the Doppler frequency shift filter is constructed as H RDPC (t,f a )=exp(-j2πτf a ) Where τ represents fast time, f a Indicates the azimuth frequency.

5. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: In range migration correction and quadratic range compression, the range-azimuth coupling term of the signal is Taylor expanded and the range-azimuth coupling term of the signal is expressed as Where τ represents fast time, f a represents the azimuth frequency, R0 represents the reference distance, θ represents the oblique angle, v represents the moving speed of the vehicle platform, and f c represents the carrier frequency, c represents the speed of light, and γ represents the modulation frequency; Expand the coupling term to get where k i represents the coefficient of Taylor expansion, and the expansion phase is Where λ represents the carrier wavelength, f aM is the maximum Doppler shift.

6. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 5, characterized in that: The distance migration filter constructed is The constructed quadratic distance compression filter is 7. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: After focusing in the range dimension, azimuth compression processing is performed to obtain a two-dimensional focused image, where the signal after focusing in the range dimension is where f r Represents the distance frequency, f a Indicates the azimuth frequency, w(f a ) represents the signal azimuth envelope, A represents the signal amplitude, T p represents the signal time width, R0 represents the reference distance, θ represents the oblique angle, v represents the moving speed of the vehicle platform, and f c represents the carrier frequency, c represents the speed of light, γ represents the modulation frequency, and x0 represents the position of the radar at this time. f aM is the maximum Doppler shift; Construct the azimuth compression filter as After azimuth compression, the two-dimensional focusing signal obtained by azimuth Fourier transform is: Among them B d represents the Doppler bandwidth, t a Indicates the slow time of radar movement.

8. The vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to claim 1, characterized in that: In stitching the sub-aperture images to obtain the full-aperture image, redundant information of each sub-image is removed and stitched together to obtain the full-aperture image according to the overlap ratio set when dividing the sub-aperture and the size of the zero padding for the sub-aperture.

9. A computer system comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the computer program is executed by a processor, the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to any one of claims 1 to 8 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the vehicle-mounted SAR real-time imaging method based on the improved range Doppler algorithm according to any one of claims 1 to 8 are implemented.

Images