Method for imaging ultra-high-speed target by ZYNQ-based step frequency radar
By using the frequency domain dual pulse group algorithm based on the ZYNQ platform in step frequency radar, the problem of accurate imaging during ultra-high-speed target imaging is solved, and efficient and real-time high-resolution distance imaging is achieved.
Patent Information
- Application Number
- CN202510000039.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-01
- Publication Date
- 2025-05-30
AI Technical Summary
Existing step frequency radars are difficult to achieve accurate imaging when facing ultra-high-speed targets, especially when the target speed is too high to use phase cancellation technology, the imaging effect is poor.
The frequency domain dual pulse group algorithm based on the ZYNQ platform is adopted to alternately transmit two sets of pulse train signals of different repetition periods, and phase cancellation is performed in the frequency domain to achieve high-resolution distance imaging.
This method has small calculation volume, is easy to implement, and has strong real-time performance. It is suitable for imaging ultra-high-speed targets. It solves the problem that the time domain dual-pulse algorithm cannot accurately image under ultra-high-speed targets, and improves imaging efficiency.
Smart Images
Figure CN120065216A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of radar high-resolution imaging, and particularly relates to an imaging method for ultra-high-speed targets by a stepped-frequency radar based on ZYNQ. Background Technique
[0002] The target range resolution of a radar refers to the minimum distance that can distinguish two scattering points. It is the effective bandwidth B of the radar signal that determines the range resolution. The wider the effective bandwidth, the better the range resolution. A stepped-frequency radar synthesizes a large-bandwidth signal by transmitting a series of pulse signals, and the operating frequency of each pulse signal is uniformly stepped based on the center frequency, thereby improving the range resolution. At the same time, the stepped-frequency signal also has the advantage of reducing the requirement for the instantaneous bandwidth of the radar signal prototype, so it has attracted much attention.
[0003] A stepped-frequency radar calculates the target range by comparing the phase differences between signals of different frequencies. However, if the target has radial motion, this phase change not only contains range information but also contains the phase change caused by the Doppler frequency shift, which will affect the imaging accuracy. Therefore, precise compensation of the phase transformation must be performed during pulse coherent synthesis, otherwise range migration and waveform distortion will occur. There are usually two methods in the prior art for precisely compensating target motion: one is to first estimate the target speed and then compensate for imaging; the other is to design the waveform and directly cancel the phase change term caused by the target speed through phase cancellation technology to finally obtain an accurate range image.
[0004] However, both of the above two methods have their limitations. The method of estimating the target speed has a large amount of calculation and a long time consumption, which is not conducive to hardware implementation. And for the method of waveform design, when the target speed is large enough to cause the movement of the echo envelope, the echo signal at the end often cannot use the phase cancellation technology. Therefore, when facing ultra-high-speed targets, the first method has a large amount of calculation, and the second method has the limitation of being unable to correctly image. Summary of the Invention
[0005] The purpose of the present invention is to provide an imaging method for ultra-high-speed targets by a stepped-frequency radar based on ZYNQ, which has a small amount of calculation, is easy to implement, has strong real-time performance, and is particularly suitable for imaging ultra-high-speed targets.
[0006] The technical solution for achieving the purpose of the present invention is: an imaging method for ultra-high-speed targets by a stepped-frequency radar based on ZYNQ, including the following steps:
[0007] (10) Radar echo data acquisition: Based on the ZYNQ platform, use the AD9238 board to collect two groups of target echo signals with different pulse repetition periods received by the radar signal prototype.
[0008] (20) Data storage: The echo data is stored in DDR4 and RAM through two channels. One channel stores the data in DDR4 through DMA transmission, and the other channel directly stores the data in RAM;
[0009] (30) Target detection: The method of pulse accumulation is used for target detection to determine the location of the target center;
[0010] (40) Algorithm implementation: The frequency-domain double pulse group algorithm is used to perform frequency-domain phase cancellation on the pulse train signals with two different repetition periods;
[0011] (50) Double pulse group imaging: The inverse Fourier transform in the slow time dimension is performed on the echo data using the FFT IP core, and then the maximum value method is used for range image stitching to obtain the high-resolution range image of the target;
[0012] (60) Data transmission: The high-resolution range image data is transmitted to the host computer through the serial port.
[0013] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of the above method are implemented.
[0014] A computer-readable storage medium stores a computer program, and when the program is executed by a processor, the steps of the above method are implemented.
[0015] Compared with the prior art, the significant advantages of the present invention are:
[0016] (1) Small computational amount, easy to implement, and strong real-time performance: The traditional method of estimating motion parameters is achieved through a search process, which has a large computational amount and is difficult for real-time imaging. This method performs phase cancellation in the frequency domain by alternately transmitting two groups of pulse train signals with different repetition periods, and does not require motion parameter estimation through an iterative algorithm, which makes the imaging efficient.
[0017] (2) Suitable for imaging ultra-high-speed targets: When the time-domain double pulse group is used for imaging ultra-high-speed targets, due to the too high target speed, the echo of the last pulse of the first group of pulse trains and the echo of the last pulse of the second group of pulse trains often do not overlap, and at this time, the time-domain double pulse group algorithm cannot be used. Different from the time-domain double pulse group algorithm, the method of the present invention uses the frequency-domain double pulse group phase cancellation method. When imaging ultra-high-speed targets, the pulse echo signals with different repetition periods overlap in the frequency domain, so this method is suitable for imaging ultra-high-speed targets.
[0018] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Description of the Drawings
[0019] Figure 1This is the main flowchart of the imaging method for ultra-high-speed targets using a stepped-frequency radar based on ZYNQ in the present invention.
[0020] Figure 2 For Figure 1 This is the flowchart of the radar echo data acquisition step in
[0021] Figure 3 For Figure 1 This is the flowchart of the data storage step in
[0022] Figure 4 For Figure 3 This is the flowchart of storing echo data into RAM in
[0023] Figure 5 For Figure 1 This is the flowchart of the target detection step in
[0024] Figure 6 For Figure 1 This is the flowchart of the algorithm implementation step in
[0025] Figure 7 For Figure 1 This is the flowchart of the dual-pulse group imaging step in
[0026] Figure 8 This is the hardware imaging diagram of the dual-pulse group in the frequency domain of ZYNQ.
[0027] Figure 9 This is the comparison diagram of MATLAB imaging and ZYNQ hardware imaging. Specific implementation mode
[0028] As Figure 1 shown, the imaging method for ultra-high-speed targets using a stepped-frequency radar based on ZYNQ in the present invention includes the following steps:
[0029] (10) Radar echo data acquisition: Based on the ZYNQ platform, use the AD9238 board to collect two groups of target echo signals with different pulse repetition periods received by the radar signal prototype;
[0030] As Figure 2 shown, the (10) radar echo data acquisition step includes:
[0031] (11) Detect the trigger signal: Detect the TRIG pulse signal input from the radar signal prototype to the ZYNQ board. When the TRIG pulse signal is the trigger signal and the time interval between two TRIGs exceeds 400 microseconds, it is judged as the first pulse data of a frame of data, and at this time, start collecting echo data;
[0032] (12) Delayed Output: According to the digital output timing of the used AD acquisition card, there is an interval of 7 sampling clock cycles between the start of data acquisition and digital output.
[0033] (20) Data Storage: The echo data is stored in DDR4 and RAM through two paths. One path is stored in DDR4 through DMA transmission, and the other path is directly stored in RAM;
[0034] As Figure 3 shown, the (20) data storage step includes:
[0035] (21) Storing Echo Data into DDR4: The digital output data of the AD acquisition card is stored in the DDR4 of the ARM side of the ZYNQ board through the DMAIP core. Since the clock frequency of the AXI data bus of the DMAIP core is not equal to the sampling clock frequency, a FIFO is used for cross-clock domain data processing between the digital output of the acquisition card and DMA;
[0036] (22) Storing Echo Data into RAM: Select a RAM memory for data grouping and caching. By judging the pulse group type, pulse sequence number, and sampling point sequence number where each sampling point is located, the data of each sampling point is written into the corresponding address space;
[0037] (23) Outputting Flag Signals: One flag signal is output for each pulse data stored in RAM, one flag signal is output after storing one pulse group, and one flag signal is output after storing one frame of data.
[0038] As Figure 4 shown, the method steps for judging the pulse group type, pulse sequence number, and sampling point sequence number where each sampling point is located in step (22) include:
[0039] Step 1: Initialize the pulse counter, sampling point counter, and number of sampling points;
[0040] Step 2: Judge whether it is the first data of a frame;
[0041] Step 3: Pulse trigger;
[0042] Step 4: Judge whether the pulse counter is less than 192. If it is greater than or equal to 192, it is determined as the second pulse train; if it is less than 192, take the remainder of the pulse counter divided by 3. If the remainder is 1, it is judged as the second pulse train; if the remainder is 0 or 2, it is judged as the first pulse train;
[0043] Step 5: Start sampling. The sampling counter is incremented by 1. After counting to the number of sampling points, the pulse counter is incremented by 1;
[0044] Step 6: Judge whether the pulse counter is less than 256;
[0045] Step 7: When the pulse counter is less than 256, repeat Steps 3 to 6; otherwise, end the loop.
[0046] (30) Target detection: The method of pulse accumulation is adopted for target detection to determine the position of the target center.
[0047] As Figure 5 shown, the (30) target detection step includes:
[0048] (31) Judging the pulse group number: Use the first group of pulses for target detection. Use the remainder method of the counter to judge whether it is Pulse Group 1. Add 1 to the counter for each incoming pulse. When the counter value is less than 192, take the remainder of 3. When the remainder is not equal to 2, it is judged as the first group of pulses.
[0049] (32) Finding the square of the modulus value: Find the square of the modulus value of the sampling data of each pulse in the first group of pulses.
[0050] (33) Accumulative summation: Accumulatively sum the squares of the modulus values of the sampling data of all pulses in the first group of pulses in the slow time dimension.
[0051] (34) Size comparison: Compare the sizes of the accumulative sum values at different sampling points.
[0052] (35) Outputting the position of the maximum value: Output the position of the sampling point with the largest accumulative sum value.
[0053] (40) Algorithm implementation: Use the frequency domain double pulse group algorithm to perform frequency domain phase cancellation on the pulse train signals with two different repetition periods.
[0054] As Figure 6 shown, the (40) algorithm implementation step includes:
[0055] (41) Setting the detection gate width: According to the radar parameters, the actual distance corresponding to adjacent sampling points is 7.5 meters, and the overall width of the airborne target is generally less than 60 meters. Therefore, 8 sampling points are selected as the detection gate.
[0056] (42) Fast time dimension Fourier transform: Send the 8 sampling point data of all pulses in the double pulse group into the FFT IP core, and configure the FFT IP core into the Fourier transform mode to perform the fast time dimension Fourier transform.
[0057] (43) Double pulse combination: Use the multiplier IP core Multiplier and the cyclic divider module to perform double pulse combination according to the formula where X 1 (f,nT r ) and X 2 (f,nT r ) are composed of r 1(mT s , nT r ) and r 2 (mT s , nT r ) is obtained by performing a Fourier transform in the fast time dimension, r 1 (mT s , nT r ) and r 2 (mT s , nT r ) are the baseband signals after mixing the nth echo signals of the first pulse group and the second pulse group with the reference signal. m is the number of sampling points, and T s is the sampling time, and T r is the pulse repetition period;
[0058] (44) Inverse Fourier transform in the fast time dimension: The FFT IP core is configured in the inverse Fourier transform mode. The X(f, nT r ) data is sent into the FFT IP core, and an inverse Fourier transform is performed along the fast time dimension to obtain x(mT s , nT r ). The frequency-domain double-pulse combined signal is converted into the time domain. Here, n represents the nth echo signal, m is the number of sampling points, and T s is the sampling time, and T r is the pulse repetition period.
[0059] (50) Double-pulse group imaging: Use the FFT IP core to perform an inverse Fourier transform on the echo data in the slow time dimension, and then use the maximum value method to perform range image stitching to obtain the target high-resolution range image;
[0060] As Figure 7 shown, the steps of the (50) double-pulse group imaging include:
[0061] (51) Inverse Fourier transform in the slow time dimension: The FFT IP core is configured in the inverse Fourier transform mode. The x(mT s , nT r ) data is sent into the FFT IP core, and an inverse Fourier transform is performed along the slow time dimension, that is, along the nT r dimension, to obtain the target high-resolution range image;
[0062] (52) Synthesize the one-dimensional range image: The high-resolution range image results obtained at each sampling point are stitched through the same-range maximum-point trace extraction algorithm. The same-range maximum method selects the maximum amplitude value as the processing result of the current range by comparing the repeated information in the high-resolution range images of each sampling point, and stitches the full-range processing results in distance order to obtain the real one-dimensional range image based on ZYNQ, as Figure 8 shown.
[0063] (60)Data transmission: Transmit the high-resolution range profile data to the host computer through the serial port.
[0064] The (60) data transmission transmits the range profile data and the original echo data to the host computer through the serial port, uses MATLAB for data analysis and verification, and compares whether the MATLAB imaging data is consistent with the ZYNQ imaging data. If they are consistent, it indicates that the method of the present invention is feasible. As Figure 9 shown.
[0065] The method of the present invention completes the frequency-domain double pulse group algorithm through ZYNQ, can realize signal processing while collecting data, has high real-time imaging efficiency, reduces the system complexity by flexibly configuring hardware resources, and the frequency-domain double pulse group algorithm has a small amount of calculation and is easy to implement. It also solves the problem that the time-domain double pulse group algorithm cannot accurately image due to the movement of the target echo envelope, improves the application range of the method of the present invention, and improves the imaging efficiency.
Claims
1. A method for ultra-high-speed target imaging using a stepped frequency radar based on ZYNQ, characterized in that: The following steps are involved: (10) Radar echo data acquisition: Based on the ZYNQ platform, the AD9238 board is used to collect two sets of target echo signals with different pulse repetition periods received by the radar signal prototype; (20) Data storage: The echo data is stored in DDR4 and RAM through two paths, one of which is stored in DDR4 through DMA transfer, and the other is directly stored in RAM; (30) Target detection: Target detection is performed using the pulse accumulation method to determine the location of the target center; (40) Algorithm implementation: Use the frequency domain dual pulse group algorithm to perform frequency domain phase cancellation on two pulse train signals with different repetition periods; (50) Dual-pulse imaging: Use the FFT IP core to perform slow-time inverse Fourier transform on the echo data, and then use the maximum value method to stitch the range images to obtain a high-resolution range image of the target; (60) Data transmission: The high-resolution range image data is transmitted to the host computer through the serial port.
2. According to claim 1, a method for ultra-high-speed target imaging using a ZYNQ-based stepped frequency radar, characterized in that: The (10) radar echo data collection step comprises: (11) Detecting trigger signals: Detecting the TRIG pulse signal input from the radar signal prototype to the ZYNQ board. The TRIG pulse signal is the trigger signal. When the interval between two TRIGs exceeds 400 microseconds, it is determined to be the first pulse data of a frame of data, and the echo data collection starts at this time. (12) Delayed output: According to the digital output timing of the AD acquisition card used, there is an interval of 7 sampling clock cycles from the start of data acquisition to digital output.
3. The method for ultra-high-speed target imaging using a ZYNQ-based stepped frequency radar according to claim 1, characterized in that: The (20) data storage step comprises: (21) Echo data is stored in DDR4: The digital output data of the AD acquisition card is stored in the DDR4 of the ZYNQ board ARM side through the DMAIP core. Since the clock frequency used by the AXI data bus of the DMAIP core is different from the sampling clock frequency, FIFO is used between the digital output of the acquisition card and DMA for cross-clock domain data processing; (22) Storing echo data into RAM: Using RAM memory for data grouping caching, the data of each sampling point is written into the corresponding address space by judging the pulse group type, pulse number, and sampling point number of each sampling point; (23) Output flag signal: A flag signal is output each time a pulse data is stored in the RAM. A flag signal is output after a pulse group is stored. A flag signal is output after a frame of data is stored.
4. A method for ultra-high-speed target imaging based on ZYNQ stepped frequency radar according to claim 3, characterized in that: The method for determining the pulse group type, pulse sequence number, and sampling point sequence number of each sampling point in step (22) comprises the following steps: Step 221, initializing the pulse counter, sampling point counter and sampling point number; Step 222, determining whether it is the first data of a frame; Step 223, pulse triggering; Step 224, determine whether the pulse counter is less than 192, if it is greater than or equal to 192, it is determined to be the second group of pulse trains; if it is less than 192, use the pulse counter to calculate the remainder of 3, if the remainder is 1, it is determined to be the second group of pulse trains, and if the remainder is 0 or 2, it is determined to be the first group of pulse trains; Step 225, start sampling, the sampling counter is incremented by 1, and after counting to the number of sampling points, the pulse counter is incremented by 1; Step 226, determine whether the pulse counter is less than 256; Step 227: When the pulse counter is less than 256, repeat steps 3 to 6; otherwise, end the loop.
5. The method for ultra-high-speed target imaging based on ZYNQ stepped frequency radar according to claim 1, characterized in that: The target detection step (30) comprises: (31) Determine the pulse group number: Use the first pulse group to detect the target, and use the counter remainder method to determine whether it is pulse group 1. The counter is incremented by 1 for each incoming pulse. When the counter value is less than 192, the remainder is calculated with respect to 3. When the remainder is not equal to 2, it is determined to be the first pulse group. (32) Calculating the square of the modulus value: calculating the square of the modulus value of each pulse sampling data of the first group of pulses; (33) Accumulation and summation: the squares of the sampled data modulus values of all pulses in the first group of pulses are accumulated and summed in the slow time dimension; (34) Size comparison: compare the size of the accumulated sum values of different sampling points; (35) Output maximum value position: Output the sampling point position with the largest cumulative sum value.
6. The method for ultra-high-speed target imaging based on ZYNQ stepped frequency radar according to claim 1, characterized in that: The (40) algorithm implementation steps include: (41) Setting the detection gate width: According to the radar parameter setting, the actual distance between adjacent sampling points is 7.5 meters, and the overall width of the aerial target is less than 60 meters, so 8 sampling points are selected as the detection gate; (42) Fast time-dimensional Fourier transform: Send the 8 sampling point data of all pulses in the dual pulse group to the FFT IP core, and configure the FFT IP core to the Fourier transform mode for fast time-dimensional Fourier transform; (43) Dual-pulse combination: Use the multiplier IP core Multiplier and the circular divider module, according to the formula Double pulse combination, where X1(f,nT r ) and X2(f,nT r ) is composed of r1(mT s ,nT r ) and r2(mT s ,nT r ) is obtained by Fourier transform in the fast time dimension, r1(mT s ,nT r ) and r2(mT s ,nT r ) is the baseband signal of the nth echo signal of the first pulse group and the second pulse group after mixing with the reference signal, m is the number of sampling points, T s is the sampling time, T r is the pulse repetition period; (44) Fast time-dimensional inverse Fourier transform: The FFT IP core is configured in inverse Fourier transform mode to transform the X(f,nT r ) data is sent to the FFT IP core, and an inverse Fourier transform is performed along the fast time dimension to obtain x(mT s ,nT r ), combine the double pulses in the frequency domain into a signal and convert it into the time domain.
7. The method for ultra-high-speed target imaging using a ZYNQ-based stepped frequency radar according to claim 6, characterized in that: The (50) dual pulse imaging step comprises: (51) Slow-time inverse Fourier transform: The FFT IP core is configured in inverse Fourier transform mode, and x(mT s ,nT r ) data is sent to the FFT IP core and an inverse Fourier transform is performed along the slow time dimension, that is, along the nT r Dimension, get the high-resolution range image of the target; (52) Synthesize one-dimensional range image: The high-resolution range image results obtained from each sampling point are stitched together using the same-distance selection algorithm. The same-distance selection algorithm compares the repeated information in the high-resolution range image of each sampling point and selects the maximum amplitude as the processing result of the current distance. The full-distance processing results are stitched together in order of distance to obtain a true one-dimensional range image.
8. The method for ultra-high-speed target imaging using a ZYNQ-based stepped frequency radar according to claim 1, characterized in that: The (60) data transmission transmits the range image data and the original echo data to the host computer through the serial port, uses MATLAB to perform data analysis and verification, and compares the MATLAB imaging data with the ZYNQ imaging data.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 8 are implemented.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.