GPU-based radar electronic countermeasure intermediate frequency data pulse batch identification and edge detection method

Abstract

The invention discloses a radar electronic countermeasure intermediate frequency data pulse batch identification and edge detection method based on a GPU, and the method comprises the steps: quickly searching a pulse segmentation flag bit in batch intermediate frequency data through the GPU, and generating a pulse segmentation index list; establishing a Hash mapping table from the intermediate frequency data to the pulse segmentation index list so as to facilitate subsequent pulse envelope extraction and edge detection; in the pulse envelope extraction process, extracting envelope data according to pulse separation information by utilizing GPU parallel calculation; a one-dimensional edge detection mode is adopted to detect the upper edge and the lower edge of the pulse so as to determine the starting position and the ending position of the pulse. According to the method, the parallel computing capability of the GPU is fully utilized, the efficiency and the real-time performance of radar electronic countermeasure intermediate frequency data processing are improved, efficient pulse recognition and edge detection are achieved, the method is suitable for processing large-scale data, and the combat requirement of a high-performance radar electronic countermeasure system can be met.

Description

Technical Field

[0001] The invention relates to the field of radar electronic countermeasure signal processing, and in particular to a radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method based on a GPU. Background Art

[0002] In radar electronic countermeasure systems, pulse detection and identification of intermediate frequency data are key steps. Intermediate frequency data is a digital signal obtained by sampling the radar target signal received by the radar electronic countermeasure system at the intermediate frequency, which contains the individual characteristics and radiation parameter characteristics of the target radar radiation source. By performing pulse detection and identification on the intermediate frequency data, the existence of the target radar can be determined and its properties can be analyzed. The traditional CPU processing method has the problems of slow processing speed and poor real-time performance, which makes it difficult to meet the operational requirements of high-performance radar electronic countermeasure systems. Summary of the invention

[0003] The purpose of the present invention is to propose a radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method based on GPU to improve the efficiency and real-time performance of radar electronic countermeasure signal processing.

[0004] The technical solution to achieve the purpose of the present invention is: a radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method based on GPU, comprising the following steps:

[0005] Step 1, generate a pulse segmentation index list: use the GPU to quickly search for the pulse segmentation flag in the batch intermediate frequency data, and generate a pulse segmentation index list in sequence;

[0006] Step 2, establish a hash mapping table from intermediate frequency data to pulse segmentation index list: establish an equal-length pulse segmentation point hash mapping table for the input intermediate frequency data, where each position in the table represents the position of the pulse data at that point in the pulse segmentation index list;

[0007] Step 3, pulse envelope extraction: According to the pulse separation information and pulse width, the pulse is dynamically divided into multiple sub-intervals, a GPU thread is started for each sub-interval, and its maximum value is calculated as the envelope value;

[0008] Step 4, pulse upper and lower edge detection: After the pulse envelope extraction is completed, the pulse upper and lower edges are detected by one-dimensional edge detection to determine the starting and ending positions of the pulse.

[0009] Further, step 1, generating a pulse segmentation index list: using the GPU to quickly search for the pulse segmentation flag in the batch intermediate frequency data, and sequentially generating a pulse segmentation index list, the specific method is:

[0010] Step 1-1: Copy the real-time received batch IF data from the CPU memory to the GPU memory, and the data length is if_lens;

[0011] Step 1-2: Generate the initial pulse segmentation index list if_seg_idx according to the minimum pulse width limit pw_min, the formula is:

[0012] if_seg_idx=zeros(if_lens / / pw_min)

[0013] Where if_seg_idx is an array of all zeros with a length of if_lens / / pw_min;

[0014] Step 1-3: Start a GPU thread for each IF data point in the GPU;

[0015] Step 1-4: In each thread, check whether the point is a pulse split flag;

[0016] Step 1-5: If yes, record its index in the pulse segmentation index list;

[0017] if_seg_idx[thread_idx / / pw_min]=thread_idx

[0018] Step 1-6: Keep only the non-zero values ​​in if_seg_idx.

[0019] Further, step 2, establish a hash mapping table from intermediate frequency data to pulse segmentation index list: establish an equal-length pulse segmentation point hash mapping table for the input intermediate frequency data, each position in the table represents the position of the pulse data at that point in the pulse segmentation index list, the specific method is:

[0020] Step 2-1: Create an array of the same length as the input IF data, called the hash mapping table if_seg_idx_hash_map from IF data to pulse segmentation index list;

[0021] Step 2-2: Initialize if_seg_idx_hash_map to 0;

[0022] Step 2-3: traverse the pulse segmentation index list if_seg_idx, and set the position data corresponding to the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list to 1;

[0023] Step 2-4: Through GPU parallel computing, the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list is accumulated and summed, and then the actual index of each pulse cutting point is calculated.

[0024] Further, step 3, pulse envelope extraction: according to the pulse separation information and pulse width, the pulse is dynamically divided into multiple sub-intervals, a GPU thread is started for each sub-interval, and its maximum value is calculated as the envelope value; the specific method is:

[0025] Step 3-1: Determine the start and end positions of each pulse according to the pulse segmentation index list;

[0026] Step 3-2: For each pulse, divide it into 512 sub-intervals;

[0027] Step 3-3: Start a GPU thread for each sub-interval, parallelize the calculation of the maximum value of the intermediate frequency data in the interval, and use it as the envelope value of the interval.

[0028] Further, step 4, pulse upper and lower edge detection: after the pulse envelope extraction is completed, the upper and lower edges of the pulse are detected by one-dimensional edge detection to determine the starting and ending positions of the pulse. The specific method is:

[0029] Step 4-1: For each segmented pulse envelope one-dimensional signal, establish a GPU sub-thread;

[0030] Step 4-2: In each sub-thread, edge detection is performed on the pulse envelope one-dimensional signal;

[0031] Step 4-3: Using a differential algorithm, when a transition from low to high or from high to low is detected, the position is recorded as the upper and lower edges of the pulse, thereby determining the start and end positions of the pulse;

[0032] Step 4-4: Post-process the upper and lower edge positions.

[0033] Furthermore, before the data is copied to the GPU memory, the batch intermediate frequency data received in real time is preprocessed, including removing noise, filtering or signal enhancement.

[0034] Furthermore, pulse envelope extraction uses Fourier transform or wavelet transform to extract the envelope of the signal, and uses a sliding window function to smooth the signal.

[0035] Furthermore, one-dimensional edge detection uses first-order or second-order difference methods to detect signal changes, or uses a convolution algorithm to find the upper and lower edges of the signal.

[0036] Furthermore, the upper and lower edge positions are post-processed, including removing repeated edges, using filters to remove noise or smooth edge positions, and using pattern recognition or machine learning algorithms to further optimize the pulse detection results.

[0037] A radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection system based on GPU implements the radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method based on GPU to achieve radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection based on GPU.

[0038] Compared with the prior art, the present invention has the following significant advantages: by making full use of the parallel computing capability of GPU, the efficiency and real-time performance of radar electronic countermeasure intermediate frequency data processing are improved, and efficient pulse recognition and edge detection are achieved. The method shows significant performance advantages when processing large-scale data, and can meet the needs of high-performance radar electronic countermeasure systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is the overall flow chart. Implementation

[0040] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0041] The present invention provides a GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method and system, and the specific steps are as follows:

[0042] Step 1: Generate a pulse segmentation index list:

[0043] Step 1-1: Copy the real-time received batch IF data from the CPU memory to the GPU memory, and the data length is if_lens;

[0044] Step 1-2: Generate an initial pulse segmentation index list if_seg_idx according to the minimum pulse width limit pw_min.

[0045] if_seg_idx=zeros(if_lens / / pw_min)

[0046] Step 1-3: Start a GPU thread for each IF data point in the GPU;

[0047] Step 1-4: In each thread, check whether the point is a pulse split flag;

[0048] Step 1-5: If yes, record its index in the pulse segmentation index list;

[0049] if_seg_idx[thread_idx / / pw_min]=thread_idx

[0050] Step 1-6: Keep only the non-zero values ​​in if_seg_idx.

[0051] Step 2: Create a hash mapping table from intermediate frequency data to pulse segmentation index list:

[0052] Step 2-1: Create an array of the same length as the input IF data, called the hash mapping table if_seg_idx_hash_map from IF data to pulse segmentation index list;

[0053] Step 2-2: Initialize if_seg_idx_hash_map to 0;

[0054] Step 2-3: traverse the pulse segmentation index list if_seg_idx, and set the position data corresponding to the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list to 1;

[0055] Step 2-4: Use GPU parallel computing to accumulate and sum the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list, and then calculate the actual index of each pulse cutting point.

[0056] Step 3, pulse envelope extraction;

[0057] Step 3-1: Determine the start and end positions of each pulse according to the pulse segmentation index list;

[0058] Step 3-2: For each pulse, divide it into 512 sub-intervals;

[0059] Step 3-3: Start a GPU thread for each sub-interval, calculate the maximum value of the intermediate frequency data in the interval, and use it as the envelope value of the interval. This step significantly improves the efficiency of envelope extraction through parallel processing. This step extracts the intensity variation characteristics of each pulse, providing key signal features for pulse upper and lower edge detection.

[0060] Step 4: Pulse rising and falling edge detection:

[0061] Step 4-1: For each segmented pulse envelope one-dimensional signal, establish a GPU sub-thread;

[0062] Step 4-2: In each sub-thread, edge detection is performed on the pulse envelope one-dimensional signal;

[0063] Step 4-3: Using a differential algorithm or other one-dimensional signal processing method, when a transition from low to high or from high to low is detected, the position is recorded as the upper and lower edges of the pulse. By using a one-dimensional edge detection method, the upper and lower edge positions of the pulse can be effectively detected, thereby determining its starting and ending positions;

[0064] Step 4-4: Post-process the upper and lower edge positions as required, such as removing duplicate edges or performing filtering to improve the accuracy and stability of detection.

[0065] Preferably, preprocessing of the intermediate frequency data: In step 1, preprocessing of the batch intermediate frequency data received in real time may be considered, such as removing noise, filtering or signal enhancement, to improve the accuracy and efficiency of subsequent processing steps. This may be done before the data is copied to the GPU memory to ensure the quality and stability of the processed data.

[0066] Preferably, in step 3, the extraction of the pulse envelope is mentioned, and Fourier transform or wavelet transform can be used to extract the envelope of the signal, and a sliding window or other window function can also be considered to smooth the signal to reduce noise interference.

[0067] Preferably, in step 4, a one-dimensional edge detection method is mentioned to determine the starting and ending positions of the pulse. The first-order or second-order difference method can be used to detect the change of the signal, and the convolution algorithm or other digital signal processing techniques can also be considered to find the upper and lower edges of the signal.

[0068] Preferably, in step 4-4, it is mentioned that the upper and lower edge positions are post-processed, and a filter can be used to remove noise or smooth the edge position, and pattern recognition or machine learning algorithms can also be considered to further optimize the pulse detection results.

[0069] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0070] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method, characterized in that: The steps include: Step 1, generate a pulse segmentation index list: use the GPU to quickly search for the pulse segmentation flag in the batch intermediate frequency data, and generate a pulse segmentation index list in sequence; Step 2, establish a hash mapping table from intermediate frequency data to pulse segmentation index list: establish an equal-length pulse segmentation point hash mapping table for the input intermediate frequency data, where each position in the table represents the position of the pulse data at that point in the pulse segmentation index list; Step 3, pulse envelope extraction: According to the pulse separation information and pulse width, the pulse is dynamically divided into multiple sub-intervals, a GPU thread is started for each sub-interval, and its maximum value is calculated as the envelope value; Step 4, pulse upper and lower edge detection: After the pulse envelope extraction is completed, the pulse upper and lower edges are detected by one-dimensional edge detection to determine the starting and ending positions of the pulse.

2. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: Step 1, generate a pulse segmentation index list: Use GPU to quickly find the pulse segmentation flag in the batch IF data, and generate a pulse segmentation index list in sequence. The specific method is as follows: Step 1-1: Copy the real-time received batch IF data from the CPU memory to the GPU memory, and the data length is if_lens; Step 1-2: Generate the initial pulse segmentation index list if_seg_idx according to the minimum pulse width limit pw_min, the formula is: if_seg_idx=zeros(if_lens / / pw_min) Where if_seg_idx is an array of all zeros with a length of if_lens / / pw_min; Step 1-3: Start a GPU thread for each IF data point in the GPU; Step 1-4: In each thread, check whether the point is a pulse split flag; Step 1-5: If yes, record its index in the pulse segmentation index list; if_seg_idx[thread_idx / / pw_min]=thread_idx Step 1-6: Keep only the non-zero values ​​in if_seg_idx.

3. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: Step 2, establish a hash mapping table from intermediate frequency data to pulse segmentation index list: establish an equal-length pulse segmentation point hash mapping table for the input intermediate frequency data, each position in the table represents the position of the pulse data at that point in the pulse segmentation index list, the specific method is as follows: Step 2-1: Create an array of the same length as the input IF data, called the hash mapping table if_seg_idx_hash_map from IF data to pulse segmentation index list; Step 2-2: Initialize if_seg_idx_hash_map to 0; Step 2-3: traverse the pulse segmentation index list if_seg_idx, and set the position data corresponding to the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list to 1; Step 2-4: Through GPU parallel computing, the hash mapping table if_seg_idx_hash_map from the intermediate frequency data to the pulse segmentation index list is accumulated and summed, and then the actual index of each pulse cutting point is calculated.

4. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: Step 3, pulse envelope extraction: According to the pulse separation information and pulse width, the pulse is dynamically divided into multiple sub-intervals, a GPU thread is started for each sub-interval, and its maximum value is calculated as the envelope value; the specific method is: Step 3-1: Determine the start and end positions of each pulse according to the pulse segmentation index list; Step 3-2: For each pulse, divide it into 512 sub-intervals; Step 3-3: Start a GPU thread for each sub-interval, parallelize the calculation of the maximum value of the intermediate frequency data in the interval, and use it as the envelope value of the interval.

5. The method for batch recognition and edge detection of intermediate frequency data pulses of radar electronic countermeasures based on GPU according to claim 1 is characterized in that: Step 4, pulse upper and lower edge detection: After the pulse envelope extraction is completed, the pulse upper and lower edges are detected by one-dimensional edge detection to determine the starting and ending positions of the pulse. The specific method is as follows: Step 4-1: For each segmented pulse envelope one-dimensional signal, establish a GPU sub-thread; Step 4-2: In each sub-thread, edge detection is performed on the pulse envelope one-dimensional signal; Step 4-3: Using a differential algorithm, when a transition from low to high or from high to low is detected, the position is recorded as the upper and lower edges of the pulse, thereby determining the start and end positions of the pulse; Step 4-4: Post-process the upper and lower edge positions.

6. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: Before the data is copied to the GPU memory, the batch IF data received in real time is preprocessed, including noise removal, filtering or signal enhancement.

7. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: Pulse envelope extraction uses Fourier transform or wavelet transform to extract the envelope of the signal and uses a sliding window function to smooth the signal.

8. The GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method according to claim 1 is characterized in that: One-dimensional edge detection uses first-order or second-order difference methods to detect signal changes, or uses convolution algorithms to find the rising and falling edges of the signal.

9. The method for batch recognition and edge detection of intermediate frequency data pulses of radar electronic countermeasures based on GPU according to claim 1 is characterized in that: Post-process the upper and lower edge positions. These include removing duplicate edges, using filters to remove noise or smooth edge positions, and using pattern recognition or machine learning algorithms to further optimize pulse detection results.

10. A radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection system based on GPU, characterized in that: Implement the GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection method described in any one of claims 1-9 to achieve GPU-based radar electronic countermeasure intermediate frequency data pulse batch recognition and edge detection.

Images