A radar sorting and identification method of multi-feature transformation

Abstract

The application discloses a radar sorting and identification method of multi-feature conversion, and relates to the field of radar signal processing; according to the multi-feature data of the radar, the method is grouped for multiple times, and the grouped discrete radar new data is converted into plane image data; after the distribution of the radar signal is acquired and determined, the Kmeans algorithm is used to perform secondary clustering and grouping on the data; then, the PRI of each signal is acquired through three data grouping of the data amount per unit time; then, the acquired PRI is used to perform conversion on the TOA data, so that each signal is gathered after conversion and separated based on this; finally, feature extraction and identification are performed on the separated data in the PRI and PW features. The application combines the carrier frequency, the pulse width, the arrival time, the amplitude and the direction angle of the radar, fully utilizes the available information in the radar signal, and rapidly maps the pulse stream to different spaces through the carrier frequency and the pulse width clustering, thereby improving the real-time performance of the radar signal sorting.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing, and more specifically to a radar sorting and identification method with multi-feature transformation. Background Technology

[0002] The statements in this section are provided only as background information in connection with this disclosure and may not constitute prior art.

[0003] As electromagnetic environments and radar modulation methods become increasingly complex, the performance of single-feature matching recognition based on traditional methods is gradually declining. In recent years, with the rapid development of image processing technology, converting radar signals into time-frequency images for sorting and recognition has become a trend.

[0004] Many existing studies follow this approach: first, they use methods such as short-time Fourier transform and Choi-Williams time-frequency distribution to perform time-frequency analysis on the signal, obtaining a time-frequency image. Then, based on information theory and domain expertise, they manually design feature vectors for the radar signal from the time-frequency image. Finally, they use machine learning methods to complete signal classification. However, this method struggles to separate signals effectively when faced with aliased signals. Therefore, the resulting time-frequency image often fails to accurately represent the true characteristics of the radar signal, making identification difficult. Summary of the Invention

[0005] The purpose of this invention is to address the problem of sorting and identifying radar signals with multiple overlapping features by providing a radar sorting and identification method based on multi-feature transformation. This method utilizes K-means data clustering and grouping based on image transformation and TOA transformation. It performs multiple groupings based on radar multi-feature data, converts the discrete new radar data into planar image data, and after determining the distribution of radar signals, uses the K-means algorithm for secondary clustering. Then, it obtains the PRI (Primary Indicator) of each signal through three data groupings per unit time. Next, it uses the obtained PRI to transform the TOA data, causing each signal to cluster after the transformation and thus achieve separation. Finally, it extracts and identifies features from the separated data based on PRI and PW (Primary Indicator and Pw) features, thereby solving the aforementioned problem.

[0006] The technical solution of the present invention is as follows: A radar sorting and identification method with multi-feature transformation, comprising: Step S1: Obtain the maximum and minimum values ​​of the radar signal pulse width, divide the pulse width region into n intervals, statistically analyze the radar pulse width data, segment the data based on the pulse width statistical data troughs, and separate the radar data according to the segmentation results to obtain the first-stage separated data. ; Step S2: Perform data volume statistics on the primary separated data at fixed time intervals to obtain the data volume per unit time. Then, perform secondary separation on the primary separated data based on the time of data abrupt change to obtain secondary separated data. ; Step S3: For each secondary separated data block, calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction of arrival (DOA), and divide the two characteristic intervals into k... The data is divided into l small intervals, and the data in each small interval is statistically analyzed. Finally, the resulting data matrix is ​​binarized, and the binarized image is obtained by filtering and erosion. Step S4: Find each independent connected region in the binarized image, and solve the corresponding secondary separation data filtering range by calculating the coordinate range. Take a portion of the data in the data center as the initial value of Kmeans to cluster the secondary separation data. Step S5: Calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction angle (DOA) of the clustered data, and use these values ​​as a range to select radar data from the first separation data as the third separation data. ; Step S6: If all independent connected regions in the current binarized image have undergone steps S4 and S5, then proceed to step S7; otherwise, return to step S4. Step S7: Calculate the TOA corresponding to the extreme point of amplitude for each three-dimensional separation data, and perform DTW matching on the TOA sequence of extreme points of each separation data. Merge the three-dimensional separation data that are less than the threshold. Step S8: Calculate the time derivative of TOA for each cluster of data and filter the data by radar parameter thresholds; obtain the period and the first and last TOA of the period by the time derivative; Step S9: Transform the three separated data TOA in step S7. First, subtract the first or last TOA of the period according to the data group situation. Then, take the remainder of TOA with respect to the period to obtain the TOA data with period variation. Step S10: Statistically analyze the periodically changing TOA data in the maximum and minimum intervals, and use k1 times the interval mean as a threshold to filter out intervals greater than k1 times the mean. Solve for the original data index based on the interval range. Calculate the derivative of the three separated TOA data after merging and recombination, and extract features from the time derivative and carrier frequency image to complete signal identification and labeling. Step S11: If all independent connected regions in the current binarized image have undergone steps S8 to S10, then proceed to step S12; otherwise, return to step S8. Step S12: If all the current single-separation data intervals have been processed, proceed to step S13; otherwise, return to step S2. Step S13: Output the separation results.

[0007] Further, step S1 includes: Step A: Obtain the maximum value of the radar signal pulse width. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step B: Divide the acquired radar data pulse width signal by the minimum interval threshold. The transformed pulse width data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step C: Set a suspected trough threshold based on the total amount of signal data, and analyze the statistical data. Divide the interval; Step D: Divide the original data according to the divided interval range to obtain the first separated data. .

[0008] Further, step S2 includes: Step S21: Obtain the maximum value of radar data arrival time. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S22: Divide the arrival time of the acquired radar data by the minimum interval threshold. The transformed time data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step S23: Obtain statistical data The differential components are determined, and a threshold is set based on the total amount of data to distinguish the moment of data mutation and the data is divided according to the mutation interval; Step S24: Divide the original data according to the divided interval range to obtain secondary separated data. .

[0009] Further, step S3 includes: Step S31: Obtain the maximum value of the radar signal carrier frequency. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S32: Obtain the maximum value of the radar signal arrival direction angle. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S33: Divide the acquired radar carrier frequency data by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. Divide the acquired radar data's direction of arrival (DOA) by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. ; Step S34: Obtain statistical data , Find the maximum values ​​k and l, and generate a k l-matrix; based on the statistical data , Combine them into two-dimensional coordinates, and perform statistics based on the generated two-dimensional coordinates and matrix subscripts to obtain unprocessed statistical matrix data; Step S35: Set a binarization threshold, binarize the matrix data, and convert it into a binarized image.

[0010] Further, step S4 includes: Step S41: Perform erosion and dilation operations on the binarized image, and then obtain each independent connected region; Step S42: By obtaining the first Coordinates of an independent connected region and , , , Data transformation is used to solve for the data range corresponding to independent connected regions.

[0011] Further, step S7 includes: Step S71: Find the extreme points of the amplitude for the three separate data sets and save them as a sequence; Step S72: Calculate the distance between all pairs of points for different sequences to form a matrix; then, starting from the top left corner of the matrix and ending at the bottom right corner, use dynamic programming to find a path that minimizes the sum of the distances between pairs of points on the path; finally, the sum of the distances between pairs of points on the path is the DTW distance between the two sequences, which reflects their similarity. Step S73: Set the DTW similarity threshold. If the similarity is less than a certain threshold, it is set as the same signal.

[0012] Furthermore, the interval Calculated using the following formula: .

[0013] Furthermore, the interval Calculated using the following formula: .

[0014] Furthermore, the interval Calculated using the following formula: .

[0015] Furthermore, the interval Calculated using the following formula: .

[0016] Compared with existing technologies, the advantages of this invention are: 1. A radar sorting and identification method based on multi-feature transformation combines multiple parameters of radar, such as carrier frequency, pulse width, arrival time, amplitude, and direction of arrival, to make full use of available information in radar signals. By clustering carrier frequency and pulse width, the pulse stream is quickly mapped to different spaces, thereby improving the real-time performance of radar signal sorting.

[0017] 2. A radar sorting and identification method based on multi-feature transformation utilizes the periodic characteristics of the LFM signal itself, avoiding the shortcomings of traditional histograms in sorting complex modulated signals, such as poor sorting effect and time consumption, thereby improving accuracy and sorting speed.

[0018] 3. A radar sorting and identification method based on multi-feature transformation improves the accuracy and robustness of classification by grouping radar signals into different layers according to different parameters and mapping them to different spaces for sorting. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall process of a radar sorting and identification method with multi-feature transformation provided by the present invention; Figure 2 This is a schematic diagram of the radar signal to binary image conversion provided by the present invention; Figure 3 This is a diagram showing the result of converting radar signals into a binarized image, as provided by the present invention. Figure 4 This is a result diagram of the radar signal time-frequency diagram provided by the present invention. Detailed Implementation

[0020] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0021] The features and performance of the present invention will be further described in detail below with reference to embodiments.

[0022] Example 1 Please see Figure 1 A radar sorting and identification method with multi-feature transformation, specifically including the following steps: Step S1: Obtain the maximum and minimum values ​​of the radar signal pulse width, divide the pulse width region into n intervals, statistically analyze the radar pulse width data, segment the data based on the pulse width statistical data troughs, and separate the radar data according to the segmentation results to obtain the first-stage separated data. ; In this embodiment, specifically, step S1 includes: Step A: Obtain the maximum value of the radar signal pulse width. and minimum value And set a minimum interval threshold. Calculate the interval and round up; interval Calculated using the following formula:

[0023] Step B: Divide the acquired radar data pulse width signal by the minimum interval threshold. The transformed pulse width data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step C: Set a suspected trough threshold based on the total amount of signal data, and analyze the statistical data. Divide the interval; Step D: Divide the original data according to the divided interval range to obtain the first separated data. .

[0024] Step S2: Perform data volume statistics on the primary separated data at fixed time intervals to obtain the data volume per unit time. Then, perform secondary separation on the primary separated data based on the time of data abrupt change to obtain secondary separated data. ; In this embodiment, specifically, step S2 includes: Step S21: Obtain the maximum value of radar data arrival time. and minimum value And set a minimum interval threshold. Calculate the interval and round up; interval Calculated using the following formula:

[0025] Step S22: Divide the arrival time of the acquired radar data by the minimum interval threshold. The transformed time data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step S23: Obtain statistical data The differential components are determined, and a threshold is set based on the total amount of data to distinguish the moment of data mutation and the data is divided according to the mutation interval; Step S24: Divide the original data according to the divided interval range to obtain secondary separated data. .

[0026] Step S3: For each secondary separated data block, calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction of arrival (DOA), and divide the two characteristic intervals into k... The data is divided into l small intervals, and the data in each small interval is statistically analyzed. Finally, the resulting data matrix is ​​binarized, and the binarized image is obtained by filtering and erosion. In this embodiment, specifically, step S3 includes: Step S31: Obtain the maximum value of the radar signal carrier frequency. and minimum value And set a minimum interval threshold. Calculate the interval and round up; interval Calculated using the following formula:

[0027] Step S32: Obtain the maximum value of the radar signal arrival direction angle. and minimum value And set a minimum interval threshold. Calculate the interval and round up; interval Calculated using the following formula:

[0028] Step S33: Divide the acquired radar carrier frequency data by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. Divide the acquired radar data's direction of arrival (DOA) by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. ; Step S34: Obtain statistical data , Find the maximum values ​​k and l, and generate a k l-matrix; based on the statistical data , Combine them into two-dimensional coordinates, and perform statistics based on the generated two-dimensional coordinates and matrix subscripts to obtain unprocessed statistical matrix data; Step S35: Set a binarization threshold, binarize the matrix data, and convert it into a binarized image.

[0029] Step S4: Find each independent connected region in the binarized image, and solve the corresponding secondary separation data filtering range by calculating the coordinate range. Take a portion of the data in the data center as the initial value of Kmeans to cluster the secondary separation data. In this embodiment, specifically, step S4 includes: Step S41: Perform erosion and dilation operations on the binarized image, and then obtain each independent connected region; Step S42: By obtaining the first Coordinates of an independent connected region and , , , Data transformation is used to solve for the data range corresponding to independent connected regions.

[0030]

[0031] Step S5: Calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction angle (DOA) of the clustered data, and use these values ​​as a range to select radar data from the first separation data as the third separation data. ; Step S6: If all independent connected regions in the current binarized image have undergone steps S4 and S5, then proceed to step S7; otherwise, return to step S4. Step S7: Calculate the TOA corresponding to the extreme point of amplitude for each three-dimensional separation data, and perform DTW matching on the TOA sequence of extreme points of each separation data. Merge the three-dimensional separation data that are less than the threshold. In this embodiment, specifically, step S7 includes: Step S71: Find the extreme points of the amplitude for the three separate data sets and save them as a sequence; Step S72: Calculate the distance between all pairs of points for different sequences to form a matrix; then, starting from the top left corner of the matrix and ending at the bottom right corner, use dynamic programming to find a path that minimizes the sum of the distances between pairs of points on the path; finally, the sum of the distances between pairs of points on the path is the DTW distance between the two sequences, which reflects their similarity. Step S73: Set the DTW similarity threshold. If the similarity is less than a certain threshold, it is set as the same signal.

[0032] Step S8: Calculate the time derivative of TOA for each cluster of data and filter the data by radar parameter thresholds; obtain the period and the first and last TOA of the period by the time derivative; Step S9: Transform the TOA of the three separate data sets. First, subtract the TOA at the beginning or end of the period according to the data set. Then, take the remainder of the TOA with respect to the period to obtain the TOA data with periodic variation. Step S10: Statistically analyze the periodically changing TOA data in the maximum and minimum intervals, and use k1 times the interval mean as a threshold to filter out intervals greater than k1 times the mean. Solve for the original data index based on the interval range. Calculate the derivative of the three separated TOA data after merging and recombination, and extract features from the time derivative and carrier frequency image to complete signal identification and labeling. Step S11: If all independent connected regions in the current binarized image have undergone steps S8 to S10, then proceed to step S12; otherwise, return to step S8. Step S12: If all the current single-separation data intervals have been processed, proceed to step S13; otherwise, return to step S2. Step S13: Output the separation results.

[0033] In this embodiment, it should be noted that the comprehensive search method for obtaining PRI by detecting the inherent period of PRI or the periodic pattern in the transform domain is flexible and applicable to signal sorting with various PRI patterns. However, in signal-dense environments, various PRI estimation methods are affected to varying degrees in practical applications. When various pulses are interleaved and mixed, the true PRI value is hidden in the higher-order TOA difference, and the computational complexity of the statistical algorithm increases geometrically with the order of the TOA difference. Therefore, this embodiment can also rely on the inherent period of PRI to transform the TOA model, generating a periodic pattern of the plane transformation point trace of the PRI fixed sequence with only one time-domain transformation processing; then, based on the difference in the point trace pattern pattern, the determination of multiple TOA interleaved PRI fixed pulse sequences is realized, thereby better achieving radar signal sorting and identification.

[0034] Example 2 See Figure 1-4 This embodiment provides a method for deinterleaving and sorting radar signals with multiple overlapping features. First, the data is separated and extracted multiple times using multi-dimensional features. Then, the same signal data are merged based on multiple features. Finally, the recognition is completed using the time-frequency image of the signal.

[0035] The flowchart of this method is as follows: Figure 1 As shown, the separation and extraction of multi-feature interleaved signals are completed through multiple data separation steps, and then identification is performed by plotting a time-frequency graph. The specific application steps are as follows: Step S1: Obtain the maximum and minimum values ​​of the radar signal pulse width, set the minimum pulse width interval to 1, divide the pulse width region into n intervals, perform statistical analysis on the radar pulse width data, and segment the data based on the valleys in the pulse width statistical data. Then, separate the radar data based on the segmentation results to obtain the first-stage separated data.

[0036] Step S2: Set a fixed time interval of 0.025s, perform data volume statistics on the first-separated data within the fixed time interval, obtain the data volume in each unit time, and perform a second separation on the first-separated data according to the data change time to obtain the second-separated data.

[0037] Step S3: Set the minimum interval for carrier frequency (RF) and minimum interval for direction of arrival (DOA) to 0.3. Set the erosion kernel size to (2, 2) and the expansion kernel size to (2, 2). For each secondary separated data block, calculate the maximum and minimum values ​​of carrier frequency and direction of arrival (DOA), and divide the two characteristic intervals into k... The data is divided into l small intervals, and the data in each small interval is statistically analyzed. Finally, the resulting data matrix is ​​binarized, and the binarized image is obtained by filtering and erosion.

[0038] Step S4: Find each independent connected region in the binarized image, and solve the corresponding secondary separation data filtering range by calculating the coordinate range. Take a portion of the data from the data center as the initial value of Kmeans to cluster the secondary separation data.

[0039] Step S5: Calculate the maximum and minimum values ​​of the carrier frequency and direction angle of the clustered data, and use these values ​​as the range to select radar data from the first separation data as the third separation data.

[0040] Step S6: If all independent connected regions in the current binarized image have undergone steps S4 and S5, proceed to step S7; otherwise, return to step S4.

[0041] Step S7: Calculate the TOA corresponding to the extreme point of amplitude for each tripartite separation data. Then, perform DTW matching on the TOA sequence of extreme points for each separation data, and merge tripartite separation data that are less than the threshold.

[0042] Step S8: Calculate the time derivative of TOA for each cluster of data and filter the data by radar parameter thresholds; obtain the period and the first and last TOA of the period by the time derivative; Step S9: Transform the TOA of the three separate data sets. First, subtract the TOA at the beginning or end of the period based on the data set. Then, take the remainder of the TOA with respect to the period. This will give you the TOA data with periodic variations.

[0043] Step S10: Statistically analyze the periodically varying TOA data across the maximum and minimum intervals, and use k1 times the interval mean as a threshold to filter out intervals greater than k1 times the mean. Calculate the original data indices based on the interval ranges. Differentiate the TOA data from the three separate data sets after merging and recombination, and extract features from the time derivative and carrier frequency images to complete signal identification and labeling.

[0044] Step S11: If all independent connected regions in the current binarized image have undergone steps S8 to S10, proceed to step S12; otherwise, return to step S8.

[0045] Step S12: If all the current single-separation data intervals have been processed, proceed to step S13; otherwise, return to step S2.

[0046] Step S13: Output the separation results.

[0047] The sorting and identification using the test set data yields the results shown in Table 1.

[0048] Table 1 Recognition Results

[0049] The embodiments described above merely illustrate specific implementation methods of this application, and while the descriptions are detailed and specific, they should not be construed as limiting the scope of protection of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the technical solution of this application, and these modifications and improvements all fall within the scope of protection of this application.

[0050] This background section is provided to generally present the context of the invention. The work of the currently named inventors, the work to the extent described in this background section, and aspects of this section that did not constitute prior art at the time of application are neither expressly nor impliedly acknowledged as prior art to the invention.

Claims

1. A radar sorting and identification method of multi-feature transformation, characterized in that, include: Step S1: Obtain the maximum and minimum values ​​of the radar signal pulse width, divide the pulse width region into n intervals, statistically analyze the radar pulse width data, segment the data based on the pulse width statistical data troughs, and separate the radar data according to the segmentation results to obtain the first-stage separated data. ; Step S2: Perform data volume statistics on the primary separated data at fixed time intervals to obtain the data volume per unit time. Then, perform secondary separation on the primary separated data based on the time of data abrupt change to obtain secondary separated data. ; Step S3: For each secondary separated data block, calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction of arrival (DOA), and divide the two characteristic intervals into k... The data is divided into l small intervals, and the data in each small interval is statistically analyzed. Finally, the resulting data matrix is ​​binarized, and the binarized image is obtained by filtering and erosion. Step S4: Find each independent connected region in the binarized image, and solve the corresponding secondary separation data filtering range by calculating the coordinate range. Take a portion of the data in the data center as the initial value of Kmeans to cluster the secondary separation data. Step S5: Calculate the maximum and minimum values ​​of the carrier frequency (RF) and direction angle (DOA) of the clustered data, and use these values ​​as the range to select radar data from the first separation data as the third separation data. ; Step S6: If all independent connected regions in the current binarized image have undergone steps S4 and S5, then proceed to step S7; otherwise, return to step S4. Step S7: Calculate the TOA corresponding to the extreme point of amplitude for each three-dimensional separation data, and perform DTW matching on the TOA sequence of extreme points of each separation data. Merge the three-dimensional separation data that are less than the threshold. Step S8: Calculate the time derivative of TOA for each cluster of data and filter the data by radar parameter thresholds; obtain the period and the first and last TOA of the period by the time derivative; Step S9: Transform the three separated data TOA in step S7. First, subtract the first or last TOA of the period according to the data group situation. Then, take the remainder of TOA with respect to the period to obtain the TOA data with periodic variation. Step S10: Statistically analyze the periodically changing TOA data in the maximum and minimum intervals, and use k1 times the interval mean as a threshold to filter out intervals greater than k1 times the mean. Solve for the original data index based on the interval range. Calculate the derivative of the three separated TOA data after merging and recombination, and extract features from the time derivative and carrier frequency image to complete signal identification and labeling. Step S11: If all independent connected regions in the current binarized image have undergone steps S8 to S10, then proceed to step S12; otherwise, return to step S8. Step S12: If all the current single-separation data intervals have been processed, proceed to step S13; otherwise, return to step S2. Step S13: Output the separation results.

2. The radar sorting and identification method based on multi-feature transformation according to claim 1, characterized in that, Step S1 includes: Step A: Obtain the maximum value of the radar signal pulse width. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step B: Divide the acquired radar data pulse width signal by the minimum interval threshold. The transformed pulse width data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step C: Set a suspected trough threshold based on the total amount of signal data, and analyze the statistical data. Divide the interval; Step D: Divide the original data according to the divided interval range to obtain the first separated data. .

3. The radar sorting and identification method based on multi-feature transformation according to claim 2, characterized in that, Step S2 includes: Step S21: Obtain the maximum value of radar data arrival time. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S22: Divide the arrival time of the acquired radar data by the minimum interval threshold. The transformed time data is then rounded up and statistically analyzed for each interval to obtain statistical data. ; Step S23: Obtain statistical data The differential components are determined, and a threshold is set based on the total amount of data to distinguish the moment of data mutation and the data is divided according to the mutation interval; Step S24: Divide the original data according to the divided interval range to obtain secondary separated data. .

4. The radar sorting and identification method based on multi-feature transformation according to claim 3, characterized in that, Step S3 includes: Step S31: Obtain the maximum value of the radar signal carrier frequency. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S32: Obtain the maximum value of the radar signal arrival direction angle. and minimum value And set a minimum interval threshold. Calculate the interval and round up; Step S33: Divide the acquired radar carrier frequency data by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. Divide the acquired radar data's direction of arrival (DOA) by the minimum interval threshold. The transformed data is then rounded up, and statistical analysis is performed on the data in each interval to obtain statistical data. ; Step S34: Obtain statistical data , Find the maximum values ​​k and l, and generate a k l-matrix; based on the statistical data , Combine them into two-dimensional coordinates, and perform statistics based on the generated two-dimensional coordinates and matrix subscripts to obtain unprocessed statistical matrix data; Step S35: Set a binarization threshold, binarize the matrix data, and convert it into a binarized image.

5. The radar sorting and identification method based on multi-feature transformation according to claim 4, characterized in that, Step S4 includes: Step S41: Perform erosion and dilation operations on the binarized image, and then obtain each independent connected region; Step S42: By obtaining the first Coordinates of an independent connected region and , , , Data transformation is used to solve for the data range corresponding to independent connected regions.

6. The radar sorting and identification method based on multi-feature transformation according to claim 5, characterized in that, Step S7 includes: Step S71: Find the extreme points of the amplitude for the three separate data sets and save them as a sequence; Step S72: Calculate the distance between all pairs of points for different sequences to form a matrix; then, starting from the top left corner of the matrix and ending at the bottom right corner, use dynamic programming to find a path that minimizes the sum of the distances between pairs of points on the path; finally, the sum of the distances between pairs of points on the path is the DTW distance between the two sequences, which reflects their similarity. Step S73: Set the DTW similarity threshold. If the similarity is less than a certain threshold, it is set as the same signal.

7. The radar sorting and identification method based on multi-feature transformation according to claim 2, characterized in that, The interval Calculated using the following formula: 。 8. The radar sorting and identification method based on multi-feature transformation according to claim 3, characterized in that, The interval Calculated using the following formula: 。 9. The radar sorting and identification method based on multi-feature transformation according to claim 4, characterized in that, The interval Calculated using the following formula: 。 10. The radar sorting and identification method based on multi-feature transformation according to claim 4, characterized in that, The interval Calculated using the following formula: 。

Images