DBCA-CFAR detection method and system based on feature clustering

Abstract

The invention discloses a DBCA-CFAR detection method and system based on feature clustering, mainly relates to the technical field of target detection, and is used for solving the problem that the detection performance is reduced due to the fact that an existing scheme is easily interfered in a multi-target or non-uniform clutter environment. Comprising the following steps: extracting a TEM feature and an amplitude feature corresponding to each preset distance unit so as to construct a two-dimensional feature vector; inputting the two-dimensional feature vector into a DBSCAN clustering algorithm to obtain a target interference point and a sea clutter point; setting the echo amplitude of the radar echo data of the preset distance unit corresponding to the target interference point as 0; based on the sliding window, determining a corresponding to-be-detected unit and a reference unit from the preset distance units; counting the amplitude mean value of the non-zero echo amplitude of the reference unit corresponding to the to-be-detected unit, and multiplying the amplitude mean value by a preset nominal factor to obtain a threshold value; and determining whether the to-be-detected unit has a target according to the threshold value and the echo amplitude of the to-be-detected unit.

Description

Technical Field

[0001] This application relates to the technical field of target detection, and in particular, to a DBCA-CFAR detection method and system based on feature clustering. Background Art

[0002] The constant false alarm rate (CFAR) detection technology is the core means of radar target detection, and its core goal is to maintain a stable false alarm probability in complex environments. Traditional mean-based CFAR detectors (such as CA-CFAR, Cell-Averaging Constant False Alarm Rate) estimate the background noise by sliding a window to statistically calculate the average power of the reference cells, and show good detection performance in homogeneous clutter environments.

[0003] However, when there are multiple strong targets in the reference window, traditional CA-CFAR will include the energy of interfering targets in the background estimation, resulting in an abnormally elevated detection threshold and triggering the "target masking effect", causing missed detection of real weak targets. Although ordered statistic-based CFAR (OS-CFAR, Ordered Statistic Constant False Alarm Rate) can partially alleviate this problem by sorting and removing outliers, it still has performance degradation in high-density multi-target scenarios. In addition, non-uniform environments such as sea clutter and meteorological clutter have time-variability and spatial correlation, and traditional methods relying on local statistical characteristics will lead to background estimation bias. For example, in the clutter edge region, CA-CFAR may generate false alarms or missed detections because the reference window spans high and low clutter regions. Existing improved algorithms (such as GO / SO-CFAR) cope with the clutter transition region by selecting the maximum or minimum reference sub-window, but at the cost of detection probability. Summary of the Invention

[0004] This application provides a DBCA-CFAR detection method and system based on feature clustering to solve the problem that existing solutions are vulnerable to interference in multi-target or non-uniform clutter environments, resulting in a decline in detection performance.

[0005] In a first aspect, this application provides a DBCA-CFAR detection method based on feature clustering. The method includes: Extract the TEM feature and amplitude feature corresponding to each preset range cell from the radar echo data, and then construct a two-dimensional feature vector; obtain the neighborhood radius and minimum number of points involved in the DBSCAN clustering algorithm, and then input the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points; set the echo amplitude of the radar echo data of the preset range cell corresponding to the target interference point to 0; based on a sliding window, determine the corresponding cells to be detected and reference cells from the preset range cells; count the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cells to be detected, multiply the amplitude mean by a preset normalization factor to obtain a threshold; determine whether there is a target in the cells to be detected according to the threshold and the echo amplitude of the cells to be detected.

[0006] In an implementation manner of the present application, extracting the TEM feature corresponding to each preset range cell specifically includes: Through the formula: , obtain the TEM feature; Wherein, N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, N + L - 1 represents the number of signals with a length of L, represents the time-domain entropy value of the signal, And , represents the i th interval normalized probability density of the signal amplitude, represents the i th interval sample value in the signal sequence of the radar echo data.

[0007] In an implementation manner of the present application, based on a sliding window, determining the corresponding cells to be detected and reference cells from the preset range cells specifically includes: The window slides in the order of the preset range cells. With the cell to be detected as the center, 1 protection cell and a preset number of reference cells are set on both sides respectively, and 2N represents the total number of preset reference cells; When the number of protection cells or reference cells corresponding to one side of the cell to be detected is insufficient, it is supplemented by zero padding or mirror extension.

[0008] In an implementation manner of the present application, counting the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cells to be detected specifically includes: Obtain the number of zero echo amplitudes P and Q corresponding to the reference cells on the left and right sides of the cell to be detected respectively; Through -P, obtain the number of non-zero echo amplitudes on the left side, and then sum and average to obtain X; wherein, represents the number of left reference cells; Through -Q, obtain the number of non-zero echo amplitudes on the right side, and then sum and average to obtain Y; where represents the number of right-side reference units; Through X + Y, take the average value to obtain the amplitude mean of the non-zero echo amplitudes of the reference unit corresponding to the unit to be detected.

[0009] In an implementation manner of the present application, by using the threshold value and the echo amplitude of the unit to be detected, it is determined whether there is a target in the unit to be detected, specifically including: When the echo amplitude of the unit to be detected is greater than the threshold value, it is determined that a target exists, otherwise it is clutter.

[0010] In an implementation manner of the present application, obtain the neighborhood radius and the minimum number of points involved in the DBSCAN clustering algorithm, specifically including: Set the neighborhood radius to 150 and the minimum number of points to 3.

[0011] In a second aspect, the present application provides a DBCA-CFAR detection system based on feature clustering. The system includes: A two-dimensional construction module, used to extract the TEM feature and amplitude feature corresponding to each preset distance unit from the radar echo data, and then construct a two-dimensional feature vector; an interference acquisition module, used to obtain the neighborhood radius and the minimum number of points involved in the DBSCAN clustering algorithm, and then input the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points; a target detection module, used to set the echo amplitude of the radar echo data of the preset distance unit corresponding to the target interference point to 0; based on a sliding window, determine the corresponding unit to be detected and reference unit from the preset distance units; count the amplitude mean of the non-zero echo amplitudes of the reference unit corresponding to the unit to be detected, multiply the amplitude mean by a preset normalization factor to obtain a threshold value; determine whether there is a target in the unit to be detected according to the threshold value and the echo amplitude of the unit to be detected.

[0012] In an implementation manner of the present application, the two-dimensional construction module includes a TEM feature calculation unit used to obtain the TEM feature through the formula: , obtain the TEM feature; where N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, N + L - 1 represents the number of signals with a length of L, represents the time-domain entropy value of the signal, and , represents the i th interval of the signal amplitude, and the normalized probability density, The i th interval sample value in the signal sequence representing radar echo data.

[0013] In one implementation of the present application, the target detection module includes a window determination unit, which is used to slide the window in the order of preset distance units. With the unit to be detected as the center, one protection unit and a preset number of reference units are respectively set on both sides, and 2N represents the total number of preset reference units; When the number of protection units or reference units corresponding to one side of the unit to be detected is insufficient, it is supplemented by zero-padding or mirror extension.

[0014] In one implementation of the present application, the target detection module includes a target determination unit, which is used to determine that a target exists if the echo amplitude of the unit to be detected is greater than the threshold value, otherwise it is clutter.

[0015] From the above technical solutions, it can be seen that the present application has the following advantages: Through the joint analysis of TEM features and amplitude features by the DBSCAN clustering algorithm, the present application can accurately identify and eliminate strong interference target points (set to amplitude 0) in the reference window, fundamentally avoiding the "target masking effect" caused by the inclusion of interference target energy in the background estimation of traditional CA-CFAR, and significantly improving the detection probability of weak targets in high-density multi-target scenarios.

[0016] Multi-dimensional feature clustering can effectively distinguish sea clutter from real targets. The density-based clustering characteristic of DBSCAN can adaptively process the clutter edge region. Compared with the fixed sub-window selection strategy of GO / SO-CFAR, it avoids sacrificing the detection probability while suppressing false alarms.

[0017] Only the amplitude mean value of the "non-interference points" after clustering is statistically calculated, reducing the amount of invalid calculations. Compared with the sorting operation (O(nlogn) complexity) of traditional OS-CFAR and the feature training cost of machine learning CFAR, this method has a lower real-time burden while ensuring accuracy.

[0018] In addition, the present application first jointly models the TEM feature (time-frequency energy modulation characteristic) and the amplitude feature, breaking through the limitation of traditional CFAR that only relies on amplitude information, and providing a new dimension of discriminant basis for target detection in complex electromagnetic environments.

[0019] In summary, since the traditional CA-CFAR method estimates the background power relying on the statistical characteristics of the local reference window, it is vulnerable to interference in multi-target or non-uniform clutter environments, resulting in a decline in detection performance. The DBCA-CFAR detector based on feature clustering in this application improves the robustness in complex scenarios by introducing data clustering analysis. First, this application extracts features from radar echo data, including amplitude and TEM; subsequently, the data is divided into target classes and background clutter classes through the DBSCAN clustering algorithm; finally, pure background samples are selected based on the clustering results to estimate the noise power, thereby optimizing the detection threshold. It solves the problem of increased false alarm rate caused by interference in the reference window in multi-target scenarios in the traditional method, and at the same time reduces the influence of clutter non-uniformity. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] To more clearly illustrate the technical solutions of the present invention, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 It is a flowchart of a DBCA-CFAR detection method based on feature clustering provided by an embodiment of this application.

[0022] Figure 2 It is a comparison diagram of the numerical distribution of the mean time-domain entropy values of targets and clutter provided by an embodiment of this application.

[0023] Figure 3 It is a comparison diagram of the histogram distribution of the mean time-domain entropy values of targets and clutter provided by an embodiment of this application.

[0024] Figure 4 It is a schematic diagram of the clustering result of #DATA1 provided by an embodiment of this application.

[0025] Figure 5 It is a data diagram of #DATA1 before screening provided by an embodiment of this application.

[0026] Figure 6 It is a data diagram of #DATA1 after screening provided by an embodiment of this application.

[0027] Figure 7 It is a test result diagram of the data of #DATA1 before screening provided by an embodiment of this application.

[0028] Figure 8 It is a test result diagram of the data of #DATA1 after screening provided by an embodiment of this application.

[0029] Figure 9 It is a schematic diagram of the clustering result of #DATA2 provided by an embodiment of this application.

[0030] Figure 10 It is the data graph before #DATA2 screening provided by the embodiment of the present application.

[0031] Figure 11 It is the data graph after #DATA2 screening provided by the embodiment of the present application.

[0032] Figure 12 It is the test result graph of the data after #DATA2 screening provided by the embodiment of the present application.

[0033] Figure 13 It is a schematic diagram of the internal structure of a DBCA-CFAR detection system based on feature clustering provided by the embodiment of the present application. Detailed implementation manners

[0034] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0035] Those skilled in the art should understand that the embodiments described below are only the preferred embodiments of the present disclosure, which does not mean that the present disclosure can only be implemented through these preferred embodiments. These preferred embodiments are only used to explain the technical principles of the present disclosure and are not used to limit the protection scope of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the preferred embodiments provided by the present disclosure without creative efforts shall still fall within the protection scope of the present disclosure.

[0036] It should also be noted that the term "including", "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, so that a process, method, commodity or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or elements inherent to such process, method, commodity or device. Without further limitation, the element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, commodity or device including the element.

[0037] Next, the technical solutions proposed in the embodiments of the present application will be described in detail through the accompanying drawings.

[0038] The embodiment provides a DBCA-CFAR detection method based on feature clustering. As Figure 1 shown, the method provided by the embodiment of the present application mainly includes the following steps: Step 110: Extract the TEM feature and amplitude feature corresponding to each preset range cell from the radar echo data, and then construct a two-dimensional feature vector.

[0039] In some embodiments, extracting the TEM feature corresponding to each preset range cell specifically includes: Through the formula: , obtain the TEM feature; where N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, N + L - 1 represents the number of signals with a length of L, represents the time-domain entropy value of the signal, and , represents the i th interval normalized probability density of the signal amplitude, represents the i th interval sample value in the signal sequence of the radar echo data.

[0040] It should be noted that the TEM feature reflects the change characteristics of the signal in time, and the amplitude feature reflects the intensity of the signal. The combination of the two can more comprehensively describe the characteristics of the target and sea clutter.

[0041] Step 120: Obtain the neighborhood radius and minimum number of points involved in the DBSCAN clustering algorithm, and then input the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points; set the echo amplitude of the radar echo data of the preset range cell corresponding to the target interference point to 0.

[0042] It should be noted that in the mean-based CFAR detection, the interference target and sea spikes far exceed the power level of sea clutter. If they are not removed from the reference cells, the reference window will be severely affected when averaging the power level of the background clutter. Therefore, accurately distinguishing interference from sea clutter is a key issue. As Figure 2 shown, it can be found that the TEM feature values of sea clutter mainly fluctuate between 0.01 and 0.15, the overall TEM feature sequence of the target fluctuates around 0.05, and the TEM feature values of sea clutter cells are all greater than those of the target. As Figure 3 shown, the distinguishability between the mean values of the time-domain entropy of the target and clutter is demonstrated through the histogram distribution. It can be seen that there is a certain distinguishability between the TEM features of the target signal and the TEM features of the sea clutter signal.

[0043] Two key parameters need to be set when using the DBSCAN clustering algorithm: the neighborhood radius (Eps) and the minimum number of points (MinPts). Through multiple tests on the measured data, it can be determined that when the neighborhood radius (Eps) is set to 150 and the minimum number of points (MinPts) is 3, the clustering results can accurately distinguish suspected targets, interference, and clutter.

[0044] Input the two-dimensional feature vectors constructed in step 110 into the DBSCAN algorithm for clustering. According to the density relationship between data points, the data points are divided into different clusters, and at the same time, the noise points are identified. During the clustering process, data points with high density and adjacent to each other will be divided into the same cluster, while isolated data points will be marked as noise points. Through clustering, suspected target interference points and sea clutter points are initially distinguished.

[0045] In the clustering results, the amplitude of the points marked as outliers is set to 0. These outliers may be caused by interference, noise, or other factors. Setting their amplitudes to zero can effectively reduce the impact on the constant false alarm detection threshold. Step 130: Based on a sliding window, determine the corresponding cells to be detected and reference cells from the preset range cells.

[0046] This step can be specifically: The window slides in the order of the preset range cells. With the cell to be detected as the center, one guard cell and a preset number of reference cells are set on both sides. 2N represents the total number of preset reference cells; When the number of guard cells or reference cells corresponding to one side of the cell to be detected is insufficient, it is supplemented by zero-padding or mirror extension.

[0047] Step 140: Calculate the mean amplitude of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected, multiply the mean amplitude by a preset normalization factor to obtain a threshold; determine whether there is a target in the cell to be detected according to the threshold and the echo amplitude of the cell to be detected.

[0048] In some embodiments, calculating the mean amplitude of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected specifically includes: Obtain the numbers P and Q of the zero echo amplitudes corresponding to the reference cells on the left and right sides of the cell to be detected respectively; Through -P, obtain the number of non-zero echo amplitudes on the left side, and then sum and average to obtain X; where represents the number of left reference cells; Through -Q, obtain the number of non-zero echo amplitudes on the right side, and then sum and average to obtain Y; where represents the number of right reference cells; The amplitude mean of the non-zero echo amplitudes of the reference unit corresponding to the unit to be detected is obtained by taking the average of X+Y.

[0049] In some embodiments, it is determined whether there is a target in the unit to be detected by the threshold value and the echo amplitude of the unit to be detected, which specifically includes: When the echo amplitude of the unit to be detected is greater than the threshold value, it is determined that a target exists; otherwise, it is clutter.

[0050] Based on the previous description, it can be seen that in this embodiment, through the fusion of the DBSCAN clustering algorithm and the CA-CFAR algorithm, the fusion is mainly reflected in that the DBSCAN algorithm can effectively process the density distribution of data points, which highly matches the differences in sea clutter characteristics and target distributions in complex marine environments. Facing diverse noise interference sources in the marine environment, the ability of DBSCAN to identify noise points can remove interference noise in radar echoes for the CA-CFAR algorithm, thereby reducing their adverse effects on the detection results of the CA-CFAR algorithm and improving the reliability of detection.

[0051] Based on the above description, as a specific embodiment: To further verify the performance of DBCA-CFAR detection, the measured data uses the X-band radar measured sea clutter data published in "Journal of Radars" in 2022. The working frequency range of the radar is 9.3~9.5 GHz, the antenna polarization mode is HH, and the antenna is in the staring working state during the test. In this embodiment, 20221113180037_stare_HH.mat is selected as #DATA1, and 20221114020008_stare_HH.mat is selected as #DATA2 to verify the algorithm performance. The data is composed of 950 range cells and 131072 pulse cells. #DATA1 records that the range cells where target 1 is located are 502-514, target 2 is at range cells 667-677, and the sea state is level 4; #DATA2 has target 1, target 2, and target 3 at range cells 502-514, 613-624, and 667-677 respectively, and the sea state is level 3.

[0052] Use DBCA-CFAR to process the echo data #DATA1. Set the reference unit N to 64, the protection units on both sides to 8 each, and the constant false alarm rate Pfa to 10-4. First, preprocess #DATA1, extract the mean value of the time-domain entropy and amplitude characteristics of each range cell, construct a two-dimensional feature vector, set the neighborhood radius (Eps) to 150, and the minimum number of points (MinPts) to 3, and use the DBSCAN clustering algorithm for clustering. The result is as Figure 4 shown. By Figure 4It can be seen that the scatter points are divided into two categories: clutter and targets. The red scatter points are the interfering target clusters, and the blue scatter points are the clutter clusters. This is because the target signals exhibit local high-density aggregation characteristics (red area) in the time-domain entropy - amplitude joint feature space, while the clutter forms a low-density discrete point set due to random distribution (blue area). Its core advantage is that it does not require prior assumptions (such as clutter distribution models, number of targets), and can dynamically adapt to complex environments, breaking through the limitations of traditional CFAR that rely on manual experience for parameter tuning. The range cells corresponding to the characteristics of the interfering targets are filtered out. Figure 5 and Figure 6 By comparison, the amplitudes of the range cells where the targets and a small number of high-power points are located are removed to a certain extent. Figure 7 and Figure 8 It can be seen that the DBCA - CFAR detection threshold obtained by estimating the local background noise power using only the samples within the clutter cluster (blue points) successfully detects two targets, while the CA - CFAR misses the right target. This is because DBCA - CFAR avoids the "threshold elevation" caused by target signals mixing into the reference cells.

[0053] The same as the above steps, the #DATA2 clustering results are as Figure 9 shown. Since the target signal amplitude is significantly higher than the clutter background, TEM is around [0.09:0.14], and the two form significant separability in the joint feature space. Figure 10 and Figure 11 The data comparison before and after screening is shown. The high-amplitude interfering points are effectively filtered out, and the amplitude distribution of the remaining clutter cells becomes more uniform. Finally, the DBCA - CFAR dynamic detection threshold is generated based on the local background power calculated from the pure clutter, as Figure 12 shown. This algorithm successfully detects three targets, while the traditional CA - CFAR detects target 2 as two targets, resulting in a false detection. This may be because the large size of target 2 causes the radar scattering intensity of the target to change at different range cells, resulting in a certain attenuation of the target amplitude. And due to the presence of target 2 and target 3, the CA - CFAR threshold is greatly increased, resulting in the problem that although target 2 can be detected, it is split.

[0054] The main reason for the performance improvement of the improved algorithm in complex marine environments lies in the effective processing of data by the DBSCAN clustering algorithm. Through DBSCAN clustering, it is possible to accurately identify the density distribution regions of sea clutter and targets, providing a more accurate background noise estimate and adaptive detection threshold adjustment for the CA-CFAR algorithm. However, the improved algorithm also has certain limitations. For example, in extremely severe sea conditions, due to the highly non-linear and complex nature of sea clutter, the performance of the algorithm will still be affected to a certain extent. In addition, the computational complexity of the algorithm has increased compared to the traditional CA-CFAR algorithm. In scenarios with extremely high real-time requirements, it may be necessary to further optimize the algorithm implementation to improve the processing speed.

[0055] In summary, the traditional CA-CFAR method relies on the statistical characteristics of the local reference window to estimate the background power, but it is vulnerable to interference in multi-target or non-uniform clutter environments, resulting in a decline in detection performance. The DBCA-CFAR detection scheme based on feature clustering improves the robustness in complex scenarios by introducing data clustering analysis. First, feature extraction is performed on the radar echo data, including amplitude and TEM; subsequently, the data is divided into target classes and background clutter classes through the DBSCAN clustering algorithm; finally, pure background samples are selected based on the clustering results to estimate the noise power, thereby optimizing the detection threshold. This method effectively solves the problem of increased false alarm rate caused by interference in the reference window in multi-target scenarios for traditional methods, while reducing the impact of clutter non-uniformity. Through verification with simulation data and measured data, it can be seen that the improved algorithm can improve the detection probability in multi-target environments and clutter edge environments, and can detect targets that cannot be detected by CA-CFAR in measured sea data, effectively solving the target occlusion problem without losing the original detection performance, indicating the effectiveness of DBCA-CFAR.

[0056] In addition, this application Figure 13 provides a DBCA-CFAR detection system based on feature clustering for the embodiments of this application. As Figure 13 shown, the system provided by the embodiments of this application mainly includes: A two-dimensional construction module 210, configured to extract the TEM feature and amplitude feature corresponding to each preset distance unit from the radar echo data, and then construct a two-dimensional feature vector.

[0057] The two-dimensional construction module 210 includes a TEM feature calculation unit, configured to obtain the TEM feature through the formula: , to obtain the TEM feature; where N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, N + L - 1 represents the number of signals with a length of L, represents the time-domain entropy value of the signal, and , represents the normalized probability density of the i th interval of the signal amplitude, represents the i th interval sample value in the signal sequence of the radar echo data.

[0058] The interference acquisition module 220 is used to obtain the neighborhood radius and the minimum number of points involved in the DBSCAN clustering algorithm, and then input the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points.

[0059] The target detection module 230 is used to set the echo amplitude of the radar echo data of the preset range cell corresponding to the target interference point to 0; based on a sliding window, determine the corresponding cells to be detected and reference cells from the preset range cells; statistically calculate the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cells to be detected, multiply the amplitude mean by a preset normalization factor to obtain a threshold value; determine whether there is a target in the cells to be detected according to the threshold value and the echo amplitude of the cells to be detected.

[0060] The target detection module 230 includes a window determination unit, which is used to slide the window in the order of the preset range cells, with the cell to be detected as the center, and set 1 protection cell and a preset number of reference cells on both sides respectively, where 2N represents the total number of preset reference cells; When the number of protection cells or reference cells corresponding to one side of the cell to be detected is insufficient, it is supplemented by zero-padding or mirror extension.

[0061] The target detection module 230 includes a target determination unit, which is used to determine that there is a target if the echo amplitude of the cell to be detected is greater than the threshold value, otherwise it is clutter.

[0062] In addition, the embodiment of the present application also provides a non-volatile computer storage medium, on which executable instructions are stored, and when the executable instructions are executed, the above-mentioned DBCA-CFAR detection method based on feature clustering is implemented.

[0063] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A DBCA-CFAR detection method based on feature clustering, characterized in that The method includes: Extracting the TEM feature and amplitude feature corresponding to each preset range cell from the radar echo data, and then constructing a two-dimensional feature vector; Obtaining the neighborhood radius and minimum number of points involved in the DBSCAN clustering algorithm, and then inputting the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points; Setting the echo amplitude of the radar echo data of the preset range cell corresponding to the target interference point to 0; Based on a sliding window, determining the corresponding cells to be detected and reference cells from the preset range cells; Statistically calculating the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected, multiplying the amplitude mean by a preset normalization factor to obtain a threshold value; determining whether there is a target in the cell to be detected according to the threshold value and the echo amplitude of the cell to be detected.

2. The DBCA-CFAR detection method based on feature clustering according to claim 1, wherein, Extracting the TEM feature corresponding to each preset range cell, specifically including: Through the formula: , TEM characteristics are obtained; Wherein, N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, and N+L-1 represents the number of signals with a length of L. represents the time-domain entropy value of the signal. and , represents the normalized probability density of the i-th interval of the signal amplitude, represents the i sample value of the j-th interval in the signal sequence of the radar echo data.

3. The DBCA-CFAR detection method based on feature clustering according to claim 1, wherein Based on a sliding window, determining the corresponding cells to be detected and reference cells from the preset range cells, specifically including: The window slides in the order of the preset distance units. With the unit to be detected as the center, one protection unit and a preset number of reference units are set on both sides respectively, and 2N represents the total number of preset reference units; When the number of protection cells or reference cells corresponding to one side of the cell to be detected is insufficient, padding with zeros or mirror extension to make up the number.

4. The DBCA-CFAR detection method based on feature clustering according to claim 1, wherein, Statistically calculating the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected, specifically including: Obtaining the numbers P and Q of the zero echo amplitudes of the reference cells on the left and right sides of the cell to be detected respectively; By -P, obtain the number of non-zero echo amplitudes on the left side, and then sum and average to obtain X; where represents the number of left reference units; By -Q, obtain the number of non-zero echo amplitudes on the right side, and then sum and average to obtain Y; where represents the number of right reference units; Obtaining the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected by taking the average of X + Y.

5. The DBCA-CFAR detection method based on feature clustering according to claim 1, wherein Judging whether there is a target in the cell to be detected through the threshold value and the echo amplitude of the cell to be detected, specifically including: When the echo amplitude of the cell to be detected is greater than the threshold value, it is determined that there is a target, otherwise it is clutter.

6. The DBCA-CFAR detection method based on feature clustering according to claim 1, wherein, Obtaining the neighborhood radius and minimum number of points involved in the DBSCAN clustering algorithm, specifically including: Setting the neighborhood radius to 150 and setting the minimum number of points to 3.

7. A DBCA-CFAR detection system based on feature clustering, characterized in that, The system includes: A two-dimensional construction module for extracting the TEM feature and amplitude feature corresponding to each preset range cell from the radar echo data, and then constructing a two-dimensional feature vector; An interference obtaining module for obtaining the neighborhood radius and minimum number of points involved in the DBSCAN clustering algorithm, and then inputting the two-dimensional feature vector into the DBSCAN clustering algorithm to obtain target interference points and sea clutter points; A target detection module for setting the echo amplitude of the radar echo data of the preset range cell corresponding to the target interference point to 0; based on a sliding window, determining the corresponding cells to be detected and reference cells from the preset range cells; statistically calculating the amplitude mean of the non-zero echo amplitudes of the reference cells corresponding to the cell to be detected, multiplying the amplitude mean by a preset normalization factor to obtain a threshold value; determining whether there is a target in the cell to be detected according to the threshold value and the echo amplitude of the cell to be detected.

8. The DBCA-CFAR detection system based on feature clustering according to claim 7, characterized in that, The two-dimensional construction module includes a TEM feature calculation unit, For passing through the formula: , TEM characteristics are obtained; Wherein, N represents the length of the time-domain signal of the radar echo data to be processed, L represents the length of the preset rectangular window, and N+L-1 represents the number of signals with a length of L. represents the time-domain entropy value of the signal. and , Indicates the normalized probability density of the i th interval of the signal amplitude, and represents the i th interval sample value in the signal sequence of the radar echo data.

9. The DBCA-CFAR detection system based on feature clustering according to claim 7, wherein The target detection module includes a window determination unit, For sliding the window in the order of preset distance units, with the unit to be detected as the center, one protection unit and a preset number of reference units are respectively set on both sides; 2N represents the total number of preset reference units When the number of protection cells or reference cells corresponding to one side of the cell to be detected is insufficient, padding with zeros or mirror extension to make up the number.

10. The DBCA-CFAR detection system based on feature clustering according to claim 7, wherein The target detection module includes a target determination unit, For when the echo amplitude of the cell to be detected is greater than the threshold value, it is determined that there is a target, otherwise it is clutter.

Images