Target detection method and device, electronic equipment and storage medium

Abstract

The application provides a target detection method and device, electronic equipment and storage medium, and relates to the technical field of radar signal processing. The target detection method comprises the following steps: performing adaptive constant false alarm rate detection on distance Doppler spectrum data of a radar target echo signal in a distance dimension and / or a Doppler dimension to obtain an initial detection result; determining data corresponding to a target resolution unit representing a target in the initial detection result in the distance Doppler spectrum data, and updating the value of the target resolution unit with the data to obtain a to-be-judged detection result; dividing the to-be-judged detection result into a set number of data segments along the distance dimension; comparing the data in each data segment with a set threshold value corresponding to the data segment, and generating a target detection result of the target echo signal based on the comparison result, wherein the set threshold value is determined based on the background clutter power of each corresponding data segment. The technical scheme provided by the application can improve the detection rate of radar target detection.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing technology, and in particular to a target detection method, apparatus, electronic device, and storage medium. Background Technology

[0002] Millimeter-wave radar is a transceiver device used for target detection. It obtains information such as the distance, speed, and angle of each target by transmitting millimeter-wave electromagnetic waves into the detection area and receiving the echo signals returned by the target. For example, in the transportation field, traffic millimeter-wave radar can be used for urban traffic control and road condition monitoring.

[0003] Taking traffic millimeter-wave radar detection as an example, after the radar receiving antenna receives the target echo signal, it can convert the target echo signal into a range-Doppler map (RDM) matrix for target detection. Target detection distinguishes the target signal from background noise, thereby detecting the target signal. In related technologies, sliding window detection of the RDM matrix involves traversing all detection units along the entire range dimension of the RDM matrix. However, the background clutter environment in traffic millimeter-wave radar applications is complex and variable. Using existing detection methods leads to a significant deterioration in detection performance as the clutter environment changes, affecting the target detection rate. Summary of the Invention

[0004] This invention provides a target detection method, apparatus, electronic device, and storage medium to address the shortcomings of low detection rates in existing radar target detection technologies and improve the detection rate of radar targets.

[0005] This invention provides a target detection method, comprising:

[0006] Acquire range-Doppler spectrum data of radar target echo signals;

[0007] Adaptive constant false alarm rate (CFAR) detection is performed on the range-Doppler spectrum data in the detection dimension to obtain initial detection results; wherein, the detection dimension includes the range dimension and / or the Doppler dimension;

[0008] The target resolution unit representing the target in the initial detection result is determined to correspond to the data in the range Doppler spectrum data, and the value of the target resolution unit is updated with the data to obtain the detection result to be determined.

[0009] The detection result to be determined is divided into a set number of data segments along the distance dimension;

[0010] The data in each data segment is compared with a set threshold corresponding to each data segment, and a target detection result of the target echo signal is generated based on the comparison result. The set threshold is determined based on the background clutter power of the corresponding data segment.

[0011] According to a target detection method provided by the present invention, the step of comparing the data in each data segment with a set threshold corresponding to each data segment, and generating a target detection result of the target echo signal based on the comparison result, includes:

[0012] In each data segment, data that is less than a predetermined threshold corresponding to that data segment is set to 0 in order to update each data segment;

[0013] All updated data segments are identified as the target detection results of the target echo signal.

[0014] According to a target detection method provided by the present invention, the step of performing adaptive constant false alarm rate (CFAR) detection on the range Doppler spectrum data in the detection dimension to obtain an initial detection result includes:

[0015] Each resolution unit in the range Doppler spectrum data is sequentially used as a detection unit, and a detection window for the detection unit is determined in the detection dimension. The detection window in the detection dimension includes the detection unit and the forward and backward sliding windows in the detection dimension where the detection unit is located.

[0016] The clutter state within the detection window and the mean ratio of the detection window are determined, and the detection threshold of the detection unit is determined based on the clutter state, the mean ratio, and a set constant false alarm rate, wherein the clutter state includes uniform clutter or non-uniform clutter.

[0017] The data from the detection unit is compared with the detection threshold, and the initial detection result is generated based on the comparison result.

[0018] According to a target detection method provided by the present invention, the detection dimension includes a distance dimension, and the step of determining the detection sliding window of the detection unit on the detection dimension includes:

[0019] Along the distance dimension of the detection unit, starting from the forward adjacent unit of the detection unit, a fifth preset number of resolution units are continuously selected to obtain the first protection unit;

[0020] Starting from the adjacent unit of the detection unit, the fifth preset number of resolution units are selected sequentially to obtain the second protection unit;

[0021] The third preset number of resolution units before the first protection unit and the third preset number of resolution units after the second protection unit are respectively defined as forward reference units and backward reference units;

[0022] The first protection unit and the forward reference unit are defined as the forward sliding window of the detection unit in the distance dimension direction, and the second protection unit and the backward reference unit are defined as the backward sliding window of the detection unit in the distance dimension direction.

[0023] The forward sliding window in the distance dimension, the detection unit, and the backward sliding window in the distance dimension are defined as the detection sliding window in the distance dimension.

[0024] According to a target detection method provided by the present invention, the detection dimension includes Doppler dimensions, and the step of determining the detection sliding window of the detection unit on the detection dimension includes:

[0025] Along the Doppler direction where the detection unit is located, determine whether there is a fourth preset number of resolution units in the forward resolution unit and the backward resolution unit of the detection unit;

[0026] If there are no fourth preset number of resolution units in the forward direction of the detection unit, all forward resolution units of the detection unit and the first supplementary resolution unit are determined as forward sliding windows of the detection unit in the Doppler direction, and the fourth preset number of backward resolution units of the detection unit are determined as backward sliding windows of the detection unit in the Doppler direction. The first supplementary resolution unit is a first number of resolution units starting from the backward boundary resolution unit of the detection unit, and the total number of all forward resolution units and the first supplementary resolution unit of the detection unit is the fourth preset number.

[0027] If the fourth preset number of resolution units does not exist in the backward direction of the detection unit, all backward resolution units of the detection unit and the second supplementary resolution units are determined as the backward sliding window of the detection unit in the Doppler direction, and the fourth preset number of resolution units in the forward direction of the detection unit are determined as the forward sliding window of the detection unit in the Doppler direction. The second supplementary resolution unit is a second number of resolution units starting from the forward boundary resolution unit of the detection unit, and the total number of all backward resolution units and the second supplementary resolution units of the detection unit is the fourth preset number.

[0028] When the fourth preset number of resolution units exist in both the forward and backward directions of the detection unit, the fourth preset number of resolution units are selected before and after the detection unit to obtain the forward sliding window and the backward sliding window of the detection unit in the Doppler direction.

[0029] The forward sliding window in the Doppler direction, the detection unit, and the backward sliding window in the Doppler direction are defined as the detection sliding window in the Doppler direction.

[0030] According to a target detection method provided by the present invention, determining the ratio of the clutter state within the detection sliding window to the mean of the detection sliding window includes:

[0031] For each of the detection dimensions, the data of all reference units within the detection sliding window of each detection dimension are sorted to obtain the sorting result;

[0032] Remove the first and second preset quantities of data from the sorting results to obtain valid reference unit data;

[0033] Based on the effective reference unit data, determine the clutter state within the detection window of each detection dimension and the mean ratio of the detection windows of each detection dimension.

[0034] According to a target detection method provided by the present invention, determining the clutter state within the detection sliding window of each detection dimension and the mean ratio of the detection sliding windows of each detection dimension based on the effective reference cell data includes:

[0035] Based on the data corresponding to the forward sliding window in the effective reference unit data, a first change index and a first power mean are determined;

[0036] Based on the data corresponding to the backward sliding window in the effective reference unit data, the second change index and the second power mean are determined;

[0037] The clutter state in the forward sliding window is determined according to the first change index, and the clutter state in the backward sliding window is determined according to the second change index.

[0038] Calculate the ratio of the first power mean to the second power mean to obtain the mean ratio of the detection sliding window for each detection dimension.

[0039] The present invention also provides a target detection device, comprising:

[0040] The acquisition module is used to acquire range Doppler spectrum data of radar target echo signals;

[0041] The detection module is used to perform adaptive constant false alarm rate detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein, the detection dimension includes the range dimension and / or the Doppler dimension;

[0042] The determination module is used to determine the data corresponding to the target resolution unit representing the target in the initial detection result in the range Doppler spectrum data, and update the value of the target resolution unit with the data to obtain the detection result to be determined;

[0043] The segmentation module is used to divide the detection result to be determined into a set number of data segments along the distance dimension;

[0044] The generation module is used to compare the data in each data segment with the set threshold corresponding to each data segment, and generate the target detection result of the target echo signal based on the comparison result.

[0045] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the target detection method as described above.

[0046] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the target detection method as described above.

[0047] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the target detection method as described above.

[0048] The target detection method, apparatus, electronic device, and storage medium provided by this invention first perform adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data of radar target echo signals in the detection dimension to obtain an initial detection result. Next, the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data is determined in the initial detection result, and the value of the target resolution unit is updated using this data to obtain a detection result to be determined. Then, the detection result to be determined is divided into a predetermined number of data segments along the range dimension. The data in each data segment is compared with a predetermined threshold corresponding to each data segment, and a target detection result of the target echo signal is generated based on the comparison result. In this way, the range-Doppler spectrum data of radar target echo signals can be segmented along the range dimension. For each range segment, a corresponding predetermined threshold can be set according to the background clutter power of that range segment. Further target determination is performed on the initial detection result based on this predetermined threshold, taking into account different background clutter environments at different range segments, thereby effectively detecting targets and improving the target detection rate. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0050] Figure 1 This is one of the flowcharts of the target detection method provided by the present invention;

[0051] Figure 2 This is a schematic diagram of the distance dimension segmentation method provided by the present invention;

[0052] Figure 3 This is a schematic diagram of the distance dimension detection sliding window provided by the present invention;

[0053] Figure 4 This is a schematic diagram illustrating the principle of the Doppler detection sliding window for determining the detection unit provided by the present invention;

[0054] Figure 5 This is the second flowchart of the target detection method provided by the present invention;

[0055] Figure 6 This is a schematic diagram illustrating the principle of adaptive constant false alarm rate detection provided by the present invention;

[0056] Figure 7 This is a schematic diagram of the target detection device provided by the present invention;

[0057] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0059] In radar signal processing, electromagnetic echo signals include background clutter such as noise, noise, and interference signals. Target detection can distinguish between target signals and background clutter from this electromagnetic echo signal, thereby detecting the true target signal. Specifically, after the radar receiving antenna receives the electromagnetic echo signal, it can convert the electromagnetic echo signal into a range-Doppler map (RDM) matrix for target detection, thus distinguishing between target signals and background clutter. RDM refers to the range-Doppler two-dimensional spectrum map containing target information obtained by mixing the target echo signal received by the radar receiving antenna with the local oscillator signal, and then performing range-dimensional Fast Fourier Transform (FFT) and Doppler-dimensional FFT processing on the mixed signal.

[0060] Constant False Alarm Rate (CFAR) detection is an important part of radar signal processing. It is a technique that, while maintaining a constant false alarm rate, distinguishes the target echo signal from the background clutter environment in the signal received by the radar receiver, determines the presence of the target echo signal, and maximizes the target detection probability.

[0061] Taking traffic millimeter-wave radar as an example, it can be applied to scenarios such as traffic control and road condition monitoring. Through millimeter-wave radar detection, it can obtain signals related to the distance, speed, angle and other information of vehicles and pedestrians in the current traffic scene in real time. Then, through signal processing, it can obtain the actual position of the target object and track its movement trajectory, thereby identifying traffic congestion, abnormal events and other events on the road, and can also count the traffic flow, queue length and other information of the corresponding road to help optimize road traffic conditions.

[0062] Because the background clutter environment in real-world traffic millimeter-wave radar applications is quite complex, including uniform clutter, multi-target detection, and edge clutter environments, the target detection performance of traffic millimeter-wave radar deteriorates significantly with changes in the background clutter environment. Current technologies simply perform sliding window detection across the entire range dimension of the RDM matrix, failing to capture the complex and varied background clutter environment at different ranges in real-world applications. Furthermore, as the detection distance increases, the energy of the echo signal gradually attenuates, and complex environmental factors lead to significant differences in background noise at different distances. This severely impacts the estimation accuracy of background clutter and the detection rate of true targets.

[0063] Based on this, embodiments of the present invention provide a target detection method that performs adaptive constant false alarm rate (CFAR) detection on range-Doppler spectrum data in the range dimension and / or the Doppler dimension to obtain an initial detection result; determines the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data in the initial detection result, and updates the value of the target resolution unit with this data to obtain a detection result to be determined; divides the detection result to be determined into a predetermined number of data segments along the range dimension; compares the data in each data segment with a predetermined threshold determined based on the background clutter power of each data segment, and generates a target detection result of the target echo signal based on the comparison result. By performing segmentation processing in the range dimension and further determining the target based on the predetermined threshold determined based on the background clutter power of each range segment, different background clutter environments at different range segments can be considered, improving the target detection detection rate and effectively detecting targets.

[0064] The following is combined Figures 1-7 The target detection method of the present invention is described below. This target detection method can be applied to electronic devices such as servers, mobile phones, and computers, and can also be applied to target detection devices installed in such electronic devices. These target detection devices can be implemented through software, hardware, or a combination of both.

[0065] Figure 1 An exemplary schematic diagram of one of the target detection methods provided in an embodiment of the present invention is shown below, with reference to... Figure 1 As shown, the target detection method may include the following steps 110 to 150.

[0066] Step 110: Acquire the range Doppler spectrum data of the radar target echo signal.

[0067] After receiving the target echo signal through its receiving antenna, the radar receiver mixes the target echo signal with the local oscillator signal. Then, it performs range-dimensional FFT and Doppler-dimensional FFT processing on the mixed signal to obtain the range-Doppler spectrum data of the target echo signal. Electronic devices can then acquire this range-Doppler spectrum data from the radar receiver.

[0068] For example, performing range-dimensional FFT and Doppler-dimensional FFT processing on the mixed signal yields a×b-dimensional range-Doppler spectrum data, where each data point represents a power value. This range-Doppler spectrum data is represented in matrix form, where 'a' represents the number of frequency points processed in the Doppler-dimensional FFT, and 'b' represents the number of frequency points processed in the range-dimensional FFT.

[0069] In this embodiment of the invention, each matrix element in the matrix of range Doppler spectrum data can be defined as a resolution unit, and the value of the resolution unit is the power value.

[0070] Step 120: Perform adaptive constant false alarm rate detection on the detection dimension of the distance Doppler spectrum data to obtain the initial detection results.

[0071] The detection dimension can include at least one of the range dimension and the Doppler dimension. That is, adaptive constant false alarm rate (CFAR) detection can be performed on the range-Doppler spectral data in the range dimension to obtain an initial detection result; or, adaptive CFAR detection can be performed on the range-Doppler spectral data in the Doppler dimension to obtain an initial detection result; or, adaptive CFAR detection can be performed on both the range dimension and the Doppler dimension of the range-Doppler spectral data to obtain an initial detection result.

[0072] For example, performing adaptive constant false alarm rate (CFAR) detection on both the range and Doppler dimensions of the range-Doppler spectrum data to obtain initial detection results can include: performing adaptive CFAR detection on the range dimension of the range-Doppler spectrum data to obtain a range dimension detection result; performing adaptive CFAR detection on the Doppler dimension of the range-Doppler spectrum data to obtain a Doppler dimension detection result; and performing a logical AND operation on the range dimension detection result and the Doppler dimension detection result to obtain the initial detection result. In this way, by comprehensively considering information from both dimensions, the accuracy and precision of adaptive CFAR detection can be improved, further increasing the target detection rate.

[0073] Constant false alarm rate (CFAR) detection can estimate the clutter power of the detected cell based on reference cells near it, thereby adaptively setting a threshold to maintain constant CFAR performance. Depending on the clutter power estimation method, CFAR detection methods can be divided into two categories: mean-based CFAR detection and ordered statistical CFAR detection. Mean-based CFAR detection can include three methods: Cell Average CFAR (CA-CFAR), Maximum Cell Average CFAR (GO-CFAR), and Minimum Cell Average CFAR (SO-CFAR). These CFAR detection methods exhibit different detection effects in various clutter environments, such as uniform clutter environments, multi-target environments, and clutter edge environments.

[0074] Therefore, in the constant false alarm rate (CFAR) detection process, a CFAR detection strategy can be adaptively selected for different clutter environments to leverage the advantages of various CFAR detection methods and improve the detection performance. This process is known as adaptive CFAR detection.

[0075] Step 130: Determine the data corresponding to the target resolution unit in the range Doppler spectrum data in the initial detection result, and update the value of the target resolution unit with the data to obtain the detection result to be determined.

[0076] For example, in the initial detection result, "1" can represent the presence of a target, and "0" can represent the absence of a target. The resolution unit corresponding to "1" can be identified as the target resolution unit. For each target resolution unit, there is corresponding data in the original range-Doppler spectrum data, namely the corresponding power value. This power value can be used to replace the "1" of the target resolution unit, while the "0" remains unchanged. This updates the initial detection result, yielding the detection result to be determined. In the updated detection result to be determined, the location with a target is represented by the power value at that location, and the location without a target is represented by "0".

[0077] Step 140: Divide the detection result to be judged into a set number of data segments along the distance dimension.

[0078] For example, the detection result to be judged can be divided into a predetermined number of uniform data segments along the distance dimension, meaning each data segment contains the same number of resolution units. Alternatively, the detection result to be judged can be divided into a predetermined number of non-uniform data segments along the distance dimension, depending on the specific application.

[0079] For example, Figure 2 An exemplary diagram illustrating the segmentation method in the distance dimension is shown below. Figure 2 As shown, the detection result to be judged can be divided into four uniform data segments, R1, R2, R3, and R4, along the distance dimension, according to segmentation method 1, with each data segment having the same number of resolution units. Alternatively, the detection result to be judged can be divided into four non-uniform data segments, R5, R6, R7, and R8, along the distance dimension, according to segmentation method 2, with each data segment having a different number of resolution units.

[0080] It should be noted that, Figure 2 This is merely an illustrative example and is not intended to limit the invention. The number of data segments and the number of resolution units in each data segment can be determined according to the actual application.

[0081] For example, the number can be preset or determined based on the distance dimension length and segmentation method of the detected distance-Doppler spectrum data. For instance, under the selected segmentation method, it is ensured that each data segment has at least 2 resolution units.

[0082] It is understandable that segmenting the detection results into distance segments along the distance dimension and then processing the data in each distance segment is equivalent to segmenting the range-Doppler spectrum data along the distance dimension.

[0083] Step 150: Compare the data in each data segment with the set threshold corresponding to that data segment, and generate the target detection result of the target echo signal based on the comparison result.

[0084] The set threshold for each data segment is determined based on the background clutter power of that data segment.

[0085] For example, in a scenario without moving targets, range-Doppler spectrum data of radar target echo signals can be acquired. The values ​​of the resolution cells corresponding to stationary targets in the zero-Doppler channel are removed from this range-Doppler spectrum data to filter out interference from stationary objects, resulting in processed range-Doppler spectrum data. Then, this processed range-Doppler spectrum data is divided in the same way as in step 140. For each data segment, the mean of all data in that segment is calculated to obtain the mean background clutter power of that segment. A set multiple of this mean background clutter power can be determined as a set threshold for that range segment, thus obtaining the set threshold corresponding to each data segment. The set multiple can be, for example, 2 times, 2.5 times, 3 to 4 times, etc., and this embodiment of the invention does not impose any special limitations on this.

[0086] After dividing the data into a set number of segments, the data in each segment that is less than the set threshold corresponding to each segment is set to 0 to update each segment. Then, all the updated data segments are determined as the target detection results of the target echo signal.

[0087] Specifically, for each data segment, the data representing the target in that segment is compared sequentially with a set threshold corresponding to that segment. If the data at the current comparison position is less than the set threshold, it is assumed that no target exists at that position, and the data at that position is set to 0. If the data at the current comparison position is greater than or equal to the set threshold, it is assumed that a target exists at that position, and the data at that position is retained. By traversing all data segments, the final target detection result can be obtained.

[0088] The target detection method provided in this invention first performs adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data of the radar target echo signal to obtain an initial detection result. Then, it determines the data corresponding to the target resolution unit in the range-Doppler spectrum data in the initial detection result and updates the value of the target resolution unit with this data to obtain a detection result to be determined. Next, the detection result to be determined is divided into a predetermined number of data segments along the range dimension. The data in each data segment is compared with a predetermined threshold corresponding to each data segment, and a target detection result for the target echo signal is generated based on the comparison result. In this way, the range-Doppler spectrum data of the radar target echo signal can be segmented along the range dimension. For each range segment, a corresponding predetermined threshold can be set according to the background clutter power of that range segment. Based on this predetermined threshold, further target determination is performed on the initial detection result. This method considers different background clutter environments at different range segments, thereby effectively detecting targets and improving the target detection rate.

[0089] based on Figure 1 In a corresponding embodiment of the target detection method, in one example, adaptive constant false alarm rate (CFAR) detection is performed on range-Doppler spectrum data in the detection dimension to obtain an initial detection result. This may include: sequentially using each resolution unit in the range-Doppler spectrum data as a detection unit, determining the detection window for that detection unit in the detection dimension; determining the clutter state within the detection window and the mean ratio of the detection window, and determining the detection threshold for that detection unit based on the clutter state, the mean ratio, and a set CFAR; comparing the data corresponding to that detection unit with the detection threshold of that detection unit, and generating a detection result based on the comparison result. The clutter state may include uniform clutter or non-uniform clutter. The detection window in the detection dimension includes the detection unit and the forward and backward sliding windows in the detection dimension where the detection unit is located.

[0090] For example, the detection threshold of the detection unit can be determined by looking up a table based on the clutter state, the mean ratio, and the set constant false alarm rate.

[0091] In this embodiment of the invention, the detection unit is a resolution unit in the data segment, that is, a matrix element unit in the data segment, and the data in the detection unit can represent a spectral power value of the target echo signal.

[0092] In this embodiment of the invention, adaptive constant false alarm rate (CFAR) detection in the range dimension can be performed on the detection unit in the range-Doppler spectrum data to obtain an initial detection result; alternatively, adaptive CFAR detection in the Doppler dimension can be performed on the detection unit in the range-Doppler spectrum data to obtain an initial detection result; or, adaptive CFAR detection in both the range dimension and the Doppler dimension can be performed on the detection unit in the range-Doppler spectrum data to obtain an initial detection result. When performing adaptive CFAR detection on any dimension of the detection unit, it is necessary to first determine the detection window of the detection unit in that dimension. The detection window may include a protection unit, a detection unit, and a reference unit.

[0093] In this embodiment of the invention, the detection window of the detection unit in the distance dimension direction can be defined as a distance dimension detection window, and the detection window of the detection unit in the Doppler dimension direction can be defined as a Doppler dimension detection window.

[0094] For example, determining the detection window of the detection unit in the distance dimension direction may include: continuously selecting a fifth preset number of resolution units forward from the forward adjacent unit of the detection unit along the distance dimension direction where the detection unit is located to obtain a first protection unit; continuously selecting a fifth preset number of resolution units backward from the backward adjacent unit of the detection unit to obtain a second protection unit; determining a third preset number of resolution units before the first protection unit and a third preset number of resolution units after the second protection unit as a forward reference unit and a backward reference unit, respectively; determining the first protection unit and the forward reference unit as a forward sliding window of the detection unit in the distance dimension direction, and determining the second protection unit and the backward reference unit as a backward sliding window of the detection unit in the distance dimension direction; and determining the forward sliding window, the detection unit, and the backward sliding window in the distance dimension direction as a detection window in the distance dimension.

[0095] For example, Figure 3 An exemplary diagram of a distance dimension detection sliding window provided in an embodiment of the present invention is shown, with reference to... Figure 3 As shown, the distance-dimensional detection sliding window 30 may include a detection unit and a forward sliding window 31 and a backward sliding window 32 in the distance-dimensional direction where the detection unit is located. The forward sliding window 31 may include a fifth preset number of first protection units and a third preset number of reference units in the distance-dimensional direction preceding the first protection unit. The first protection unit may, for example, be a forward adjacent unit 311 of the detection unit. The backward sliding window 32 may include a fifth preset number of second protection units and a third preset number of reference units in the distance-dimensional direction following the second protection unit. The second protection unit may, for example, be a backward adjacent unit 321 of the detection unit. For example, in... Figure 3 In the middle, both the forward sliding window 31 and the backward sliding window 32 include one protection unit and three reference units.

[0096] Understandably, in the distance dimension, when the detection unit approaches the distance dimension boundary, the sliding window on the side closer to the boundary will be smaller than or disappear on the other side. For example, in... Figure 3 In the process, when the detection unit is close to the right boundary of the distance dimension, the forward sliding window 21 of the formed distance dimension detection sliding window 20 is reduced to only one reference cell, which is smaller than the backward sliding window 22. It is understandable that for the detection unit close to the boundary of the distance dimension, the number of resolution cells of the sliding window taken on the side close to the boundary may be insufficient. In this case, all resolution cells on the side close to the boundary can be determined as the sliding window on that side to meet the selection requirements of the detection sliding window as much as possible.

[0097] When performing sliding window detection on Doppler units near the boundary, there may be situations where the number of reference units on the Doppler boundary side of the current detection unit is small or non-existent. In such cases, the only available reference units can be used directly for background clutter estimation. However, this method suffers from insufficient training samples, leading to inaccurate background clutter estimation around detection units near the Doppler boundary and reducing the target detection rate of these units. Furthermore, real-world application scenarios often involve highly complex background clutter environments, making it difficult for a single clutter estimation model or algorithm to accurately fit the actual background clutter conditions.

[0098] Based on this, in one exemplary embodiment of the present invention, the Doppler detection window corresponding to a detection unit can be determined based on the number of forward and backward resolving units along the Doppler dimension in which the detection unit is located. For example, based on the circular convolution characteristics of velocity, for a detection unit located near the boundary of the Doppler dimension in the Doppler spectrum data, resolving units on the boundary side of the Doppler dimension away from the boundary can be taken as a supplement to the missing reference units of the detection unit, increasing the number of reference units. This allows for accurate estimation of the background clutter threshold, effectively improving the robustness and detection rate of target detection.

[0099] For example, for the Doppler direction, determining the detection sliding window of the detection unit in the Doppler direction may include:

[0100] Along the Doppler direction where the detection unit is located, it is determined whether there is a fourth preset number of resolution units in the forward and backward resolution units of the detection unit; if there is no fourth preset number of resolution units in the forward direction of the detection unit, all the forward resolution units of the detection unit and the first supplementary resolution unit are determined as the forward sliding window of the detection unit in the Doppler direction, and the fourth preset number of backward resolution units of the detection unit are determined as the backward sliding window of the detection unit in the Doppler direction. The first supplementary data unit is the first number of resolution units starting from the backward boundary resolution unit of the detection unit, and the total number of all the forward resolution units and the first supplementary resolution units of the detection unit is the fourth preset number.

[0101] If there is no fourth preset number of resolution units in the backward direction of the detection unit, all backward resolution units of the detection unit and the second supplementary resolution units are determined as backward sliding windows of the detection unit in the Doppler direction, and the fourth preset number of forward resolution units of the detection unit are determined as forward sliding windows of the detection unit in the Doppler direction. The second supplementary resolution units are the second number of resolution units starting from the forward boundary resolution unit of the detection unit, and the total number of all backward resolution units and the second supplementary resolution units of the detection unit is the fourth preset number.

[0102] When there are a fourth preset number of resolution units in both the forward and backward directions of the detection unit, a fourth preset number of resolution units are selected before and after the detection unit to obtain the forward sliding window and the backward sliding window of the detection unit in the Doppler direction.

[0103] The forward sliding window, the detection unit, and the backward sliding window in the Doppler direction are defined as the detection sliding window in the Doppler direction.

[0104] For example, Figure 4 An exemplary schematic diagram of the Doppler detection sliding window for determining the detection unit is shown, with reference to... Figure 4 As shown, the arrows pointing in the Doppler direction indicate forward direction, and vice versa. Taking a Doppler detection sliding window with a length of 9 resolution units, 3 reference units on one side, and 1 protection unit on one side as an example, then for each detection unit, 4 resolution units are needed on both sides to form a Doppler detection sliding window.

[0105] For detection unit A1, there are 5 forward resolving units from the upper boundary and 6 backward resolving units from the lower boundary, both of which are greater than 4. Therefore, 4 resolving units can be selected before and after detection unit A1 to form the Doppler detection sliding window 41 of detection unit A1 together with detection unit A1.

[0106] For detection unit A2, there are 10 forward resolving units from the upper boundary. The number of forward resolving units meets the requirements for forming a Doppler detection sliding window. Four resolving units in front of detection unit A2 can be directly selected. The resolving units adjacent to detection unit A2 in the front are used as protection units, and the other three are used as reference units. There is only one backward resolving unit from the lower boundary, which is less than 4. It is necessary to add three resolving units as reference units, starting from the upper boundary resolving unit (i.e., the forward boundary resolving unit) 421, namely resolving unit 421, resolving unit 422, and resolving unit 423, to finally form the Doppler detection sliding window 42 of detection unit A2.

[0107] For detection unit A3, there is only one forward resolving unit from the upper boundary, which is less than 4. It is necessary to add 3 resolving units upward from the lower boundary resolving unit (i.e., the backward boundary resolving unit) 433 as reference units, namely resolving unit 433, resolving unit 432, and resolving unit 431. There are 10 backward resolving units from the lower boundary. The number of backward resolving units meets the requirements for forming a Doppler detection sliding window. Four resolving units backward of detection unit A3 can be directly selected. The resolving units adjacent to the backward resolving unit of detection unit A3 are used as protection units, and the other three are used as reference units, thus forming the Doppler detection sliding window 43 of detection unit A3.

[0108] In this way, for the situation where there is insufficient or no reference cell on the side of the Doppler dimension near the boundary of the detection unit, the circular convolution property of velocity can be used to supplement the reference cell by using the resolution cell on the other side of the current Doppler dimension of the detection unit. This increases the number of training samples for the detection unit near the boundary, making the estimation of background clutter more accurate, thereby improving the target detection rate of the Doppler dimension near the boundary detection unit.

[0109] In one example embodiment, determining the clutter state within a detection window and the mean ratio of the detection window may include: for each detection dimension, sorting the data of all reference cells within the detection window of each detection dimension to obtain a sorting result; removing the first and second preset numbers of data from the sorting result to obtain valid reference cell data; and determining the clutter state within the detection window of each detection dimension and the mean ratio of the detection window of each detection dimension based on the valid reference cell data. This can eliminate interference signals with high and low clutter power, improving the accuracy of clutter state detection.

[0110] It is understandable that, for each detection unit, when the detection dimensions are distance dimension and Doppler dimension, the data of all reference units in the distance dimension detection window and the data of all reference units in the Doppler dimension detection window are sorted separately to obtain the sorting results of the two windows, thereby determining the effective reference unit data, clutter state and mean ratio of each.

[0111] In one example embodiment, determining the clutter state within the detection window of each detection dimension and the mean ratio of the detection windows for each detection dimension based on effective reference cell data may include: determining a first change index and a first power mean based on data corresponding to the forward sliding window in the effective reference cell data; determining a second change index and a second power mean based on data corresponding to the backward sliding window in the effective reference cell data; determining the clutter state within the forward sliding window based on the first change index, and determining the clutter state within the backward sliding window based on the second change index; and calculating the ratio of the first power mean and the second power mean to obtain the mean ratio of the detection windows for each detection dimension.

[0112] For example, the first change index and the second change index can be compared with a set threshold value. If they are less than or equal to the set threshold value, the corresponding clutter state is determined to be uniform clutter. If they are greater than the set threshold value, the corresponding clutter state is determined to be non-uniform clutter.

[0113] Based on the methods of the above embodiments, the following is combined with... Figure 5 and Figure 6 The target detection method provided in the embodiments of the present invention will be further illustrated with examples.

[0114] Figure 5 This is an exemplary schematic diagram of the target detection method provided in an embodiment of the present invention, with reference to... Figure 5 As shown, the target detection method may include the following steps 501 to 513.

[0115] Step 501: Acquire the range Doppler spectrum data of the radar target echo signal.

[0116] Traffic millimeter-wave radar transmits frequency-modulated continuous waves via a transmitting antenna. After the receiving antenna receives the target echo signal, electronic equipment can perform frequency mixing processing on the target echo signal to obtain an intermediate frequency (IF) signal. Then, it performs Fast-Time FFT processing on the IF signal in both the fast and slow time dimensions, i.e., range-dimensional and Doppler-dimensional FFT processing, to obtain the range-Doppler spectrum data corresponding to each frame of the signal. This range-Doppler spectrum data can be represented in matrix form, for example, obtaining an a×b dimensional range-Doppler spectrum data matrix M. RDM .

[0117] After obtaining the range Doppler spectrum data of the radar target echo signal, steps 502 and 506 can be executed respectively.

[0118] Step 502: Determine the distance dimension detection sliding window for each detection unit in the distance Doppler spectrum data.

[0119] The distance Doppler spectrum data matrix M RDM Each resolution unit in the array serves as a detection unit in turn, along the distance dimension, i.e., M. RDM In the row direction, a total of (2P+2N) resolution units are selected on both sides of the detection unit to form the distance dimension detection sliding window of the detection unit together with the detection unit. Wherein, P is the number of protection units on one side, N is the number of reference units on one side, and P and N are both positive integers.

[0120] If the detection unit is close to the boundary of the distance dimension, the number of resolution units of the sliding window taken on the side close to the boundary of the distance dimension may be insufficient. In this case, try to meet the selection requirements of the distance dimension detection sliding window, and all resolution units on the side close to the boundary can be determined as the sliding window on that side.

[0121] Step 503: Remove interference clutter and determine the clutter state within the distance dimension detection window and the mean ratio of the distance dimension detection window.

[0122] For each detection unit, after determining the range-dimensional detection window for that unit, the data (power values) in all reference units on both sides of the detection unit are sorted by size. Then, the first K1 and last K2 data points in the sorted results are removed to eliminate interference with large and small power. The remaining data are then placed back in their original reference unit positions. Finally, the clutter state within the range-dimensional detection window is determined using the reference units and data corresponding to the remaining data. Here, K1 and K2 are positive integers, and they can be equal or unequal.

[0123] Figure 6 An exemplary schematic diagram of the principle of adaptive constant false alarm rate detection is shown below. Figure 6 As shown, d is the detection unit, p1 and p2 are the protection units on the front and rear sides of the detection unit d, and x1~x N / 2 For the reference cell data within the forward sliding window, x N / 2+1 ~x N After interference and clutter removal processing, the reference cell data within the backward sliding window can be processed from x1 to x N / 2 and x N / 2+1 ~x N The effective reference cell data are determined separately. Then, the sum of the effective reference cell data within the forward sliding window, ∑, can be calculated. l and the first power average μ l Calculate the sum of the effective reference cell data within the sliding window ∑r and the second power mean μ r And calculate the sum of all valid reference cell data ∑ l +∑ r Among them, μ l and μ r It can be calculated using formula (1) and formula (2) respectively.

[0124] Formula (1) is

[0125] Formula (2) is

[0126] Where n represents the total number of valid reference cell data within the forward sliding window, and x i This represents the i-th valid reference cell data within the forward sliding window; m represents the total number of valid reference cell data within the backward sliding window, x j This represents the j-th valid reference cell data within the backward sliding window.

[0127] The mean μ is obtained l and μ r Then, the first change index V of the forward sliding window can be calculated using the following formulas (3) and (4) respectively. l And the second change index V of the backward sliding window r .

[0128] Formula (3) is

[0129] Formula (4) is

[0130] Where n represents the total number of valid reference cell data within the forward sliding window, x i This represents the i-th valid reference cell data within the forward sliding window; m represents the total number of valid reference cell data within the backward sliding window, x j This represents the j-th valid reference cell data within the backward sliding window.

[0131] Then, V l and V r The clutter state within the forward and backward sliding windows is determined by comparing it with the change index threshold. If the change index threshold is less than or equal to the threshold, the clutter state is determined to be uniform clutter; otherwise, the clutter state is determined to be non-uniform clutter. The change index threshold can be determined based on engineering experiments.

[0132] Combination Figure 6 The mean μ is obtained. l and μ r Then, the mean ratio V of the power mean of the sliding windows before and after the detection unit d can be calculated using formula (5). MR, which is the ratio of the mean values ​​of the sliding window used for distance dimension detection.

[0133] Formula (5) is V MR =μ l / μ r .

[0134] Get V MR After that, V can be used MR The threshold K is the ratio of the mean. MR The power average of the front and rear sliding windows is compared to determine if they are the same. If V MR satisfy The mean clutter power of the front and rear sliding windows is the same. or V MR >K MR The mean clutter power of the front and rear sliding windows differs. Where K... MR Values ​​can be obtained based on engineering experiments.

[0135] Step 504: Determine the distance dimension detection threshold for each detection unit based on the clutter state, mean ratio, and set constant false alarm rate.

[0136] False alarm rate (CFAR) refers to the probability that a radar detection method, using a threshold detection method, will detect a target when it is not actually present, due to the prevalence and fluctuations of noise. A constant false alarm rate (CFAR) aims to maintain a constant CFAR. For example, based on Table 1 below, the equivalent CFAR strategy and the corresponding background product constant C can be determined according to the clutter state in the forward sliding window, the clutter state in the backward sliding window, and the determination of whether the average power values ​​of the forward and backward sliding windows are the same. num The background product constant C num This can represent determining the adaptive threshold T. VI The number of effective reference units used. Specifically, referring to Table 1, ∑ l ∑ represents the sum of the effective reference cell data within the forward sliding window. r This represents the sum of effective reference cell data within the backward sliding window. An equivalent CFAR strategy can be determined based on the clutter state within the forward and backward sliding windows and the power mean values ​​of the forward and backward sliding windows being the same. This strategy is then applied according to the "adaptive threshold T". VI The column corresponding to the content determines the sliding window and the corresponding C used when estimating the background clutter power. num When both forward and backward sliding windows are needed, C num C represents the number of all valid reference cells within the forward and backward sliding windows used to estimate background clutter power; when only one side of the forward and backward sliding windows is used, C num This indicates the number of effective reference cells within one sliding window used to estimate background clutter power. For example, if the clutter state within both the forward and backward sliding windows is non-uniform and the mean values ​​of the forward and backward sliding windows are the same, then according to the "adaptive threshold T"...VI The content corresponding to column C num (∑ l +∑ r It can be determined that forward sliding window and backward sliding window are needed when estimating background clutter power, then the corresponding C num This represents the number of all valid reference cells within the forward and backward sliding windows used to estimate background clutter power. For example, if the clutter state within the forward sliding window is uniform and the clutter state within the backward sliding window is non-uniform, based on the "adaptive threshold T..." VI The content corresponding to column C num ∑ r It can be determined that a backward sliding window is needed when estimating background clutter power, then the corresponding C num This indicates the number of effective reference cells in the backward sliding window used to estimate background clutter power.

[0137] Table 1

[0138]

[0139] Where, max(∑ l ,∑ r ) represents ∑ l and ∑ r The maximum value, min(∑ l ,∑ r ) represents ∑ l and ∑ r The minimum value.

[0140] According to Table 1, the corresponding adaptive threshold T can be selected based on the determination results of the forward sliding window clutter state (clutter state within the forward sliding window), the backward sliding window clutter state (clutter state within the backward sliding window), and whether the mean values ​​of the forward and backward sliding windows are the same. VI The calculation method is determined, and the equivalent CFAR policy is identified. Then, it can be applied according to the selected adaptive threshold T. VI The adaptive threshold T is calculated using the calculation method. VI And based on the set constant false alarm rate P fa Using the threshold factor solution method expressed by formula (6), the threshold factor α of the corresponding equivalent CFAR strategy is obtained, and then the distance dimension detection threshold T = αT of the detection unit can be obtained. VI .

[0141] Formula (6) is

[0142] Step 505: Determine the distance dimension detection result D based on the distance dimension detection threshold. R .

[0143] Combination Figure 6For each detection unit, after obtaining the range-dimensional detection threshold T, the data (power value) of that detection unit is compared with T. If the data of the detection unit is greater than T, it is determined that the detection unit has a target, and the value of the detection unit is set to 1; if the data of the detection unit is less than or equal to T, it is determined that the detection unit has no target, and the value of the detection unit is set to 0. The range-Doppler spectrum data matrix M is traversed along the range-dimensional direction. RDM All detection units in the array can yield the distance dimension detection result D. R The D R It is a matrix.

[0144] Step 506: Determine the Doppler detection sliding window of the detection unit.

[0145] The distance Doppler spectrum data matrix M is obtained according to step 501. RDM Then, along the Doppler direction, M RDM Each resolution unit serves as a detection unit in sequence, and can be based on Figure 4 The principle of determining the Doppler window of the detection unit as described in the corresponding embodiment is used to determine the Doppler window of the detection unit, which will not be repeated here.

[0146] Step 507: Remove interference clutter and determine the clutter state within the Doppler detection window and the mean ratio of the Doppler detection window.

[0147] Step 508: Determine the Doppler detection threshold of the detection unit based on the clutter state, mean ratio, and set constant false alarm rate.

[0148] The specific implementation process of steps 507 and 508 is similar to the processing process on the distance dimension, and can be referred to steps 503 and 504, which will not be repeated here.

[0149] Step 509: Determine the Doppler detection result D based on the Doppler detection threshold. D .

[0150] For each detection unit, after obtaining the Doppler detection threshold of the detection unit, the data (power value) of the detection unit is compared with its corresponding Doppler detection threshold. If it is greater than the Doppler detection threshold, it is determined that the detection unit has a target and the value of the detection unit is set to 1. Otherwise, it is determined that the detection unit has no target and the value of the detection unit is set to 0.

[0151] Traverse the distance-Doppler spectrum data matrix M obtained in step 501 RDM All detection units in the array can be used to obtain the distance Doppler spectrum data matrix M. RDM Doppler test results D D The DD It is a matrix.

[0152] Step 510: Detect the distance dimension D R And Doppler detection results D D Perform a logical AND operation to obtain the initial detection result.

[0153] For example, the distance dimension detection result D can be... R And Doppler detection results D D Each detection unit performs a logical AND operation to obtain the initial detection result of the target echo signal. This initial detection result is in matrix form and can be denoted as matrix D. If the value of a detection unit in the initial detection result is 1, it is preliminarily determined that there is a target at that location; if the value of a detection unit in the initial detection result is 0, it is preliminarily determined that there is no target at that location.

[0154] Step 511: Convert the initial detection results into detection results to be judged.

[0155] After obtaining the initial detection result matrix D, the coordinates of the positions with a value of 1 in the initial detection result matrix D can be found sequentially, and then compared with the distance Doppler spectrum data matrix M. RDM The power value corresponding to the coordinate is updated to 1 at the corresponding position in the matrix D, thus obtaining the detection result matrix Dx to be determined.

[0156] Step 512: Divide the detection result to be judged into a set number of data segments along the distance dimension, and obtain the set threshold corresponding to each data segment.

[0157] For example, the detection result matrix Dx can be divided into M (M is an integer greater than 1) uniform data segments along the distance dimension. Alternatively, it can be divided into M non-uniform data segments according to actual application requirements. The segmented data segments can be represented as R1, R2, ..., R... M It is understandable that dividing the detection result matrix Dx into its range dimension is equivalent to dividing the range-Doppler spectrum data matrix M. RDM The distance dimension was divided.

[0158] During the segmentation process, the rows of the detection result matrix Dx to be judged are also divided into a set number of distance segments. For each distance segment, the resolution unit occupied by its distance dimension is at least 2. Based on engineering experience and combined with the average background clutter power of different distance segments, different threshold values ​​can be set for each data segment, such as R1, R2, ..., R M The corresponding threshold values ​​are T1, T2, ..., T MFor example, for each range segment, the set threshold for that range segment can be set as a multiple of the average background clutter power of that range segment, such as 2 to 4 times. The average background clutter power of the range segment can be referred to the description in step 150, and will not be repeated here.

[0159] Step 513: Compare the data in each data segment with the set threshold corresponding to that data segment, and generate the target detection result of the target echo signal based on the comparison result.

[0160] The specific implementation process of step 513 can be referred to step 150 above, and will not be repeated here.

[0161] After generating the target echo signal, the target detection result can be used for subsequent processing such as multi-target clustering and tracking of traffic millimeter-wave radar.

[0162] The target detection method provided in this invention, on the one hand, segments the range-dimensional Doppler spectrum data of target echo signals from traffic millimeter-wave radar in the range dimension. For each range segment, a corresponding threshold is set based on the background clutter power of that range segment. Based on this threshold, further target determination is performed on the initial detection results. This method considers different background clutter environments at different range segments, enabling better fitting of the actual background clutter environment at different range segments in traffic application scenarios, effectively improving the background clutter estimation accuracy of traffic millimeter-wave radar. Furthermore, by further segmenting the adaptive constant false alarm rate (CFAR) detection results in the range and Doppler dimensions, false targets in the adaptive CFAR detection results are effectively suppressed, improving the target accuracy. On the one hand, the target detection rate is improved; on the other hand, adaptive constant false alarm rate (CFAR) detection is performed on both the range dimension and the Doppler dimension, and the target detection result of the target echo signal can be obtained by combining the detection results of the two dimensions. By comprehensively considering the information of the two dimensions, the target detection rate is further improved. Furthermore, in the adaptive CFAR detection process in the Doppler dimension direction, for the detection unit near the Doppler dimension boundary in the range-Doppler spectrum data, the resolution unit on the opposite side of the nearby Doppler dimension boundary can be used as a supplement to the missing reference unit based on the circular convolution characteristics of the velocity. This makes up for the deficiency of the reference unit and improves the accuracy of the background clutter threshold estimation in the Doppler dimension direction, thereby improving the robustness and detection rate of target detection.

[0163] The target detection method provided in this invention can effectively improve the target detection rate of traffic millimeter-wave radar while ensuring a constant false alarm rate, thus providing a foundation for subsequent target clustering and tracking.

[0164] The target detection device provided by the present invention is described below. The target detection device described below and the target detection method described above can be referred to in correspondence.

[0165] Figure 7 An exemplary schematic diagram of the target detection device provided in an embodiment of the present invention is shown, with reference to... Figure 7 As shown, the target detection device 700 may include: an acquisition module 710 for acquiring range-Doppler spectrum data of radar target echo signals; a detection module 720 for performing adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein the detection dimension includes the range dimension and / or the Doppler dimension; a determination module 730 for determining the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data in the initial detection result, and updating the value of the target resolution unit with the data to obtain a detection result to be determined; a segmentation module 740 for dividing the detection result to be determined into a set number of data segments along the range dimension; and a generation module 750 for comparing the data in each data segment with a set threshold corresponding to each data segment, and generating a target detection result of the target echo signal based on the comparison result.

[0166] In one example embodiment, the generation module 750 may be specifically used to: set the data in each data segment that is less than the set threshold corresponding to each data segment to 0, so as to update each data segment; and determine all the updated data segments as the target detection results of the target echo signal.

[0167] In one example embodiment, the detection module 720 may include: a first determining unit, configured to sequentially use each resolution unit in the range Doppler spectrum data as a detection unit, and determine the detection sliding window of the detection unit in the detection dimension, wherein the detection sliding window in the detection dimension includes the detection unit and the forward sliding window and the backward sliding window in the detection dimension where the detection unit is located; a second determining unit, configured to determine the clutter state within the detection sliding window and the mean ratio of the detection sliding window, and determine the detection threshold of the detection unit based on the clutter state, the mean ratio and a set constant false alarm rate, wherein the clutter state includes uniform clutter or non-uniform clutter; and a comparison unit, configured to compare the data of the detection unit with the detection threshold, and generate an initial detection result based on the comparison result.

[0168] In one example embodiment, the first determining unit may include: a first selection subunit, configured to continuously select a fifth preset number of resolving units forward from the forward adjacent unit of the detection unit along the distance dimension direction where the detection unit is located, to obtain a first protection unit; a second selection subunit, configured to continuously select a fifth preset number of resolving units backward from the backward adjacent unit of the detection unit, to obtain a second protection unit; a first determining subunit, configured to determine a third preset number of resolving units before the first protection unit and a third preset number of resolving units after the second protection unit as a forward reference unit and a backward reference unit, respectively; a second determining subunit, configured to determine the first protection unit and the forward reference unit as a forward sliding window of the detection unit in the distance dimension direction, and to determine the second protection unit and the backward reference unit as a backward sliding window of the detection unit in the distance dimension direction; and a third determining subunit, configured to determine the forward sliding window in the distance dimension direction, the detection unit, and the backward sliding window in the distance dimension direction as a detection sliding window in the distance dimension.

[0169] In one example embodiment, the first determining unit may further include: a judging subunit, configured to judge, along the Doppler direction where the detection unit is located, whether there is a fourth preset number of judging units in the forward and backward resolving units of the detection unit; a fourth determining subunit, configured to, when there is no fourth preset number of judging units in the forward direction of the detection unit, determine all forward judging units and the first supplementary judging units of the detection unit as forward sliding windows of the detection unit in the Doppler direction, and determine the fourth preset number of backward judging units of the detection unit as backward sliding windows of the detection unit in the Doppler direction, wherein the first supplementary judging units are a first number of judging units starting from the backward boundary judging units of the detection unit, and the total number of all forward judging units and the first supplementary judging units of the detection unit is the fourth preset number; a fifth determining subunit, configured to, when there is no fourth preset number of judging units in the backward direction of the detection unit... In the case of a detection unit, all backward resolution units and the second supplementary resolution unit of the detection unit are determined as backward sliding windows of the detection unit in the Doppler direction, and a fourth preset number of forward resolution units of the detection unit are determined as forward sliding windows of the detection unit in the Doppler direction. The second supplementary resolution unit is a second number of resolution units starting from the forward boundary resolution unit of the detection unit, and the total number of all backward resolution units and the second supplementary resolution units of the detection unit is the fourth preset number. The third selection subunit is used to select a fourth preset number of resolution units before and after the detection unit when there are a fourth preset number of resolution units in both the forward and backward directions of the detection unit, so as to obtain forward sliding windows and backward sliding windows of the detection unit in the Doppler direction. The sixth determination subunit is used to determine the forward sliding window in the Doppler direction, the detection unit, and the backward sliding window in the Doppler direction as a detection sliding window in the Doppler direction.

[0170] In one example embodiment, the second determining unit may include: a sorting subunit, used to sort the data of all reference units within the detection window of each detection dimension for each detection dimension, to obtain a sorting result; a removal subunit, used to remove the first first preset number and the last second preset number of data in the sorting result, to obtain valid reference unit data; and a seventh determining subunit, used to determine the clutter state within the detection window of each detection dimension and the mean ratio of the detection window of each detection dimension based on the valid reference unit data.

[0171] In one example embodiment, the seventh determining subunit may be specifically used to: determine a first change index and a first power mean based on the data corresponding to the forward sliding window in the effective reference unit data; determine a second change index and a second power mean based on the data corresponding to the backward sliding window in the effective reference unit data; determine the clutter state in the forward sliding window according to the first change index, and determine the clutter state in the backward sliding window according to the second change index; calculate the ratio of the first power mean and the second power mean to obtain the mean ratio of the detection sliding windows for each detection dimension.

[0172] Figure 8 An example is a schematic diagram of the structure of an electronic device, such as... Figure 8 As shown, the electronic device 800 may include: a processor 810, a communication interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call logic instructions in the memory 830 to execute the target detection method provided in the above-described method embodiments. This method may include, for example,: acquiring range-Doppler spectrum data of radar target echo signals; performing adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein the detection dimension includes the range dimension and / or the Doppler dimension; determining the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data in the initial detection result, and updating the value of the target resolution unit with the data to obtain a detection result to be determined; dividing the detection result to be determined into a set number of data segments along the range dimension; comparing the data in each data segment with a set threshold corresponding to each data segment, and generating a target detection result for the target echo signal based on the comparison result; wherein the set threshold is determined based on the background clutter power of each corresponding data segment.

[0173] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0174] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the target detection method provided in the above-described method embodiments. This method may include, for example, acquiring range-Doppler spectrum data of radar target echo signals; performing adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein the detection dimension includes the range dimension and / or the Doppler dimension; determining the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data in the initial detection result, and updating the value of the target resolution unit with the data to obtain a detection result to be determined; dividing the detection result to be determined into a set number of data segments along the range dimension; comparing the data in each data segment with a set threshold corresponding to each data segment, and generating a target detection result of the target echo signal based on the comparison result, wherein the set threshold is determined based on the background clutter power of each corresponding data segment.

[0175] In another aspect, the present invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the target detection method provided in the above-described method embodiments. This method may include, for example,: acquiring range-Doppler spectrum data of radar target echo signals; performing adaptive constant false alarm rate (CFAR) detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein the detection dimension includes the range dimension and / or the Doppler dimension; determining the data corresponding to the target resolution unit representing the target in the range-Doppler spectrum data in the initial detection result, and updating the value of the target resolution unit with the data to obtain a detection result to be determined; dividing the detection result to be determined into a predetermined number of data segments along the range dimension; comparing the data in each data segment with a predetermined threshold corresponding to each data segment, and generating a target detection result for the target echo signal based on the comparison result, wherein the predetermined threshold is determined based on the background clutter power of each corresponding data segment.

[0176] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0177] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A target detection method, characterized in that, include: Acquire range-Doppler spectrum data of radar target echo signals; Adaptive constant false alarm rate (CFAR) detection is performed on the range-Doppler spectrum data in the detection dimension to obtain initial detection results; wherein, the detection dimension includes the range dimension and / or the Doppler dimension; The target resolution unit representing the target in the initial detection result is determined to correspond to the data in the range Doppler spectrum data, and the value of the target resolution unit is updated with the data to obtain the detection result to be determined. The detection result to be determined is divided into a set number of data segments along the distance dimension; The data in each data segment is compared with a set threshold corresponding to each data segment, and a target detection result of the target echo signal is generated based on the comparison result. The set threshold is determined based on the background clutter power of the corresponding data segment. The adaptive constant false alarm rate (CFAR) detection on the detection dimension of the range Doppler spectrum data is performed to obtain the initial detection result, including: Each resolution unit in the range-Doppler spectrum data is sequentially used as a detection unit, and a detection sliding window for the detection unit is determined on the range dimension detection dimension and / or the Doppler dimension detection dimension, wherein the detection sliding window on the detection dimension includes the detection unit and the forward sliding window and the backward sliding window on the detection dimension where the detection unit is located; The initial detection result is determined based on the detection sliding window; The step of determining the detection sliding window of the detection unit on the distance dimension detection dimension includes: Along the distance dimension of the detection unit, starting from the forward adjacent unit of the detection unit, a fifth preset number of resolution units are continuously selected to obtain the first protection unit; Starting from the adjacent unit of the detection unit, the fifth preset number of resolution units are selected sequentially to obtain the second protection unit; The third preset number of resolution units before the first protection unit and the third preset number of resolution units after the second protection unit are respectively defined as forward reference units and backward reference units; The first protection unit and the forward reference unit are defined as the forward sliding window of the detection unit in the distance dimension direction, and the second protection unit and the backward reference unit are defined as the backward sliding window of the detection unit in the distance dimension direction. The forward sliding window in the distance dimension, the detection unit, and the backward sliding window in the distance dimension are defined as the detection sliding window in the distance dimension.

2. The target detection method according to claim 1, characterized in that, The step of comparing the data in each data segment with a set threshold corresponding to each data segment, and generating a target detection result of the target echo signal based on the comparison result, includes: In each data segment, data that is less than a predetermined threshold corresponding to that data segment is set to 0 in order to update each data segment; All updated data segments are identified as the target detection results of the target echo signal.

3. The target detection method according to claim 1, characterized in that, Determining the initial detection result based on the detection sliding window includes: The clutter state within the detection window and the mean ratio of the detection window are determined, and the detection threshold of the detection unit is determined based on the clutter state, the mean ratio, and a set constant false alarm rate, wherein the clutter state includes uniform clutter or non-uniform clutter. The data from the detection unit is compared with the detection threshold, and the initial detection result is generated based on the comparison result.

4. The target detection method according to claim 1, characterized in that, The step of determining the detection sliding window of the detection unit in the Doppler detection dimension includes: Along the Doppler direction where the detection unit is located, determine whether there is a fourth preset number of resolution units in the forward resolution unit and the backward resolution unit of the detection unit; If there are no fourth preset number of resolution units in the forward direction of the detection unit, all forward resolution units of the detection unit and the first supplementary resolution unit are determined as forward sliding windows of the detection unit in the Doppler direction, and the fourth preset number of backward resolution units of the detection unit are determined as backward sliding windows of the detection unit in the Doppler direction. The first supplementary resolution unit is a first number of resolution units starting from the backward boundary resolution unit of the detection unit, and the total number of all forward resolution units and the first supplementary resolution unit of the detection unit is the fourth preset number. If the fourth preset number of resolution units does not exist in the backward direction of the detection unit, all backward resolution units of the detection unit and the second supplementary resolution units are determined as the backward sliding window of the detection unit in the Doppler direction, and the fourth preset number of resolution units in the forward direction of the detection unit are determined as the forward sliding window of the detection unit in the Doppler direction. The second supplementary resolution unit is a second number of resolution units starting from the forward boundary resolution unit of the detection unit, and the total number of all backward resolution units and the second supplementary resolution units of the detection unit is the fourth preset number. When the fourth preset number of resolution units exist in both the forward and backward directions of the detection unit, the fourth preset number of resolution units are selected before and after the detection unit to obtain the forward sliding window and the backward sliding window of the detection unit in the Doppler direction. The forward sliding window in the Doppler direction, the detection unit, and the backward sliding window in the Doppler direction are defined as the detection sliding window in the Doppler direction.

5. The target detection method according to claim 3 or 4, characterized in that, Determining the ratio of the clutter state within the detection window to the mean of the detection window includes: For each of the detection dimensions, the data of all reference units within the detection sliding window of each detection dimension are sorted to obtain the sorting result; Remove the first and second preset quantities of data from the sorting results to obtain valid reference unit data; Based on the effective reference unit data, determine the clutter state within the detection window of each detection dimension and the mean ratio of the detection windows of each detection dimension.

6. The target detection method according to claim 5, characterized in that, The step of determining the clutter state within the detection window of each detection dimension and the mean ratio of the detection windows of each detection dimension based on the effective reference unit data includes: Based on the data corresponding to the forward sliding window in the effective reference unit data, a first change index and a first power mean are determined; Based on the data corresponding to the backward sliding window in the effective reference unit data, the second change index and the second power mean are determined; The clutter state in the forward sliding window is determined according to the first change index, and the clutter state in the backward sliding window is determined according to the second change index. Calculate the ratio of the first power mean to the second power mean to obtain the mean ratio of the detection sliding window for each detection dimension.

7. A target detection device, characterized in that, The method for implementing any one of claims 1-6 includes: The acquisition module is used to acquire range Doppler spectrum data of radar target echo signals; The detection module is used to perform adaptive constant false alarm rate detection on the range-Doppler spectrum data in the detection dimension to obtain an initial detection result; wherein, the detection dimension includes the range dimension and / or the Doppler dimension; The determination module is used to determine the data corresponding to the target resolution unit representing the target in the initial detection result in the range Doppler spectrum data, and update the value of the target resolution unit with the data to obtain the detection result to be determined; The segmentation module is used to divide the detection result to be determined into a set number of data segments along the distance dimension; The generation module is used to compare the data in each data segment with the set threshold corresponding to each data segment, and generate the target detection result of the target echo signal based on the comparison result.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the target detection method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the target detection method as described in any one of claims 1 to 6.

Images