Clutter detection method, radar device, and computer-readable storage medium

Abstract

The application relates to the technical field of radar signal processing, and discloses a clutter detection method, a radar device and a computer readable storage medium. The method is applied to a radar device and comprises the following steps: acquiring radar detection data, processing the radar detection data, generating a first radar observation graph, analyzing the energy of each position in the first radar observation graph, determining the energy distribution characteristics of the first radar observation graph, determining a target clutter type according to the energy distribution characteristics, searching a preset clutter database according to the target clutter type, acquiring a target clutter window corresponding to the target clutter type, wherein the clutter database stores a plurality of clutter types and a clutter window corresponding to each clutter type, and the target clutter window is used for analyzing the first radar observation graph to generate a target clutter graph. Through the above method, the edge of the clutter is accurately and finely described.

Description

Technical Field

[0001] This application relates to the field of radar signal processing technology, specifically to a clutter detection method, radar equipment, and computer-readable storage medium. Background Technology

[0002] When processing radar signals, radar clutter can affect the detection and tracking of real targets. It is usually necessary to identify clutter areas from radar observation maps (e.g., range-Doppler maps, range-azimuth maps, range-elevation maps, etc.) to improve the accuracy of real target identification.

[0003] Currently, the main approach relies on statistical analysis to determine threshold values, and then uses these threshold values ​​to detect targets on radar images in order to identify real targets. However, this method focuses more on clutter suppression than clutter edge localization, resulting in poor accuracy in clutter edge localization, which can easily affect the identification of real targets, especially those near clutter edges, which are prone to being missed. Summary of the Invention

[0004] In view of the above problems, embodiments of this application provide a clutter detection method, radar equipment, and computer-readable storage medium to solve the problem in the prior art that poor clutter edge positioning accuracy can easily affect the identification of real targets.

[0005] According to one aspect of the embodiments of this application, a clutter detection method is provided, applied to a radar device. The method includes: acquiring radar detection data and processing the radar detection data to generate a first radar observation map; analyzing the energy at each location in the first radar observation map to determine the energy distribution characteristics of the first radar observation map, and determining the target clutter type based on the energy distribution characteristics; searching a preset clutter database according to the target clutter type to obtain a target clutter window corresponding to the target clutter type, wherein the clutter database stores multiple clutter types and clutter windows corresponding to each clutter type; and analyzing the first radar observation map using the target clutter window to generate a target clutter map.

[0006] In one optional approach, the first radar observation map includes a range dimension and an observation dimension. Analyzing the energy at each location in the first radar observation map to determine its energy distribution characteristics specifically includes: sliding a preset window along the observation dimension and performing sliding window calculations on the energy at each location in the first radar observation map to determine the energy threshold corresponding to each distance value in the range dimension; analyzing the energy at each location in the first radar observation map based on the energy threshold corresponding to each distance value in the range dimension to determine multiple detection points in the first radar observation map; and analyzing the locations of the multiple detection points to determine the energy distribution characteristics of the first radar observation map.

[0007] In one optional approach, a preset window is slid along the observation dimension, and a sliding window calculation is performed on the energy at each location in the first radar observation image to determine the energy threshold corresponding to each distance value in the range dimension. Specifically, this includes: sliding the preset window along the observation dimension, and performing a sliding window calculation on the energy at each location in the first radar observation image to obtain the average energy value corresponding to each distance value in the range dimension; and determining the energy threshold corresponding to each distance value based on the preset threshold and the average energy value corresponding to each distance value.

[0008] In one alternative approach, the energy at each location in the first radar observation map is analyzed based on the energy threshold corresponding to each distance value in the range dimension to determine multiple detection points in the first radar observation map. Specifically, this includes: performing peak detection along the range dimension on the first radar observation map to obtain multiple peak points corresponding to each distance value in the range dimension; comparing the peak values ​​of the multiple peak points corresponding to each distance value with the energy threshold; and if the peak value of a peak point is greater than the energy threshold, then the peak point is determined as a detection point.

[0009] In one alternative approach, the locations of multiple detection points are analyzed to determine the energy distribution characteristics of the first radar observation map. Specifically, this includes: counting the number of detection points corresponding to each observation value in the observation dimension, generating a histogram corresponding to each detection point; and analyzing the peak regions in the histogram to determine the energy distribution characteristics of the first radar observation map.

[0010] In one alternative approach, a target clutter window is used to analyze the first radar observation map to generate a target clutter map. Specifically, this includes: aligning the center point of the target clutter window with each detection point, calculating the clutter level corresponding to each position in the first radar observation map according to the target clutter window, and generating an initial clutter map; and performing dilation processing on the initial clutter map to generate a target clutter map.

[0011] In one optional embodiment, the method further includes: acquiring clutter window sizes corresponding to multiple clutter types; acquiring the operating parameters of the radar equipment and the environmental parameters of the detection area, and determining the range data and observation data corresponding to each clutter type based on the operating parameters and environmental parameters; generating clutter windows corresponding to each clutter type based on the clutter window sizes, range data, observation data, and a preset weighting formula; and storing the clutter windows corresponding to each clutter type in a clutter database.

[0012] In one optional approach, clutter window sizes corresponding to multiple clutter types are obtained, specifically including: acquiring radar detection data corresponding to each clutter type, and generating a second radar observation map corresponding to each clutter type based on the radar detection data corresponding to each clutter radar; for each second radar observation map corresponding to a clutter type, the following steps are performed: analyzing the energy at each location in the second radar observation map based on a first preset energy threshold to determine scatterer characteristic information; dividing the second radar observation map into a real target area and a clutter target area based on a second preset energy threshold; analyzing the energy at each location in the real target area based on a third preset energy threshold to determine the distribution density of the real target; analyzing the energy at each location in the clutter target area based on a fourth preset energy threshold to determine the distribution density of the clutter target; and determining the clutter window size corresponding to the clutter type based on the scatterer characteristic information, the distribution density of the real target, and the distribution density of the clutter target.

[0013] According to another aspect of the embodiments of this application, a radar device is provided, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the clutter detection method described in any of the above claims.

[0014] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, implements the clutter detection method described in any of the preceding claims.

[0015] This application embodiment analyzes a first radar observation image to determine the clutter type corresponding to the first radar observation image based on its energy distribution characteristics and the physical properties of different clutter types. A corresponding clutter window is then obtained based on the clutter type to achieve precise and accurate characterization of clutter edges. In CFAR processing, the selection of reference cells is adjusted based on accurate clutter edge selection, effectively avoiding the signal-to-noise ratio reduction caused by clutter energy falling into the reference window, and significantly improving the detection rate of true targets at clutter edges. Simultaneously, the use of a sparse clutter window structure effectively reduces the amount of iterative computation, meeting the real-time requirements of engineering implementation.

[0016] The above description is merely an overview of the technical solutions of the embodiments of this application. In order to better understand the technical means of the embodiments of this application and to implement them in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the embodiments of this application more obvious and understandable, specific implementation methods of this application are described below. Attached Figure Description

[0017] The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 A schematic flowchart of the clutter detection method provided in an embodiment of this application is shown; Figure 2a This illustration shows a first structural schematic diagram of a first radar observation map provided in an embodiment of this application; Figure 2b This paper shows a second structural schematic diagram of the first radar observation map provided in an embodiment of this application; Figure 2c This illustration shows a third structural schematic diagram of the first radar observation map provided in an embodiment of this application; Figure 3a A schematic diagram of the structure of the first clutter window provided in an embodiment of this application is shown; Figure 3b A schematic diagram of the structure of the second clutter window provided in an embodiment of this application is shown; Figure 3c A schematic diagram of the structure of the third clutter window provided in the embodiments of this application is shown; Figure 4a This illustration shows a schematic diagram of the structure of the first type of target clutter map provided in an embodiment of this application; Figure 4b This illustration shows a schematic diagram of the structure of the second type of target clutter map provided in an embodiment of this application; Figure 4c This illustration shows a structural schematic diagram of the third type of target clutter map provided in an embodiment of this application; Figure 5 A flowchart illustrating a clutter detection method according to another embodiment of this application is shown; Figure 6 A schematic diagram of the structure of the radar device provided in an embodiment of this application is shown. Detailed Implementation

[0018] Exemplary embodiments of the present application will now be described in more detail with reference to the accompanying drawings. Although exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application may be implemented in various forms and should not be limited to the embodiments set forth herein.

[0019] Radar clutter refers to radar echoes received by radar equipment from non-real targets such as the ground, sea surface, rain, snow, and birds. This radar clutter can affect the detection and tracking of real targets, and in severe cases, it can overwhelm real targets or generate a large number of false alarms. Radar echoes from surface or volume targets such as the ground, sea surface, rain, and snow will form clutter areas on the radar observation map. The clutter edge refers to the area where the energy of the clutter area attenuates and transitions to background noise.

[0020] When performing constant false alarm rate (CFAR) detection on real targets located near clutter edges, clutter energy falls into the CFAR reference window, causing a decrease in the signal-to-noise ratio of the real target and resulting in missed detections. Therefore, accurately locating clutter edges can effectively improve the detection rate of real targets at clutter edges.

[0021] Currently, the main approach is to statistically analyze the noise energy of each range cell in the radar observation map, calculate the detection threshold for each range cell, and then use this threshold to detect targets within the radar's detection area. However, this detection threshold focuses on suppressing clutter energy rather than locating clutter edges. Consequently, the accuracy of clutter edge localization is poor when identifying real targets, making it difficult to accurately distinguish real targets at the edges of clutter.

[0022] Based on this, to improve the accuracy of clutter edge localization, this application provides a clutter detection method. First, by analyzing the distribution characteristics of different types of clutter (e.g., ground clutter, rain / snow clutter, sea clutter, etc.), clutter windows corresponding to each clutter type are pre-generated. When clutter detection is required, the energy distribution characteristics of the radar observation image are determined by analysis, and the clutter type is determined based on these characteristics. Then, a corresponding target clutter window is selected from the pre-generated clutter windows according to the clutter type. Finally, the target clutter window is used to analyze the radar observation image to generate a clutter map, through which clutter edges are obtained.

[0023] In this approach, by determining the clutter type and obtaining the corresponding clutter window based on that type, the radar observation image can be accurately analyzed using the clutter window to generate a precise clutter map, thereby depicting a more refined and accurate clutter edge. When performing CFAR on real targets near the clutter edge, the noise calculation method can be adjusted based on the clutter edge to prevent clutter energy from falling into the reference window, thus improving the accuracy of the real target signal-to-noise ratio and avoiding missed detections.

[0024] According to a first aspect of the embodiments of this application, a clutter detection method is provided, such as... Figure 1 As shown, Figure 1 A flowchart of a clutter detection method provided in an embodiment of this application is shown. This method is applied to and executed by a radar device. A radar device is a hardware device capable of transmitting and receiving electromagnetic waves and acquiring target echoes. It typically includes an antenna, transmitter, receiver, and corresponding radio frequency and timing modules. It is the physical basis for radar systems to perform target detection (such as CFAR) and parameter measurements (such as range, velocity, and angle). Figure 1 As shown, the method includes the following steps: Step S110: Acquire radar detection data and process the radar detection data to generate the first radar observation map.

[0025] Radar detection data consists of the raw echo signals received by the radar equipment's antenna. These signals contain electromagnetic wave information emitted by various scattering objects in space (such as drones, vehicles, the ground, rain, snow, and other objects capable of emitting electromagnetic waves). This data serves as the fundamental data source for subsequent signal processing and target detection by the radar equipment. The radar equipment can perform analog-to-digital conversion through its radio frequency front-end and data acquisition system, transforming the received analog echo signals into digital signals for subsequent calculations by a digital signal processor.

[0026] The first radar observation map is a two-dimensional feature map generated by the radar equipment based on radar detection data. Specifically, it can be a range-Doppler map, a range-azimuth map, a range-elevation map, etc. After acquiring the radar detection data, the radar equipment can sequentially perform preprocessing operations on the raw echo, such as pulse compression, moving target indication (MTI), and fast Fourier transform (FFT). Taking the range-Doppler map as an example, the radar equipment can convert its acquired time-domain signal into a two-dimensional frequency domain with range and Doppler dimensions, thereby enabling the differentiation of targets at different distances and speeds.

[0027] The first radar observation map includes the range dimension and the observation dimension (i.e., Doppler dimension, angle dimension, etc.), which specifically reflect the two-dimensional distribution of the target in the range dimension and the observation dimension. The range dimension is the dimension in the radar echo signal that reflects the continuous spatial distance between the target and the radar equipment. Its resolution is determined by the bandwidth of the radar transmitted signal. The observation dimension is the dimension used by the radar equipment to distinguish different observation angles or velocities when performing spatial scanning or signal processing.

[0028] Taking the first radar observation image as a range-Doppler image as an example, such as Figure 2a As shown, Figure 2a This diagram illustrates the structure of the first radar observation map corresponding to the first type of clutter. The horizontal axis represents Doppler dimensions, and the vertical axis represents the range dimension. The first radar observation map reflects the two-dimensional distribution of targets in both the range and Doppler dimensions (i.e., the velocity dimension). The coordinates of different targets on the map are shown below. Energy is . This reflects the distance between the target and the radar equipment. This reflects the radial relative velocity between the target and the radar equipment. To represent the strength of the target energy, different colors in the diagram indicate different energy intensities. Figure 2a For example, in the diagram, the energy in the yellow area is stronger than that in the green area, and the energy in the green area is stronger than that in the blue area.

[0029] As an example, after the radar equipment is powered on, it can continuously collect spatial echo signals and sample the data of each pulse repetition cycle through its signal processing module. Then, pulse compression is performed along the range dimension to improve range resolution, and FFT is performed along the observation dimension to extract observation information (such as velocity information, angle information, etc.). Finally, a two-dimensional matrix containing range information, observation information, and energy information (i.e., the first radar observation data) is output.

[0030] This allows the raw echoes acquired by radar equipment to be transformed into intuitive two-dimensional images, revealing the spectral broadening characteristics of clutter in both the range and observation dimensions. This provides a data foundation for subsequent clutter classification and edge localization based on physical distribution characteristics.

[0031] Step S120: Analyze the energy at each location in the first radar observation map, determine the energy distribution characteristics of the first radar observation map, and determine the target clutter type based on the energy distribution characteristics.

[0032] Specifically, the energy distribution characteristics refer to the clustering state and histogram statistical characteristics of high-energy points in the range and observation dimensions in the first radar observation map. These characteristics reflect the spatial distribution pattern of clutter scatterers and are the key basis for distinguishing different clutter types.

[0033] Specifically, radar equipment can identify and extract high-energy points from a first radar observation map, and then perform statistical analysis on the extracted high-energy points to obtain the energy distribution pattern on the first radar observation map. For example, radar equipment can traverse the energy at each location on the first radar observation map and compare the traversed energy with a preset threshold. If the energy is greater than the preset threshold, the location corresponding to that energy is determined as a high-energy point.

[0034] Clutter types can include ground clutter, rain / snow clutter, or sea clutter, among others, with different physical scattering characteristics. Different clutter types exhibit significant differences in broadening and distribution patterns across the range and observation dimensions. Figure 2a The ground clutter shown and Figure 2b Taking the rain and snow clutter shown as an example, the observation dimension in the figure is the Doppler dimension (i.e., the velocity dimension).

[0035] like Figure 2a As shown, ground clutter refers to radar echoes received by radar equipment from various ground scattering objects. There are numerous stationary surface and volume targets on the ground at different distances. When the radar equipment is stationary, the relative velocity between the targets and the radar equipment is zero, meaning all ground clutter has the same velocity. On the first radar observation map, the high-energy points corresponding to the target will be concentrated near zero velocity and distributed along the range dimension. When the radar equipment moves, the relative velocity between the stationary target on the ground and the radar equipment is the radial projection of the radar equipment's moving velocity (i.e., radial velocity). On the first radar observation map, the high-energy points corresponding to the target will be concentrated near a certain velocity value and distributed along the range dimension.

[0036] like Figure 2b As shown, rain and snow clutter refers to the radar echoes reflected by falling raindrops and snowflakes received by radar equipment. Raindrops and snowflakes are volume targets with a wide spatial distribution, thus distributed across multiple cells along the range dimension. Furthermore, raindrops and snowflakes possess a certain velocity during their fall, generating a radial velocity component relative to the radar equipment. The actual falling velocity of rain and snow depends on factors such as particle size, shape, wind speed, and turbulence. Taking ordinary rainfall as an example, raindrops falling at different angles will produce different radial components. Therefore, high-energy points of raindrop targets will be distributed along the range dimension in the first radar observation map, spanning multiple consecutive velocity values.

[0037] Sea clutter refers to radar echoes received by radar equipment that are reflected from sea surface waves. Sea clutter is clutter caused by surface targets, and the characteristics of waves, white waves, and breaking waves are not entirely the same. Taking nearshore waves as an example, such as... Figure 2c As shown, Figure 2c The diagram illustrates a third structural representation of the first radar observation image. Waves spread towards the shore in a wavy pattern at a uniform velocity, exhibiting crests and troughs. The distribution is relatively uniform. This wavy wave clutter is similar in distribution to ground clutter, primarily along the range dimension. Furthermore, high-energy points of the sea clutter may cluster near specific velocity values, forming multiple high-energy clusters. Additionally, when radar equipment illuminates the sea surface, the undulating waves produce specular reflection, resulting in relatively high energy levels for the sea clutter.

[0038] Therefore, radar equipment can determine the type of target clutter based on the energy distribution characteristics of the first radar observation map. Taking the first radar observation map as a range-Doppler image as an example, after analyzing the energy at each location on the first radar observation map, if high-energy points are found to be clustered near a certain velocity value in Doppler, the target clutter type can be determined to be ground clutter; if the clustering area of ​​high-energy points is found to span multiple consecutive velocity values, the target clutter type can be determined to be rain / snow clutter; if high-energy points are found to be clustered near several velocity values, forming multiple clustering areas, the target clutter type can be determined to be sea clutter.

[0039] Furthermore, in order to improve the accuracy of energy distribution characteristics, the radar equipment can also set a corresponding threshold according to the actual situation of the first radar observation map. Specifically, step S120 also includes the following steps (steps S121 to S123): Step S121: Slide the preset window along the observation dimension and perform sliding window calculation on the energy at each position in the first radar observation map to determine the energy threshold corresponding to each distance value in the distance dimension.

[0040] Step S122: Analyze the energy at each location in the first radar observation map based on the energy threshold corresponding to each distance value in the distance dimension, and determine multiple detection points in the first radar observation map.

[0041] Step S123: Analyze the locations of multiple detection points to determine the energy distribution characteristics of the first radar observation map.

[0042] The preset window is a rectangle or weighted calculation window of a specific length. The preset window is translated along the direction of the observation dimension by a preset step size. During the sliding of the preset window, the energy values ​​of each position contained in the current window are calculated (such as statistical averaging or weighted summation), and the energy threshold is determined based on the calculation results.

[0043] like Figure 2a As shown, assuming the translation step size of the preset window is 1, the preset window can be translated unit by unit along the direction of the observation dimension (i.e., the horizontal direction), and the energy value of each position covered by the current window can be read sequentially. The energy level of the region can be characterized by the statistics of the local data within the sliding window. Each time the sliding window moves one unit, a local statistical mean is output. Thus, after traversing each row of observation dimension data sequence, the energy benchmark of each row of observation dimension data sequence can be extracted, that is, the energy threshold corresponding to each distance value.

[0044] After determining the energy threshold, the energy value at each coordinate position in the first radar observation map is compared one by one with the energy threshold corresponding to the distance value at that position. Points with low noise below the energy threshold are filtered out, and high-energy points exceeding the energy threshold are retained. The high-energy points retained after energy threshold comparison and screening are the multiple detection points in the first radar observation map. These detection points represent the energy peak positions of suspected clutter targets.

[0045] Next, based on the coordinate information of all detection points in the distance and observation dimensions, the distribution patterns of these points in the two-dimensional plane are statistically analyzed. For example, histograms are used to reveal the clustering patterns of high-energy points, and different clustering patterns correspond to clutter distributions with different physical characteristics. As an example, from... Figure 2a The first radar observation map shown clearly indicates that high-energy points of ground clutter are typically concentrated near a certain observation value in the observation dimension and span multiple distance values; from Figure 2bAs can be clearly seen from the first radar observation map, the high-energy points of rain and snow clutter are usually distributed in a two-dimensional diffusion pattern. Therefore, the energy distribution characteristics of the first radar observation map can reflect the spatial distribution pattern of clutter scatterers and provide an accurate basis for subsequent identification of target clutter types.

[0046] As an example, the radar equipment can perform sliding window mean calculation on the first radar observation map along the Doppler dimension (i.e., the observation dimension) to obtain the row noise floor corresponding to each range value. The row noise floor is then superimposed with a preset fixed value to obtain the energy threshold of each range gate. Subsequently, the energy of each point in the first radar observation map is compared with the energy threshold of its corresponding range value, and coarse search points (i.e., detection points) that exceed the energy threshold are selected. Finally, the histogram of these coarse search points in the Doppler dimension is calculated. If it is found that these coarse search points are mainly clustered near a certain velocity, the current energy distribution characteristics can be determined to be the energy distribution characteristics of ground clutter.

[0047] Through steps S121 to S123, the radar equipment can calculate along the observation dimension sliding window and dynamically determine the energy threshold of each range gate, realizing adaptive and accurate tracking of noise floor fluctuations, avoiding false alarms or missed detections caused by non-uniform backgrounds. Based on the dynamic threshold, the detection points are extracted and the positional distribution characteristics of the detection points are analyzed, which can accurately reveal the physical aggregation law of clutter, thereby providing reliable feature data for subsequent accurate determination of clutter type and selection of clutter window, and further assisting in depicting a more refined and accurate clutter edge.

[0048] After step S120, step S130 is executed: search the preset clutter database according to the target clutter type, and obtain the target clutter window corresponding to the target clutter type. The clutter database stores multiple clutter types and clutter windows corresponding to each clutter type.

[0049] The clutter database is a dataset stored inside the radar equipment, which pre-stores various clutter windows generated based on the actual physical distribution characteristics of clutter. Each clutter window is a template of a small clutter region composed of several clutter targets. The selection of the clutter window also determines the direction and range of clutter edge localization matching. The size of the clutter window determines the matching range, while the shape and weight of the clutter window determine the matching direction.

[0050] by Figures 3a to 3c Taking the clutter window shown as an example, Figures 3a to 3c The clutter windows corresponding to the three clutter types are shown respectively. Figure 3a This is the clutter window corresponding to ground clutter. Figure 3b For clutter windows corresponding to rain and snow clutter, Figure 3c The clutter window represents the clutter pattern of the sea clutter, with different colors in the figure representing different weights.

[0051] like Figure 2aAs shown, considering the characteristics of ground clutter mainly distributed along the range dimension and the potential presence of micro-Doppler, the clutter window is mainly distributed along the range dimension, and the scattering characteristics of different ground scatterers also vary significantly. Taking into account the main lobe broadening and energy leakage of clutter targets, such as... Figure 3a As shown, a certain weight variation can be set in the clutter window corresponding to the ground clutter, and the weight can be set to gradually increase from the center to the outer periphery, that is, the weight of the yellow area in the figure is greater than the weight of the green area, so as to suppress the main lobe broadening and energy leakage of the clutter target.

[0052] like Figure 2b As shown, raindrops and snowflakes have a wide spatial distribution and are distributed in two dimensions along the distance and Doppler dimensions. Regarding the two-dimensional distribution characteristics of rain and snow clutter, such as... Figure 3b As shown, the clutter window corresponding to rain and snow clutter is designed as a sparse clutter window with a diffuse distribution. Since rain and snow clutter are volume targets, the scattering characteristics of different positions are the same, so the same weights can be set in the clutter window.

[0053] like Figure 2c As shown, the distribution of wavy ocean waves is similar to that of ground clutter, mainly distributed along the range dimension. However, the types and distribution of ground targets are relatively more complex, while the scattering characteristics of ocean clutter are similar at different locations. Regarding the distribution characteristics of ocean clutter, such as... Figure 3c As shown, the clutter window corresponding to sea clutter is designed as a clutter window distributed along the distance dimension, and the weights are smoother than those of the ground clutter window.

[0054] Clutter windows can be pre-generated by relevant technicians through analysis of the distribution characteristics of various clutter types, or they can be automatically generated by radar equipment after analyzing its operating parameters and operating environment, or after collecting and analyzing multiple sets of radar observation images.

[0055] After determining the clutter type corresponding to the first radar observation map, the radar equipment can search the database to read the corresponding clutter window (i.e., parameters such as the size, shape, and weight distribution of the clutter window) based on the identified clutter type index.

[0056] Step S140: Analyze the first radar observation map using the target clutter window to generate a target clutter map.

[0057] The target clutter window is a two-dimensional distribution map that identifies the clutter region's extent, clutter type, and clutter level. It visually represents the clutter edge where energy attenuates and transitions to background noise. For example... Figures 4a to 4c As shown, Figures 4a to 4c The diagram shows three different target clutter maps, with different colors representing different clutter levels. Based on the clutter level, the clutter edge can be accurately located. Figure 4a for Figure 2aThe target clutter map corresponding to the first radar observation map shown is as follows: Figure 4b for Figure 2b The target clutter map corresponding to the first radar observation map shown is as follows: Figure 4c for Figure 2c The target clutter map corresponding to the first radar observation map shown.

[0058] The radar equipment can use a target clutter window to perform weighted matching and calculation on the first radar observation map to calculate the clutter level at each location in the first radar observation map, thereby generating a target clutter map. As an example, step S140 may include the following steps (steps S141 to S142): Step S141: Align the center point of the target clutter window with each detection point, calculate the clutter level corresponding to each position in the first radar observation map according to the target clutter window, and generate an initial clutter map.

[0059] Step S142: Dilate the initial clutter map to generate the target clutter map.

[0060] The center point of the target clutter window is its geometric center or weighted center, serving as the coordinate reference origin during the alignment and matching process. It represents the position of the clutter target within the clutter region. The center point of the target clutter region is used for spatial coordinate alignment with the detection point to be detected. This alignment ensures that the shape and weight distribution of the target clutter window accurately cover the detection point's area.

[0061] Clutter level is a quantitative indicator used to measure the strength and density of clutter energy at a certain location. Its value reflects the confidence that the location belongs to a clutter zone and is a core parameter constituting the pixel map of the clutter image. Specifically, each non-zero weight within the target clutter window is weighted and calculated with the energy value of the corresponding location in the first radar observation image. The weight distribution of the target clutter window is used to integrate the energy peak value in the detection point's domain. The weighted calculation result reflects the degree of matching between the detection point and the current clutter type and the energy intensity, thus deriving the clutter level. If the clutter level is higher than a preset level threshold, the detection point is determined to be the echo energy reflected by a clutter target. If the clutter level is lower than the preset level threshold, the detection point is determined to be the echo energy reflected by a real target (i.e., a vehicle, drone, or other target of interest).

[0062] The initial clutter map is a two-dimensional matrix generated after weighted matching calculation using a clutter window. This matrix records the clutter level distribution of all detection points and their neighborhoods, but morphological void filling has not yet been performed, and there may be local data gaps or discontinuous areas. Specifically, by traversing all detection points through the target clutter window and mapping the clutter levels calculated for the neighborhoods of all detection points onto a two-dimensional coordinate plane, and setting the level of the calculation region to zero, the discrete weighted matching structure can be recombined into a two-dimensional matrix to form a preliminary description of the clutter spatial distribution (i.e., the initial clutter map).

[0063] Dilation processing involves scanning the image using a structuring element. Whenever the structuring element intersects with the image using non-zero elements, the image position corresponding to the origin of the structuring element is set to a valid value to fill in scattered voids within the clutter region. Specifically, radar equipment can use a cross-shaped window as the structuring element. A sliding window judgment is performed on the initial clutter map. If a non-zero clutter level value exists within the cross-shaped window's coverage area, the clutter level at the center of the cross-shaped window is filled with the maximum value in the neighborhood or a preset value. This morphological dilation operation connects adjacent clutter regions, filling in clutter point voids caused by sparse sampling or threshold truncation, making the clutter region continuous and complete. After dilation processing, the final clutter distribution matrix (i.e., the target clutter map) is output.

[0064] As an example, the radar equipment can align the center of the target clutter window with a detection point in the first radar observation map, and perform a weighted sum of the weights within the target clutter window and the energy in the corresponding location in the first radar observation map to obtain the clutter level of that detection point. Then, the target clutter window is traversed through all detection points to complete the calculation. Finally, a cross-shaped morphological dilation operation is performed on the clutter level matrix (i.e., the initial clutter map) to eliminate sporadic voids in the clutter region and generate the final target clutter map. This morphological processing eliminates the breaks within the clutter region, making the transition boundary from the clutter region to the non-clutter region clearer, thereby accurately depicting the clutter edge where clutter energy attenuates and transitions to background noise.

[0065] Through steps S141 to S142, the clutter level assessment can be made more accurate by aligning the target clutter window with the detection point and performing weighted calculations, thereby finely characterizing the clutter edge. At the same time, by dilating the initial clutter map, the voids inside the clutter region are effectively filled, making the generated target clutter edge more complete and accurate. This provides precise position guidance for adjusting the reference cell during subsequent CFAR detection, avoiding the problem of missing the real target due to clutter energy falling into the reference window.

[0066] The operation of the radar equipment is described in detail below. After the radar equipment is powered on, it continuously collects echoes and generates a first radar observation map. Then, the radar equipment analyzes the first radar observation map to determine the target clutter type. For example, it calculates the energy threshold along the observation dimension using a sliding window and extracts detection points based on the energy threshold. The distribution of the detection points determines the energy distribution characteristics of the first radar observation map, and then the target clutter type is determined based on the energy distribution characteristics. Next, the target clutter window corresponding to the target clutter type is retrieved from the preset clutter database. Finally, the target clutter window is used to perform weighted matching on the area of ​​the detection points to calculate the clutter level and generate a target clutter map. When a real target at the edge of the clutter is detected, the radar equipment can adjust the noise calculation method according to the clutter edge in the target clutter map to reduce the impact of clutter energy on CFAR detection.

[0067] In this embodiment, the first radar observation image is analyzed to determine the clutter type based on its energy distribution characteristics and the physical properties of different clutter types. A corresponding clutter window is then obtained based on the clutter type, enabling precise and accurate characterization of clutter edges. In CFAR processing, the selection of reference cells is adjusted based on accurate clutter edge selection, effectively avoiding the signal-to-noise ratio reduction caused by clutter energy falling into the reference window, thus significantly improving the detection rate of true targets at clutter edges. Simultaneously, the use of a sparse clutter window effectively reduces the amount of iterative computation, meeting the real-time requirements of engineering implementation.

[0068] Furthermore, to improve the accuracy of energy distribution characteristics, the calculation result obtained through sliding window calculation can be supplemented with a fixed offset or multiplied by a scaling factor to generate an energy threshold, thereby obtaining a more accurate energy threshold and improving the accuracy of energy distribution characteristics. Specifically, step S121 may include the following steps: Step S210: Slide the preset window along the observation dimension and perform sliding window calculation on the energy at each position in the first radar observation map to obtain the average energy value corresponding to each distance value in the distance dimension.

[0069] Step S220: Determine the energy threshold corresponding to each distance value based on the preset threshold and the average energy value corresponding to each distance value.

[0070] Specifically, during the sliding of the preset window, the energy values ​​of all data points within the current window's coverage area are accumulated and averaged to smooth out abnormal fluctuations through statistical mean, thereby reflecting the noise floor energy level of the local area. For each distance value in the distance dimension, after completing a full sliding window mean calculation along the observation dimension, a scalar value (i.e., the energy mean) representing the noise floor level of that distance value is output.

[0071] The preset threshold is a threshold offset of the energy threshold, used to raise the decision threshold based on the mean amplitude to control the false alarm probability. Specifically, the mean energy of each distance value is linearly superimposed with the preset threshold. That is, the mean energy calculated by the sliding window is used as the basic reference, and the preset threshold factor (i.e., the preset threshold) is superimposed to dynamically generate a decision threshold for each distance value that can adapt to changes in background noise, i.e., the final decision threshold of the detection point.

[0072] In addition to being a fixed energy value, the preset threshold can also be a proportionality coefficient, and the product of this proportionality coefficient and the average energy value is used as the energy threshold. Of course, the preset threshold can also include a proportionality coefficient and a fixed energy value. For example, the product of the average energy value and a preset proportionality coefficient (such as 10%) can be calculated, and then the product, the fixed energy value, and the average energy value can be added together to obtain the energy threshold.

[0073] As an example, a radar device can slide a preset window of length L along the observation dimension and calculate the average energy of the data within the preset window point by point to obtain the average row noise energy for each distance value; then the average row noise energy is added to a fixed offset value (i.e., a preset threshold) to dynamically determine the energy threshold specific to each distance value.

[0074] In the above embodiments, the average energy value is obtained by sliding window calculation, which enables adaptive and accurate tracking of noise floor fluctuations and avoids false alarms or missed detections caused by non-uniform backgrounds. The energy threshold of each distance gate is dynamically determined based on the average energy value and a preset threshold, so that the threshold setting is more in line with the real noise floor energy level, thereby providing a reliable threshold for subsequent accurate extraction of detection points and ensuring the accuracy of the energy distribution characteristics determined based on the detection points.

[0075] Furthermore, to improve the accuracy of energy distribution characteristics, step S122 specifically includes the following steps: Step S310: Perform peak detection along the range dimension on the first radar observation map to obtain multiple peak points corresponding to each range value in the range dimension.

[0076] Step S320: Compare the peak values ​​of multiple peak points corresponding to each distance value with the energy threshold.

[0077] Step S330: If the peak value of the peak point is greater than the energy threshold, then the peak point is determined as the detection point.

[0078] The peak point is a data point within a local area of ​​the first radar observation map where the energy value reaches its maximum compared to its neighboring cells. It represents the highest local response of the echo energy of a suspected real target or clutter scatterer and is a candidate for target confirmation in radar signal processing. Specifically, the radar equipment can scan the first radar observation map and compare the energy magnitude of each data point with its neighboring data points to extract all local energy maximum points. This eliminates smooth regions and noise floor fluctuations caused by main lobe expansion, retaining the true scatterer response peak.

[0079] Next, the energy value of each extracted peak point is compared with the dynamic energy threshold corresponding to the distance value of the peak point. The dynamic threshold is used to adaptively filter out false peaks that are lower than the background noise level, and retain the effective signal with a sufficiently high signal-to-noise ratio.

[0080] Finally, detection points are determined based on the comparison results. When the peak energy is greater than the calculated dynamic energy threshold, the state of that peak point is marked as a detection point. For example, if there are three peak points with energies of 10, 25, and 40 within a certain distance value, and the energy threshold for that distance value is 20, then the two peak points with energies of 25 and 40 are determined as detection points because their energies are greater than the energy threshold, while the peak point with energies of 10 is filtered out.

[0081] Through steps S310 to S330, peak detection can be performed on the first radar observation image and the dynamic energy threshold can be used for filtering to extract clutter targets. This effectively avoids false peak interference caused by noise floor fluctuations. The high-energy peak points after filtering are confirmed as detection points, ensuring the data purity of subsequent energy distribution feature statistics, thereby accurately determining the clutter type.

[0082] Furthermore, to improve the accuracy of energy distribution characteristics, step S123 specifically includes the following steps: Step S410: Count the number of detection points corresponding to each observation value in the observation dimension, and generate a histogram corresponding to the detection points.

[0083] Step S420: Analyze the peak regions in the histogram to determine the energy distribution characteristics of the first radar observation map.

[0084] Among them, a histogram is a statistical report chart that uses a series of vertical stripes or line segments of varying heights to represent the distribution of data. In this embodiment, a two-dimensional statistical chart is constructed with each observation value in the observation dimension as the horizontal axis and the number of detection points within each observation value as the vertical axis, which is used to intuitively display the aggregation state of high-energy points in the observation dimension.

[0085] The peak region is the segment in the histogram where the statistical value is significantly higher than that of the adjacent interval and is continuously distributed. It is used to indicate the interval of observation values ​​where the detection points are dense in the histogram, reflecting the spectral broadening and energy concentration range of a specific type of clutter in the observation dimension.

[0086] Specifically, by traversing the coordinate data of all detection points and classifying them according to their respective observations, the number of high-energy points in each observation channel is quantified, transforming the discrete two-dimensional point set into a one-dimensional statistical vector, thereby revealing the distribution density of clutter in the observation dimension.

[0087] Next, by calculating data such as the breadth range, center location, or amplitude ratio of the peak region in the histogram, the envelope shape of the histogram is analyzed to obtain the physical expansion characteristics of clutter in the observation dimension. Finally, the energy distribution characteristics are determined based on the physical expansion characteristics of the peak region, and the clutter type is determined through the energy distribution characteristics. For example, ground clutter appears as a single narrow peak located near a specific observation value, while rain and snow clutter appears as a wide peak spanning multiple observation values, and sea clutter appears as multiple narrow peaks located near different observation values.

[0088] As an example, after generating the histogram, the peak points of the histogram can be detected. If the histogram contains only one peak point, the clutter type can be determined to be ground clutter. If the histogram contains multiple peak points, and the number of detection points between the multiple peak points is also relatively large, the clutter type can be determined to be rain / snow clutter. If the histogram contains multiple peak points, and the number of detection points is relatively large only at the locations near the peak points, while the number of detection points at other locations is relatively small, the clutter type can be determined to be sea clutter.

[0089] In the above embodiments, by generating a histogram by statistically analyzing the number of detection points in the observation dimension, the distribution pattern of clutter in the observation dimension is converted into statistical features, effectively suppressing the interference caused by random scattered points. Then, through precise analysis of the peak region of the histogram, the physical broadening characteristics of different clutter in the observation dimension can be accurately revealed, thereby providing a reliable feature basis for subsequent accurate determination of clutter type and selection of target clutter window.

[0090] Furthermore, to improve the accuracy of target clutter maps, radar equipment can generate corresponding clutter windows based on its operating parameters and environmental parameters, providing an accurate data foundation for subsequent target clutter map generation, such as... Figure 5 As shown, Figure 5 A flowchart of another clutter detection method is shown, which further includes the following steps: Step S510: Obtain the clutter window size corresponding to multiple clutter types.

[0091] The clutter window size is used to characterize the coverage of the clutter window in the range dimension and the observation dimension. The design of the clutter window size needs to ensure that multiple clutter targets fall within the same clutter window, while multiple non-clutter targets do not fall within the same clutter window, so as to avoid misjudging non-clutter targets as clutter targets.

[0092] The clutter window size can be user-preset data, or it can be determined by the radar equipment based on the distribution density of clutter targets, the distribution density of real targets, and the sampling interval and sampling range of the scatterer characteristics in the entire radar observation map. Specifically, the radar equipment can automatically acquire relevant generation parameters to determine the clutter window size after activation. Step S510 may include the following steps (steps S511 to S516): Step S511: Obtain radar detection data corresponding to each clutter type, and generate a second radar observation map corresponding to each clutter type based on the radar detection data corresponding to each clutter radar.

[0093] For each type of clutter, the following steps are performed on the second radar observation map: Step S512: Analyze the energy at each location in the second radar observation map according to the first preset energy threshold to determine the scatterer characteristic information.

[0094] Step S513: Divide the second radar observation map into a real target area and a clutter target area according to the second preset energy threshold.

[0095] Step S514: Analyze the energy at each location in the real target area according to the third preset energy threshold to determine the distribution density of the real target.

[0096] Step S515: Analyze the energy at each location in the clutter target area according to the fourth preset energy threshold to determine the distribution density of the clutter target.

[0097] Step S516: Determine the clutter window size corresponding to the clutter type based on the scatterer characteristic information, the distribution density of the real target, and the distribution density of the clutter target.

[0098] The second radar observation map is a radar observation map obtained by preprocessing radar echoes collected by radar equipment for a specific clutter type, and is used for offline or preliminary analysis of clutter physical characteristics.

[0099] The first preset energy threshold is used to initially distinguish between effective scatterer echoes and system noise floor. Its main goal is to filter out interference from pure noise regions and retain effective signals that reflect the presence and distribution range of scatterers. Scatterer characteristic information refers to the characteristic parameters of objects that can reflect radar signals, such as distribution range, sampling interval, and spectral broadening, as shown on the radar observation map. These parameters reflect the coverage breadth and radial motion characteristics (or azimuth characteristics) of the scatterer in space.

[0100] The second preset energy threshold is used to distinguish between targets of interest and non-targets. Based on the differences in radar cross-section and echo energy intensity between real targets and clutter targets, the radar observation map is divided into two different types of regions. The real target region refers to the high-energy region in the radar observation map where the energy originates from the detected target (such as a vehicle or drone). The echo points in this region represent signals that the radar needs to focus on and extract. The clutter target region refers to the non-target region in the radar observation map where the energy is higher than the noise floor. The echo points in this region represent clutter reflected from surface or volume targets such as ground, sea, rain, or snow.

[0101] The third preset energy threshold refers to the energy threshold used to statistically determine the density of effective target points within the real target area. This threshold can be used to detect false alarms or weak interference, ensuring that the statistical distribution density truly reflects the density of detected targets. The distribution density of real targets is the proportion of high-energy points within the real target area. This parameter reflects the spatial dispersion of detected targets in the range and observation dimensions and is a key constraint in determining the clutter window size to avoid multiple real targets falling into the same clutter window.

[0102] The fourth preset energy threshold refers to the energy threshold used to statistically determine the effective clutter point density within the clutter target area. This threshold eliminates edge points with attenuation exceeding half, retaining strong scattering points that represent the density of the clutter core area. The clutter target distribution density refers to the proportion of high-energy points within the clutter target area. This parameter reflects the spatial clustering of clutter scatterers in both the range and observation dimensions, and is the core basis for determining the clutter window size to ensure that multiple clutter targets fall within the window.

[0103] As an example, radar equipment collects raw echoes under different geographical environments or weather conditions to establish dedicated datasets for typical scenarios such as ground clutter, rain and snow clutter, and sea clutter, providing a data foundation for subsequent analysis of the physical distribution characteristics of various clutter types. Next, the radar equipment sequentially performs pulse compression and fast Fourier transform on the collected raw echoes, converting the time-domain echoes to a two-dimensional frequency domain of range and Doppler dimensions, enabling the differentiation of scatterers at different distances and velocities and forming a visual image.

[0104] Then, the radar equipment compares the energy value of each point in the second radar observation map with the first preset energy threshold one by one. In order to eliminate invalid data in the noise area through threshold comparison, retain the echo energy distribution of the effective scatterer, and calculate the coverage and sampling interval of the pixels that exceed the first preset energy threshold in the range dimension and observation dimension, so as to quantify the spread and distribution state of the scatterer in the two-dimensional plane and provide data basis for clutter window shape design.

[0105] Next, using the second preset energy threshold as a boundary, the pixel coordinates with energy higher than the threshold are assigned to the real target area, and the pixel coordinates with energy lower than the threshold but higher than the background noise are assigned to the clutter target area. The energy difference is used to separate the target of interest from the background clutter.

[0106] Then, within the real target area, the effective detection target points are further filtered using the third preset energy threshold, and the ratio of the number of points exceeding the third preset energy threshold in the real target area to the total number of points in the area is calculated; within the clutter target area, the high-energy points in the clutter core area are extracted using the fourth preset energy threshold, and the ratio of the number of points exceeding the fourth preset energy threshold in the clutter target area to the total number of points in the area is calculated.

[0107] Finally, the window boundary is calculated by combining the coverage of the scatterer, the sparsity of the real target, and the density of the clutter target to ensure that multiple clutter targets fall within the same clutter window, while multiple non-clutter targets do not fall within the same clutter window. At the same time, the sampling interval and sampling range of the scatterer characteristics are combined to limit the limit size of the window. Specifically, the clutter window size needs to meet the following formula requirements:

[0108] in, For clutter window size, The distribution density of clutter targets, The distribution density of the real target. The sampling interval for the scatterer characteristics. The sampling range represents the characteristics of the scatterer.

[0109] Through steps S511 to S516, the radar observation map is refined through multi-level energy threshold analysis to accurately extract scatterer feature information, real target density, and clutter target density, providing accurate data for determining the clutter window size. This ensures that the generated clutter window size can both enclose enough clutter targets to accurately characterize the clutter edge and avoid multiple real targets falling into the same clutter window, thus preventing misjudgment.

[0110] After step S510, step S520 is executed: the operating parameters of the radar equipment and the environmental parameters of the detection area are obtained, and the distance data and observation data corresponding to each clutter type are determined based on the operating parameters and environmental parameters.

[0111] Among them, the operating parameters of radar equipment refer to the radar system configuration data that affect the frequency shift and range broadening of clutter observations, including but not limited to the radar equipment's movement speed, azimuth angle, and beamwidth. These parameters directly determine the projection speed and range of clutter in the observation dimension.

[0112] As an example, since there are a large number of stationary surface targets and volume targets on the ground, when radar identifies a moving target, the relative velocity between the target and the radar equipment is mainly determined by the radial component of the radar equipment's velocity. Let's assume the radar equipment's velocity is... The beamwidth is azimuth angle is The radial projection of the ground clutter velocity in the Doppler view is: ,in, and It can be determined using the following formula: , .

[0113] The environmental parameters of the detection area refer to the parameters that reflect the physical characteristics of the weather or sea conditions in the detection scene, including but not limited to the falling speed of rain and snow, wind speed, turbulence, and the diffusion speed of ocean waves. These parameters determine the actual motion state of the volume target or surface target in space.

[0114] As an example, raindrops and snowflakes possess a certain velocity during their fall. Therefore, targets such as raindrops and snowflakes will generate a radial velocity component relative to radar equipment. The actual falling velocity of rain and snow depends on factors such as particle size, shape, wind speed, and turbulence. Taking ordinary rainfall as an example, raindrop diameters range from 0.5 to 2.0 mm, with falling velocities ranging from 2.0 to 6.0 m / s. The velocity of most raindrop particles is concentrated between 2.0 and 4.8 m / s, with a peak of approximately 2.8 m / s. Different radial components will be generated at different angles. Therefore, the velocity projection of raindrops on the Doppler plane depends on the rain / snow velocity, radar velocity resolution, radar elevation beamwidth, and pointing, and is typically determined by more than five parameters.

[0115] Range data refers to the distribution range and number of elements of clutter scatterers in the range dimension. It is usually determined by the radar resolution and the spatial distribution breadth of the scatterers in environmental parameters. For example, rain and snow clutter, as volume targets, are usually distributed across multiple range values ​​in the range dimension. Observation data refers to the projection set and number of elements of clutter scatterers in the observation dimension. It is usually calculated from radar operating parameters and environmental parameters. For example, the velocity of ground clutter is the radial projection of the radar equipment's moving velocity, while the velocity of rain and snow clutter depends on the falling speed of rain and snow, radar velocity resolution, radar elevation beamwidth and pointing, etc., and is usually determined by more than five parameters.

[0116] Radar equipment can read bus data from the radar system and input data from external sensors (such as weather radar) to obtain dynamic variables affecting clutter distribution in real time, providing real-time input for subsequent calculations of the actual physical distribution of clutter.

[0117] Next, the radar equipment substitutes its operating parameters (such as radar velocity, beamwidth, etc.) and environmental parameters (such as raindrop falling velocity, etc.) into the clutter physical distribution model for calculation. Based on the kinematic projection principle, it derives the distance data and observation data of each clutter on the second radar observation map. For example, the ground clutter velocity is the radial projection of the radar velocity.

[0118] Step S530: Generate clutter windows corresponding to each clutter type based on the clutter window size, distance data, observation data, and preset weight formula.

[0119] The preset weight formula refers to a mathematical expression pre-stored based on the clutter type for calculating the weights at different positions within the clutter window. The weight calculation primarily depends on the distribution characteristics of different types of clutter targets in the range dimension and velocity dimension, while also considering energy leakage and main lobe broadening of the central target. Furthermore, the smoothness of the weights within the clutter window needs to be adjusted based on the specific scatterer characteristics, for example... Figure 2a and Figure 2c As shown, although the distribution characteristics of ground clutter are similar to those of wavy sea clutter, the characteristics of scatterers on the ground are more complex. Therefore, the weights within the clutter window corresponding to ground clutter are as follows: Figure 3a The diagram is more complex, and the weights within the clutter window corresponding to sea clutter are as follows: Figure 3c The result is smoother.

[0120] Within a defined clutter window size range, radar equipment defines the shape using range data and observation data. Starting from the center of the clutter window, it calculates the weights at different locations within the detection window according to a preset weight formula. This ensures that the shape and weight distribution of the clutter window correspond to the distribution of clutter targets within the clutter region. Each non-zero weight represents a high-energy peak value in the clutter target domain at the center of the clutter window. For example, ground clutter windows are mainly distributed along the range dimension, with weights decreasing closer to the center. Rain and snow clutter windows are designed as sparse clutter windows with a diffuse distribution, while sea clutter windows are designed as clutter windows with smoother weights distributed along the range dimension.

[0121] Step S540: Store the clutter windows corresponding to each clutter type in the clutter database.

[0122] After generating clutter windows, the electronic device stores the size parameters, shape matrix, and weight matrix of various types of clutter windows in non-volatile memory or cache according to the clutter type index to build a prior knowledge base. This allows the corresponding clutter window to be directly retrieved once the clutter type is identified during real-time clutter detection, avoiding computational delays caused by real-time generation.

[0123] In the above embodiments, clutter windows are dynamically generated based on actual working parameters and environmental parameters, making the size and shape of the clutter windows more closely match the actual physical distribution characteristics of clutter in the current detection scenario, thus achieving accurate quantification of prior information. Then, the weights within the clutter window are calculated using a preset weight formula, fully considering the energy leakage of the central target and the broadening of the main lobe, making the generated clutter window more sensitive to clutter edges. Finally, the generated clutter windows are pre-stored in the database, enabling direct retrieval and matching during real-time detection. This effectively reduces the amount of real-time computation while ensuring fine characterization of clutter edges, thereby meeting the real-time requirements of engineering implementation.

[0124] It should be noted that the process of generating clutter windows in steps S510 to S540 can be carried out by collecting detection data corresponding to multiple clutter types when the radar device is turned on for the first time, and generating clutter windows corresponding to each clutter type; or by generating the corresponding clutter window when a clutter type is detected for the first time. For example, the radar device can generate clutter windows corresponding to rain and snow clutter based on the detection data it collects for the first time in a rain and snow environment; or by collecting detection data corresponding to multiple clutter types in advance, and then processing these detection data offline to generate clutter windows corresponding to each clutter type.

[0125] Figure 6 The diagram shows a structural schematic of a radar device provided in an embodiment of this application. The specific embodiments of this application do not limit the specific implementation of the radar device.

[0126] like Figure 6As shown, the radar device 1 may include a processor 11 and a memory 12.

[0127] The memory 12 is used to store the computer program 13. The memory 12 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device. The computer program 13 may include computer-executable instructions.

[0128] The processor 11 is used to execute the computer program 13 to implement the above-described clutter detection method embodiment.

[0129] The processor 11 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The radar device 1 includes one or more processors, which may be processors of the same type, such as one or more CPUs; or they may be processors of different types, such as one or more CPUs and one or more ASICs.

[0130] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described clutter detection method embodiment.

[0131] This application provides a computer program that can be executed by a processor to implement the above-described clutter detection method embodiments.

[0132] This application provides a computer program product, which includes a computer program that, when executed by a processor, implements the above-described clutter detection method embodiment.

[0133] In the several embodiments provided in this application, any function, if implemented as a software functional module / unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the technical solution of this application 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 other electronic device) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing computer program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0134] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used in conjunction with the teachings herein. The required structure for constructing such systems is apparent from the above description. Furthermore, the embodiments of this application are not directed to any particular programming language. It should be understood that the content of this application described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing the best mode of implementation of this application.

[0135] It should be noted that the above embodiments are illustrative of this application and not restrictive, and those skilled in the art can devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses should not be construed as limiting the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "a" or "an" preceding an element does not exclude the presence of a plurality of such elements. This application can be implemented by means of hardware comprising several different elements and by means of a suitably programmed computer. In claims enumerating several means, several units or modules of these means may be embodied by the same item of hardware. The use of the words first, second, and third, etc., does not indicate any order. These words can be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be construed as limiting the order of execution.

[0136] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A clutter detection method, characterized in that, Applied to radar equipment, the method includes: Acquire radar detection data and process the radar detection data to generate a first radar observation map; The energy at each location in the first radar observation map is analyzed to determine the energy distribution characteristics of the first radar observation map, and the target clutter type is determined based on the energy distribution characteristics. According to the target clutter type, a preset clutter database is searched to obtain the target clutter window corresponding to the target clutter type. The clutter database stores multiple clutter types and clutter windows corresponding to each clutter type. The target clutter window is used to analyze the first radar observation map to generate a target clutter map.

2. The clutter detection method according to claim 1, characterized in that, The first radar observation map includes the range dimension and the observation dimension; The step of analyzing the energy at each location in the first radar observation map to determine the energy distribution characteristics of the first radar observation map specifically includes: The preset window is slid along the observation dimension, and the energy at each position in the first radar observation map is calculated by sliding the window to determine the energy threshold corresponding to each distance value in the distance dimension. The energy at each location in the first radar observation map is analyzed based on the energy threshold corresponding to each distance value in the distance dimension to determine multiple detection points in the first radar observation map; The locations of the multiple detection points are analyzed to determine the energy distribution characteristics of the first radar observation map.

3. The clutter detection method according to claim 2, characterized in that, The step of sliding a preset window along the observation dimension and performing sliding window calculations on the energy at each location in the first radar observation map to determine the energy threshold corresponding to each distance value in the range dimension specifically includes: The preset window is slid along the observation dimension, and the energy at each position in the first radar observation map is calculated by sliding the window to obtain the average energy value corresponding to each distance value in the distance dimension; The energy threshold corresponding to each distance value is determined based on a preset threshold and the average energy value corresponding to each distance value.

4. The clutter detection method according to claim 2, characterized in that, The step of analyzing the energy at each location in the first radar observation map based on the energy threshold corresponding to each distance value in the distance dimension, and determining multiple detection points in the first radar observation map, specifically includes: Peak detection is performed on the first radar observation map along the distance dimension to obtain multiple peak points corresponding to each distance value in the distance dimension; Compare the peak values ​​of multiple peak points corresponding to each distance value with the energy threshold; If the peak value of the peak point is greater than the energy threshold, then the peak point is determined as the detection point.

5. The clutter detection method according to claim 2, characterized in that, The step of analyzing the locations of the multiple detection points to determine the energy distribution characteristics of the first radar observation map specifically includes: Count the number of detection points corresponding to each observation value in the observation dimension, and generate a histogram corresponding to the detection points; The peak regions in the histogram are analyzed to determine the energy distribution characteristics of the first radar observation map.

6. The clutter detection method according to claim 1, characterized in that, The step of analyzing the first radar observation map using the target clutter window to generate a target clutter map specifically includes: Align the center point of the target clutter window with each of the detection points, and calculate the clutter level corresponding to each position in the first radar observation map according to the target clutter window to generate an initial clutter map; The initial clutter map is dilated to generate the target clutter map.

7. The clutter detection method according to any one of claims 1 to 6, characterized in that, The method further includes: Obtain the clutter window size corresponding to multiple clutter types; The operating parameters of the radar equipment and the environmental parameters of the detection area are obtained, and the range data and observation data corresponding to each clutter type are determined based on the operating parameters and the environmental parameters. Based on the clutter window size, distance data, observation data, and preset weighting formula corresponding to each clutter type, a clutter window corresponding to each clutter type is generated; The clutter windows corresponding to each of the aforementioned clutter types are stored in the clutter database.

8. The clutter detection method according to claim 7, characterized in that, The step of obtaining the clutter window size corresponding to multiple clutter types specifically includes: Acquire radar detection data corresponding to each of the aforementioned clutter types, and generate a second radar observation map corresponding to each of the aforementioned clutter types based on the radar detection data corresponding to each of the aforementioned clutter radars; For each of the aforementioned clutter types, the following steps are performed on the corresponding second radar observation map: The energy at each location in the second radar observation map is analyzed based on the first preset energy threshold to determine the scatterer characteristic information; The second radar observation map is divided into a real target area and a clutter target area according to the second preset energy threshold. The energy at each location in the real target area is analyzed based on a third preset energy threshold to determine the distribution density of the real target. The energy at each location in the clutter target region is analyzed based on the fourth preset energy threshold to determine the distribution density of the clutter target. The clutter window size corresponding to the clutter type is determined based on the scatterer characteristic information, the distribution density of the real target, and the distribution density of the clutter target.

9. A radar device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the clutter detection method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the clutter detection method according to any one of claims 1 to 8.

Images