Target point cloud clustering method based on radar multi-dimensional feature information and related equipment

Abstract

The invention provides a target point cloud clustering method based on radar multi-dimensional feature information and related equipment. The target point cloud clustering method based on the radar multi-dimensional feature information comprises the steps of determining an echo signal and a transmitting signal of a radar, performing frequency mixing on the echo signal and the transmitting signal, and determining an intermediate frequency signal; according to the intermediate frequency signal, position information of any target point cloud is determined through distance-Doppler FFT processing; determining a radial distance between the target point cloud and the radar according to the position information, carrying out classification processing on the target point cloud according to the radial distance, and determining a short-distance target point cloud and a long-distance target point cloud; clustering the close-range target point clouds through a clustering method based on space coordinate dimensions, and determining a close-range point cloud clustering result; clustering the long-distance target point cloud through a clustering method based on speed radial distance dimension, and determining a long-distance point cloud clustering result; and fusing the short-distance point cloud clustering result and the long-distance point cloud clustering result.

Description

Technical Field

[0001] This application relates to the technical field of radar signal processing and data processing, and particularly to a method for clustering target point clouds based on radar multi-dimensional feature information and related devices. Background Art

[0002] Radar target recognition technology not only significantly improves traffic safety and efficiency, but also plays an important role in applications such as aviation, navigation, weather forecasting, and intelligent transportation systems. Millimeter-wave radar can construct three-dimensional point cloud images of the environment and targets by accurately sensing the surrounding environment, and accurately achieve target recognition through the three-dimensional point cloud images.

[0003] However, in the prior art, there are problems in target recognition where the calculation results are inaccurate, leading to inaccurate target recognition results. Summary of the Invention

[0004] In view of this, the purpose of this application is to propose a method for clustering target point clouds based on radar multi-dimensional feature information and related devices.

[0005] Based on the above purpose, this application provides a method for clustering target point clouds based on radar multi-dimensional feature information for target recognition, including:

[0006] Determine the echo signal and the transmitted signal of the radar, mix the echo signal with the transmitted signal, and determine the intermediate frequency signal;

[0007] According to the intermediate frequency signal, determine the position information of any target point cloud through range-Doppler FFT processing;

[0008] According to the position information, determine the radial distance between the target point cloud and the radar, and classify the target point cloud according to the radial distance to determine the near-distance target point cloud and the far-distance target point cloud;

[0009] Cluster the near-distance target point cloud through a clustering method based on the spatial coordinate dimension to determine the near-distance point cloud clustering result;

[0010] Cluster the far-distance target point cloud through a clustering method based on the velocity radial distance dimension to determine the far-distance point cloud clustering result;

[0011] Fuse the near-distance point cloud clustering result and the far-distance point cloud clustering result, and perform target recognition according to the fusion result.

[0012] Optionally, the step of determining the position information of any target point cloud through range-Doppler FFT processing according to the intermediate frequency signal includes:

[0013] Extract information from the intermediate-frequency signal through the distance-Doppler FFT method to determine the distance information, velocity information, and azimuth information of the target point cloud;

[0014] Determine the position information according to the distance information and the azimuth information.

[0015] Optionally, the method further includes:

[0016] Perform clutter suppression processing and denoising processing on the distance information, the velocity information, and the azimuth information through 2D-CFAR.

[0017] Optionally, the classifying the target point cloud according to the radial distance to determine a near-distance target point cloud and a far-distance target point cloud includes:

[0018] In response to determining that the radial distance is not greater than a preset distance, determine the target point cloud as the near-distance target point cloud;

[0019] In response to determining that the radial distance is greater than the preset distance, determine the target point cloud as the far-distance target point cloud.

[0020] Optionally, the clustering the near-distance target point cloud through a clustering method based on the spatial coordinate dimension to determine a near-distance point cloud clustering result includes:

[0021] According to the position information of the near-distance target point cloud and a preset clustering domain;

[0022] Screen the data point clouds in the preset clustering domain according to the position information, and determine the data objects to be clustered according to the screening result; wherein, the data objects to be clustered include core objects and non-core objects, and the screening includes: determining the relationship between any one of the data point clouds and the near-distance target point cloud, and taking the data point clouds with the relationship of density direct reach or density reachable or density connected as the data objects to be clustered;

[0023] Perform data clustering on the data objects to be clustered through a clustering method based on the spatial coordinate dimension to determine the near-distance point cloud clustering result; wherein, the clustering method based on the spatial coordinate dimension is the DBSCAN clustering method based on the spatial coordinate dimension.

[0024] Optionally, the clustering the far-distance target point cloud through a clustering method based on the velocity radial distance dimension to determine a far-distance point cloud clustering result includes:

[0025] Determine the data objects to be clustered according to the near-distance target point cloud and a preset clustering domain; wherein, the data objects to be clustered include core objects and non-core objects;

[0026] Based on the data objects to be clustered, data clustering is performed through a clustering method based on the velocity radial distance dimension to determine the clustering result of the far-distance point cloud; wherein, the clustering method based on the velocity radial distance dimension is the DBSCAN clustering method of the velocity radial distance dimension.

[0027] Based on the same inventive concept, an embodiment of the present application further provides a target point cloud clustering device based on radar multi-dimensional feature information, including:

[0028] A mixing module, configured to determine an echo signal and a transmission signal of the radar, mix the echo signal with the transmission signal to determine an intermediate frequency signal;

[0029] A signal processing module, configured to determine position information through range-Doppler FFT processing according to the intermediate frequency signal;

[0030] A classification module, configured to determine the radial distance between the target point cloud and the radar according to the position information, and perform classification processing on the target point cloud according to the radial distance to determine a near-distance target point cloud and a far-distance target point cloud;

[0031] A first clustering module, configured to cluster the near-distance target point cloud through a clustering method based on the spatial coordinate dimension to determine the clustering result of the near-distance point cloud;

[0032] A second clustering module, configured to cluster the far-distance target point cloud through a clustering method based on the velocity radial distance dimension to determine the clustering result of the far-distance point cloud;

[0033] A fusion module, configured to fuse the clustering result of the near-distance point cloud and the clustering result of the far-distance point cloud, and perform target recognition according to the fusion result.

[0034] Based on the same inventive concept, an embodiment of the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the program, it implements the target point cloud clustering based on radar multi-dimensional feature information as described in any one of the above.

[0035] Based on the same inventive concept, an embodiment of the present application further provides a non-transitory computer-readable storage medium, where the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to cause a computer to execute the target point cloud clustering based on radar multi-dimensional feature information as described in any one of the above.

[0036] Based on the same inventive concept, an embodiment of the present application further provides a computer program product, including computer program instructions, which when running on a computer, cause the computer to execute any one of the above-mentioned target point cloud clustering methods based on radar multi-dimensional feature information.

[0037] As can be seen from the above, a target point cloud clustering method and related devices provided by the present application, through the clustering recognition of point cloud targets within the radar field of view, make full use of the feature information of the radar to obtain target points. By adding the speed-radial distance dimension clustering, the accuracy of long-distance target recognition is improved, and it is fused with the spatial coordinate system feature information used for close distances, greatly improving the accuracy of target recognition throughout the field of view. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only the embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 It is a schematic flowchart of the target point cloud clustering method based on radar multi-dimensional feature information according to an embodiment of the present application;

[0040] Figure 2a It is a schematic diagram of the experimental radar of the target point cloud clustering method based on radar multi-dimensional feature information according to an embodiment of the present application;

[0041] Figure 2b It is a schematic diagram of the experimental scene of the target point cloud clustering method based on radar multi-dimensional feature information according to an embodiment of the present application;

[0042] Figure 3 It is a schematic diagram of the original point cloud map according to an embodiment of the present application;

[0043] Figure 4 It is a schematic diagram of the clustering result of the existing clustering method according to an embodiment of the present application;

[0044] Figure 5 It is a schematic diagram of the clustering result in the speed-radial distance dimension according to an embodiment of the present application;

[0045] Figure 6 It is a schematic diagram of the point cloud clustering of the present solution according to an embodiment of the present application;

[0046] Figure 7 It is a schematic diagram of the target point cloud clustering device based on radar multi-dimensional feature information according to an embodiment of the present application;

[0047] Figure 8 This is a schematic structural diagram of the electronic device according to an embodiment of the present application. Detailed implementation manners

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0049] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should have the ordinary meanings understood by those of ordinary skill in the art to which the present application belongs. The "first", "second" and similar terms used in the embodiments of the present application do not denote any order, quantity or importance, but are only used to distinguish different components. The terms such as "include" or "comprise" mean that the elements or objects appearing before this word cover the elements or objects listed after this word and their equivalents, without excluding other elements or objects. The terms such as "connect" or "couple" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0050] It can be understood that, before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to users and the authorization of users should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0051] In order to make the technical solutions of the present disclosure clearer and easier to understand, the method for clustering target point clouds based on radar multi-dimensional feature information provided in the embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings.

[0052] As described in the background art section, radar target recognition technology not only significantly improves traffic safety and efficiency, but also plays an important role in applications such as aviation, navigation, weather forecasting, and intelligent transportation systems. Millimeter-wave radar can construct three-dimensional point cloud images of the environment and targets by accurately sensing the surrounding environment, and accurately realize target recognition through the three-dimensional point cloud images.

[0053] However, in the prior art, there is a problem that the calculation results in target recognition are inaccurate, resulting in inaccurate target recognition results.

[0054] In view of this, an embodiment of the present application provides a method, apparatus, electronic device, storage medium, and program product for clustering target point clouds based on radar multi-dimensional feature information. The method for clustering target point clouds based on radar multi-dimensional feature information is implemented based on an FMCW radar system, and includes the following steps: First, perform operations such as range-Doppler FFT and 2D-CFAR on the radar echo signal to obtain feature information such as the three-dimensional spatial coordinates, velocity, and radial distance of the target point cloud. Then, first use the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm to cluster the point cloud in the three-dimensional coordinate space, and then establish a velocity-distance space to perform clustering again. Since the different distances between the same target and the radar will affect the quantity and density distribution of the point cloud, appropriate clustering results need to be selected according to the distance between the two. When the distance is short, the clustering in the coordinate space dimension is given priority, and when the distance is long, the clustering in the velocity and radial distance spaces is emphasized. The two are combined to achieve the final point cloud clustering. By making full use of multi-dimensional information such as the distance, velocity, and spatial coordinates of the target point cloud, this method effectively improves the accuracy of point cloud category determination and enhances the clustering effect of the target point cloud.

[0055] It can be seen that by clustering and identifying the point cloud targets within the radar field of view, fully utilizing the feature information of the target points obtained by the radar, and improving the accuracy of long-distance target recognition by adding clustering in the velocity-radial distance dimension and fusing it with the clustering of the feature information of the spatial coordinate system used at close range, the accuracy of target recognition throughout the field of view is greatly improved.

[0056] Such as Figure 1 is a flowchart showing a method for clustering target point clouds based on radar multi-dimensional feature information according to an embodiment of the present application.

[0057] Such as Figure 1 As shown, the method for clustering target point clouds based on radar multi-dimensional feature information includes:

[0058] Step S102: Determine the echo signal and the transmission signal of the radar, mix the echo signal with the transmission signal, and determine the intermediate frequency signal;

[0059] Step S104: According to the intermediate frequency signal, determine the position information of any target point cloud through range-Doppler FFT processing;

[0060] Step S106: According to the position information, determine the radial distance between the target point cloud and the radar, and classify the target point cloud according to the radial distance to determine the close-range target point cloud and the long-range target point cloud;

[0061] Step S108: Cluster the short-range target point cloud through a clustering method based on the spatial coordinate dimension to determine the short-range point cloud clustering result;

[0062] Step S110: Cluster the long-range target point cloud through a clustering method based on the velocity radial distance dimension to determine the long-range point cloud clustering result;

[0063] Step S112: Fuse the short-range point cloud clustering result and the long-range point cloud clustering result, and perform target recognition based on the fusion result.

[0064] In step S102, signal transmission and signal reception are performed by an FMCW radar system, and the transmitted signal is as follows:

[0065]

[0066] where m ∈ [0,..., N - 1], let is defined as the signal time in the fast time dimension (range dimension), t s = mT is defined as the signal time in the slow time dimension (Doppler dimension), f c is the carrier center frequency, T is the total transmission duration of a Chirp signal, γ is the frequency modulation slope, T p is the sweep period, and j is the imaginary unit.

[0067] Since the subsequent processing of this application only processes within the range of [mT, mT + T p ) where there is a signal, the above formula is rewritten as follows:

[0068]

[0069] where f c is the carrier center frequency, γ is the frequency modulation slope, and j is the imaginary unit.

[0070] In some embodiments, after the radar transmits a radio frequency (RF) signal, the target echo signal is received by the receiver and down-converted to an intermediate frequency (IF) signal by a mixer for subsequent digital signal processing.

[0071] Furthermore, the extracted intermediate frequency signal is as follows:

[0072]

[0073] Therefore, it can be seen that the first term of the frequency difference in the intermediate frequency signal contains the distance information r, and the second term contains the velocity information v. Where f c is the carrier center frequency, γ is the frequency modulation slope, j is the imaginary unit, and c is the speed of light.

[0074] In some alternative embodiments, mixing can be achieved through analog mixers, digital mixing, SDR, zero-IF architectures, and superheterodyne architectures.

[0075] In some embodiments, in a target recognition scenario, a radar device as described in Figure 2a is used to recognize a target in a moving state as described in Figure 2b . As shown in Figure 2b , 4 people are walking within a radar azimuth of 60° and a pitch angle of 60° in an open area. Among them, 2 people are approaching the radar head-on from a distance, and 2 people are moving away from the radar with their backs turned from a short distance. Exemplarily, during the entire experiment, the distance between the human body and the y-axis of the radar is set to 0 - 25 m.

[0076] In some embodiments, according to the intermediate frequency signal, through range-Doppler FFT processing, determining the position information of any target point cloud includes:

[0077] Extracting information from the intermediate frequency signal through the range-Doppler FFT method to determine the range information, velocity information, and azimuth information of the target point cloud;

[0078] According to the range information and the azimuth information, determining the position information. Wherein, the position information is position information including three-dimensional space coordinate points (x, y, z coordinates).

[0079] Furthermore, through 2D-CFAR, clutter suppression processing and denoising processing are performed on the range information, the velocity information, and the azimuth angle information.

[0080] In addition, in order to enhance the readability and intuitiveness of the point cloud data, the point cloud data obtained after 2D-CFAR is mapped to a spatial coordinate system to determine the original point cloud map as shown in Figure 3 .

[0081] In some embodiments, the intermediate frequency analog signal is digitized through analog-to-digital conversion (ADC) to generate discrete time series data.

[0082] In some embodiments, according to the radial distance, classifying the target point cloud to determine a near-range target point cloud and a far-range target point cloud includes:

[0083] In response to determining that the radial distance is not greater than a preset distance, determining the target point cloud as the near-range target point cloud;

[0084] In response to determining that the radial distance is greater than the preset distance, determining the target point cloud as the far-range target point cloud.

[0085] Specifically, setting a target point (i.e., as shown in Figure 2bThe radial distance between the person in the moving state shown in the figure and the radar version is R, and the distance resolution threshold R the = 10m. If R ≤ R the , then select the close-range - spatial coordinate dimension for clustering. If R > R the , then select the long-range - velocity radial dimension for clustering, so as to combine the advantages of point cloud clustering of the two methods in different distance cases and improve the clustering accuracy.

[0086] In some embodiments, clustering the close-range target point cloud by the clustering method based on the spatial coordinate dimension to determine the close-range point cloud clustering result includes:

[0087] According to the position information of the close-range target point cloud and the preset clustering domain;

[0088] Filter the data point cloud in the preset clustering domain according to the position information, and determine the data object to be clustered according to the filtering result; wherein, the data object to be clustered includes a core object and a non-core object, and the filtering includes: determining the relationship between any data point cloud and the close-range target point cloud, and taking the data point cloud with the relationship of density direct reach or density reachable or density connected as the data object to be clustered;

[0089] Cluster the data according to the data object to be clustered by the clustering method based on the spatial coordinate dimension to determine the close-range point cloud clustering result; wherein, the clustering method based on the spatial coordinate dimension is the DBSCAN clustering method based on the spatial coordinate dimension.

[0090] In some embodiments, clustering the long-range target point cloud by the clustering method based on the velocity radial distance dimension to determine the long-range point cloud clustering result includes:

[0091] Determine the data object to be clustered according to the close-range target point cloud and the preset clustering domain; wherein, the data object to be clustered includes a core object and a non-core object;

[0092] Cluster the data according to the data object to be clustered by the clustering method based on the velocity radial distance dimension to determine the long-range point cloud clustering result; wherein, the clustering method based on the velocity radial distance dimension is the DBSCAN clustering method of the velocity radial distance dimension.

[0093] Specifically, perform DBSCAN clustering (key parameter - used to define the minimum number of neighborhood points required for a point to become a core point) according to the current position information of the close-range target point cloud and the empirically set specific domain r1 and (MinPoints)1 to obtain the close-range point cloud clustering result.

[0094] Including: First, for x j ∈D, where the r-neighborhood contains the samples in the sample set D that are at a distance less than the neighborhood distance r from x j , that is, N r (x j ) = {x j ∈D | dist(x i , x j ) ≤ r}. If the number of all data objects in the r-neighborhood of x j is at least MinPoints, that is, |N r (x j )| ≥ MinPts, then the data object x j is called a core object. If the data object x j is located in the r-neighborhood of the data object x i , and x i is a core object, then it is said that x j is directly density-reachable from x i . Among them, "MinPoints" can be set according to specific situations. "If the number of all data objects in the r-neighborhood of x j is at least MinPoints" means that if the number of all data objects in the region centered at the point x j with a radius of r is at least the preset MinPoints

[0095] For the data object x j and the data object x i , if there exists a sample sequence P1, P2,... P n , where P1 = x i , P n = x j and P i+1 is directly density-reachable from P i , then it is said that x j is density-reachable from x i (that is, x i → x j ). That is, starting from a point, if another point can be reached through a series of points (each point is a directly density-reachable point of the previous point), then the latter is said to be density-reachable from the former. Among them, this is an asymmetric relationship, that is, if point B is density-reachable from point A, it does not mean that point A is necessarily density-reachable from point B.

[0096] For the data object x j and the data object x i , if there exists x k such that both x i and x j are density-reachable from x k , then it is said that xj Connected by x i density (x i →x k →x j ). That is, if there exists a point o such that both point p and point q can be density-reachable from point o, either directly or indirectly through other core points, then we say that point p and point q are density-connected.

[0097] If the data object x j is a point that does not belong to any cluster and is density-unreachable from any core point, it is called an outlier or anomaly point.

[0098] Optionally, based on the empirical experience obtained from a large amount of actual data verification for the domain distance r1 and the minimum number of data points (MinPoints)1 within the r-neighborhood, for clustering in the spatial coordinate system, setting r1 to 0.09 and (MinPoints)1 to 20 is the best DBSCAN parameter setting.

[0099] In some embodiments, in order to make full use of the target point information obtained by the radar and improve the accuracy of target point cloud clustering recognition, in the case of attenuation of the density and quantity of the far-distance point cloud, clustering in the velocity-radial distance dimension space is introduced. After obtaining the velocity and distance information of each point, it is mapped in a two-dimensional coordinate system. By judging the Euclidean distance and density between points, setting r2 to 0.5 and (MinPoints)2 to 20, the best clustering effect at a long distance can be obtained.

[0100] In some embodiments, such as Figure 4 for DBSCAN clustering that only uses the dimension information of the spatial coordinate system, it can be found that there is a loss of targets at a relatively long distance, that is, only three targets are recognized. Therefore, clustering is performed through the features of the velocity-radial distance dimension, and the distance threshold is set to 10m; that is, the clustering result of the spatial coordinate system dimension information is adopted at 0 - 10m (close distance), and the clustering result of the velocity-radial distance dimension is used when it is greater than 10m (long distance), as Figure 5 shown, it can be seen that the point cluster information of two targets is recognized in the distance.

[0101] Finally, the clustering results of the spatial coordinate system dimension information and the velocity-radial distance dimension are summarized, as Figure 6 shown, and finally the target recognition information within the entire field of view is obtained, which is consistent with the set true situation of the scene, that is, 4 targets set in the experiment process are recognized, improving the accuracy of point cloud target clustering recognition.

[0102] As can be seen from the above, for the clustering and recognition of point cloud targets within the radar's field of view, the present application makes full use of the feature information of the target points obtained by the radar. By adding the velocity-radial distance dimension for clustering, the accuracy of long-distance target recognition is improved, and it is fused with the clustering of the feature information of the spatial coordinate system used for short distances, greatly enhancing the accuracy of target recognition throughout the field of view.

[0103] It should be noted that the method of the embodiment of the present application can be executed by a single device, such as a computer or a server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In such a distributed scenario, one of the multiple devices can only execute one or more steps of the method of the embodiment of the present application, and these multiple devices will interact with each other to complete the described method.

[0104] It should be noted that some embodiments of the present application have been described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be executed in a different order from that in the above embodiments and still achieve the desired result. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0105] Based on the same inventive concept, corresponding to the method of any of the above embodiments, the present application also provides a target point cloud clustering device based on radar multi-dimensional feature information.

[0106] Refer to Figure 7 , the target point cloud clustering device based on radar multi-dimensional feature information includes:

[0107] The mixing module 702 is configured to determine the echo signal and the transmit signal of the radar, mix the echo signal with the transmit signal, and determine the intermediate frequency signal;

[0108] The signal processing module 704 is configured to determine the position information of any target point cloud through range-Doppler FFT processing according to the intermediate frequency signal;

[0109] The classification module 706 is configured to determine the radial distance between the target point cloud and the radar according to the position information, and classify the target point cloud according to the radial distance to determine the short-distance target point cloud and the long-distance target point cloud;

[0110] The first clustering module 708 is configured to cluster the short-distance target point cloud by a clustering method based on the spatial coordinate dimension to determine the short-distance point cloud clustering result;

[0111] The second clustering module 710 is configured to cluster the long-distance target point cloud through a clustering method based on the velocity radial distance dimension to determine the long-distance point cloud clustering result;

[0112] The fusion module 712 is configured to fuse the short-distance point cloud clustering result and the long-distance point cloud clustering result, and perform target recognition according to the fusion result.

[0113] Optionally, the signal processing module is further configured to:

[0114] Extract information from the intermediate frequency signal through the distance-Doppler FFT method to determine the distance information, velocity information, and azimuth information of the target point cloud;

[0115] Determine the position information according to the distance information and the azimuth information.

[0116] Optionally, the signal processing module is further configured to:

[0117] Perform clutter suppression processing and denoising processing on the distance information, the velocity information, and the azimuth information through 2D-CFAR.

[0118] Optionally, the classification module is further configured to:

[0119] In response to determining that the radial distance is not greater than a preset distance, determine the target point cloud as the short-distance target point cloud;

[0120] In response to determining that the radial distance is greater than the preset distance, determine the target point cloud as the long-distance target point cloud.

[0121] Optionally, the first clustering module is further configured to:

[0122] According to the position information of the short-distance target point cloud and the preset clustering area;

[0123] Screen the data point cloud in the preset clustering area according to the position information, and determine the data object to be clustered according to the screening result; wherein, the data object to be clustered includes a core object and a non-core object, and the screening includes: determining the relationship between any data point cloud and the short-distance target point cloud, and using the data point cloud with a relationship of density direct reach or density reachable or density connected as the data object to be clustered.

[0124] Optionally, the second clustering module is further configured to:

[0125] Determine the data object to be clustered according to the short-distance target point cloud and the preset clustering area; wherein, the data object to be clustered includes a core object and a non-core object;

[0126] According to the data object to be clustered, data clustering is performed by a clustering method based on the velocity radial distance dimension to determine the clustering result of the far-distance point cloud; wherein, the clustering method based on the velocity radial distance dimension is the DBSCAN clustering method of the velocity radial distance dimension.

[0127] For the convenience of description, when describing the above device, various modules are described separately according to their functions. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0128] The device in the above embodiment is used to implement the corresponding target point cloud clustering method based on radar multi-dimensional feature information in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0129] Based on the same inventive concept, corresponding to the method in any of the above embodiments, the present application further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the target point cloud clustering method based on radar multi-dimensional feature information in any of the above embodiments.

[0130] Figure 8 FIG. shows a more specific schematic diagram of the hardware structure of the electronic device provided in this embodiment. The device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. Among them, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are communicatively connected to each other inside the device through the bus 1050.

[0131] The processor 1010 may be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0132] The memory 1020 may be implemented in the form of a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0133] The input / output interface 1030 is used to connect to the input / output module to achieve information input and output. The input / output module can be configured as a component in the device (not shown in the figure), or can be externally connected to the device to provide corresponding functions. Among them, the input device can include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device can include a display, a speaker, a vibrator, an indicator light, etc.

[0134] The communication interface 1040 is used to connect to the communication module (not shown in the figure) to achieve communication interaction between this device and other devices. Among them, the communication module can achieve communication through wired means (such as USB, network cable, etc.), or can also achieve communication through wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0135] The bus 1050 includes a path for transmitting information between various components of the device (such as the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0136] It should be noted that although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040, and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device may also only include the components necessary to implement the solution of the embodiments of this specification, and do not necessarily include all the components shown in the figure.

[0137] The electronic device of the above embodiment is used to implement the corresponding target point cloud clustering method based on radar multi-dimensional feature information in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0138] Based on the same inventive concept, corresponding to the method of any of the above embodiments, the present application also provides a non-transitory computer-readable storage medium, and the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to make the computer execute the target point cloud clustering method based on radar multi-dimensional feature information as described in any of the foregoing embodiments.

[0139] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device.

[0140] The computer instructions stored in the storage medium of the above embodiment are used to cause the computer to execute the target point cloud clustering method based on radar multi-dimensional feature information as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0141] Based on the same inventive concept, corresponding to the target point cloud clustering method based on radar multi-dimensional feature information described in any of the above embodiments, the present disclosure also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of the computer to cause the computer and / or the processor to execute the target point cloud clustering method based on radar multi-dimensional feature information. Corresponding to the execution subjects corresponding to the steps in each embodiment of the target point cloud clustering method based on radar multi-dimensional feature information, the processor executing the corresponding steps can belong to the corresponding execution subject.

[0142] The computer program product of the above embodiment is used to cause the computer and / or the processor to execute the target point cloud clustering method based on radar multi-dimensional feature information as described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here.

[0143] Those of ordinary skill in the art should understand that the discussion of any of the above embodiments is only exemplary and is not intended to imply that the scope of the present application (including the claims) is limited to these examples; under the concept of the present application, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of brevity.

[0144] In addition, for simplicity of explanation and discussion, and so as not to make the embodiments of the present application difficult to understand, well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Further, the devices may be shown in block diagram form in order to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that details of the implementation of such block diagram devices are highly dependent on the platform on which the embodiments of the present application are to be implemented (i.e., these details should be entirely within the understanding of those skilled in the art). In cases where specific details (such as circuits) are set forth to describe exemplary embodiments of the present application, it will be apparent to those skilled in the art that the embodiments of the present application may be practiced without these specific details or with variations of these specific details. Accordingly, these descriptions should be regarded as illustrative rather than restrictive.

[0145] Although the present application has been described in connection with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art in light of the foregoing description. For example, other memory architectures (such as dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0146] Embodiments of the present application are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present application shall be included within the protection scope of the present application.

Claims

1. A target point cloud clustering method based on radar multi-dimensional feature information, characterized in that, For target recognition, including: Determine the echo signal and the transmitted signal of the radar, mix the echo signal with the transmitted signal, and determine the intermediate frequency signal; According to the intermediate frequency signal, through range-Doppler FFT processing, determine the position information of any target point cloud; According to the position information, determine the radial distance between the target point cloud and the radar, and classify the target point cloud according to the radial distance to determine the near-range target point cloud and the far-range target point cloud; Cluster the near-range target point cloud by a clustering method based on the spatial coordinate dimension to determine the near-range point cloud clustering result; Cluster the far-range target point cloud by a clustering method based on the velocity radial distance dimension to determine the far-range point cloud clustering result; Fuse the near-range point cloud clustering result and the far-range point cloud clustering result, and perform target recognition according to the fusion result.

2. The method according to claim 1, wherein The determining the position information by performing range-Doppler FFT processing according to the intermediate frequency signal includes: Extract information from the intermediate frequency signal by the range-Doppler FFT method to determine the range information, velocity information, and azimuth information of the target point cloud; Determine the position information according to the range information and the azimuth information.

3. The method according to claim 2, wherein The method further includes: Perform clutter suppression processing and denoising processing on the range information, the velocity information, and the azimuth information through 2D-CFAR.

4. The method according to claim 1, wherein The classifying the target point cloud according to the radial distance to determine the near-range target point cloud and the far-range target point cloud includes: In response to determining that the radial distance is not greater than a preset distance, determine the target point cloud as the near-range target point cloud; In response to determining that the radial distance is greater than the preset distance, determine the target point cloud as the far-range target point cloud.

5. The method according to claim 1, characterized in that The clustering the near-range target point cloud by a clustering method based on the spatial coordinate dimension to determine the near-range point cloud clustering result includes: According to the position information of the near-range target point cloud and a preset clustering domain; Screen the data point cloud in the preset clustering domain according to the position information, and determine the data object to be clustered according to the screening result; wherein, the data object to be clustered includes a core object and a non-core object, and the screening includes: determining the relationship between any data point cloud and the near-range target point cloud, and using the data point cloud with a relationship of density direct reach or density reachable or density connected as the data object to be clustered; Cluster the data according to the data object to be clustered by a clustering method based on the spatial coordinate dimension to determine the near-range point cloud clustering result; wherein, the clustering method based on the spatial coordinate dimension is the DBSCAN clustering method based on the spatial coordinate dimension.

6. The method according to claim 1, wherein The clustering the far-range target point cloud by a clustering method based on the velocity radial distance dimension to determine the far-range point cloud clustering result includes: According to the near-range target point cloud and a preset clustering domain, determine the data object to be clustered; wherein, the data object to be clustered includes a core object and a non-core object; Based on the data objects to be clustered, data clustering is performed through a clustering method based on the velocity radial distance dimension to determine the clustering result of the far-distance point cloud; wherein, the clustering method based on the velocity radial distance dimension is the DBSCAN clustering method for the velocity radial distance dimension.

7. A target point cloud clustering device based on radar multi-dimensional feature information, characterized in that, It includes: A mixing module configured to determine an echo signal and a transmission signal of a radar, mix the echo signal with the transmission signal, and determine an intermediate frequency signal; A signal processing module configured to determine the position information of any target point cloud through range-Doppler FFT processing according to the intermediate frequency signal; A classification module configured to determine the radial distance between the target point cloud and the radar according to the position information, and perform classification processing on the target point cloud according to the radial distance to determine a near-distance target point cloud and a far-distance target point cloud; A first clustering module configured to cluster the near-distance target point cloud through a clustering method based on the spatial coordinate dimension to determine the clustering result of the near-distance point cloud; A second clustering module configured to cluster the far-distance target point cloud through a clustering method based on the velocity radial distance dimension to determine the clustering result of the far-distance point cloud; A fusion module configured to fuse the clustering result of the near-distance point cloud and the clustering result of the far-distance point cloud, and perform target recognition according to the fusion result.

8. An electronic device, including a memory, a processor, and a computer program stored on the memory and executable by the processor, wherein the processor implements the method according to any one of claims 1 to 6 when executing the computer program.

9. A non-transitory computer-readable storage medium storing computer instructions for causing a computer to execute the method according to any one of claims 1 to 6.

10. A computer program product including computer program instructions, which when running on a computer, cause the computer to execute the method according to any one of claims 1 - 6.

Images