Millimeter wave radar environment rasterization target attribute identification method, equipment and medium

Abstract

The invention discloses a millimeter wave radar environment rasterization target attribute identification method, equipment and a medium, and belongs to the technical field of millimeter wave radar target detection. The method comprises the following steps: acquiring driving information of a vehicle and trace point data of a millimeter wave radar; according to the driving information, multiple frames of trace points in the trace point data are converted to a current vehicle body coordinate system; under the current vehicle body coordinate system, rasterizing the measurement area, and extracting grid features; and inputting the extracted features into a trained target attribute recognition model to obtain a predicted value of the grid target attribute. According to the method, the overall features of the trace points in the grids are summarized on the grid level, the grid environment features are constructed, the features in the grids and around the grids are fully extracted, the features are mined by using a deep learning method, and the relative relationship between the grid features and the target attributes is established, so that the features of the target can be effectively identified, and the target recognition efficiency is improved. And prior information is provided for follow-up target tracking, recognition and calculation of available areas.

Description

Technical Field

[0001] The present invention relates to the field of millimeter wave radar target detection technology, and in particular to a method, device and medium for identifying target attributes in a millimeter wave radar environment rasterization. Background Art

[0002] In recent years, vehicle-mounted millimeter-wave radar has developed rapidly, its performance has gradually improved, and the installation rate of millimeter-wave radar in various types of vehicles has also been continuously improved. In addition to completing the detection of moving targets, it can also detect obstacles in the environment through multi-frame accumulation, and is applied to the identification of movable areas, the construction of electronic maps, and the identification of parking spaces. However, since the detection results of millimeter-wave radar are interfered by ground reflection clutter, multipath reflection clutter, and frequency doubling clutter, and the point trace characteristics of the real target are also affected by environmental noise, process noise, angle measurement capability, resolution and other unfavorable factors, the identification of target point trace attributes and the elimination of false point traces by millimeter-wave radar have become difficult problems that need to be solved in current applications.

[0003] At present, in the field of target feature recognition of millimeter-wave radar, the target feature recognition is usually completed by using information from other sensors or simple feature extraction, threshold judgment and logic stacking. With the rapid development and widespread application of deep learning algorithms, there are also some methods that use the backbone network of deep learning itself to mine features, but the point cloud density of millimeter-wave radar is not as good as that of lidar, and its overall sample size is small. The training effect of deep learning models with more layers on small samples is not ideal, resulting in insufficient feature mining of its backbone network. The existing technical solution proposes a method for feature extraction based on grids, but it still uses the currently common mean extraction and feature pyramid network methods. It neither expands the point trace features nor extracts multi-factor interactive features, nor extracts grid environment features. Therefore, this method cannot fully mine the information carried by the point traces themselves and the grid local information brought by multi-frame accumulation. Summary of the invention

[0004] In order to solve at least one of the technical problems existing in the prior art to a certain extent, the purpose of the present invention is to provide a method, device and medium for rasterizing target attributes in a millimeter-wave radar environment based on deep learning.

[0005] The first technical solution adopted by the present invention is:

[0006] A method for identifying target attributes in a millimeter-wave radar environment by rasterization includes the following steps:

[0007] Obtain vehicle driving information and millimeter wave radar point data;

[0008] According to the driving information, multiple frames of point traces in the point trace data are converted to the current vehicle body coordinate system;

[0009] In the current vehicle body coordinate system, the measurement area is rasterized and raster features are extracted;

[0010] The extracted features are input into the trained target attribute recognition model to obtain the predicted values ​​of the raster target attributes.

[0011] Furthermore, the driving information includes the vehicle speed V car and yaw angular velocity Yaw;

[0012] The point data includes 11-dimensional features, namely, azimuth angle θ, distance Range, pitch angle φ, radial speed rSpeed, horizontal ground speed abSpeed, point coordinates x, y, z, reflection cross section RCS, amplitude amp, and signal-to-noise ratio SNR.

[0013] Furthermore, the step of converting multiple frames of point traces in the point trace data to the current vehicle body coordinate system according to the driving information includes:

[0014] Calculate the vehicle's motion displacement dx and dy between the previous frame and the current frame based on the driving information, and compensate for the vehicle's displacement and angle transformation for multiple frames of historical point data;

[0015] Among them, the motion displacement dx and dy are calculated as follows:

[0016] dx=V car *T*cos(Yaw*T / 2)

[0017] dy=V car *T*sin(Yaw*T / 2)

[0018] Where T is the frame period.

[0019] Furthermore, the step of rasterizing the measurement area and extracting raster features includes:

[0020] The measurement area is rasterized, multiple frames of historical points are traversed, and the relationship between the points and the grid is established according to the coordinates x and y of the points;

[0021] Traverse all grids, and perform point feature extraction and grid overall feature extraction for all points in each grid.

[0022] Furthermore, the point trace features are extracted in the following manner:

[0023] First, the 11 features of each point in the grid are discretized into single variables, and the continuous variables are rounded down to obtain 11 new discretized features;

[0024] Secondly, multi-factor feature extraction is performed based on the radiation-reflection characteristics of the radar: M features to be interacted are selected, assuming that the dimension of the interactive features is n, where n≤M. First, a variable is selected from the M features to be interacted as the numerator, and then n-1 variables are selected from the M-1 features to be interacted and multiplied as the denominator, so there are a total of There are permutations and combinations; the value of n ranges from 2 to M, so a total of MultiVarNum multi-factor features will be created:

[0025]

[0026] Furthermore, the overall grid features are extracted in the following manner:

[0027] After extracting the features of a single point in the grid, perform feature extraction of mean, variance, standard deviation, maximum, minimum, median, and kurtosis on all features of multiple points in the grid.

[0028] The local environmental features around the grid are obtained, and the environmental features include grid connectivity and grid similarity; wherein the grid connectivity is used to measure the isolation degree of the grid, and the grid similarity is used to measure the similarity between the grid and the surrounding grids.

[0029] Furthermore, the calculation formula of the grid connectivity is as follows:

[0030]

[0031] Where N is the number of grids around the i-th grid; TotalNum is the total number of points in the grid; Bool(TotalNum j >0) indicates whether the total number of points in the surrounding j-th grid is greater than 0. If it is true, the result of the expression is 1, otherwise it is 0;

[0032] The calculation formula of the grid similarity is as follows:

[0033]

[0034] Where, Value i,k is the value of feature k of the i-th grid, Value j,k is the value of feature k of the j-th grid.

[0035] Furthermore, the target attribute recognition model is a deep learning model, and the target attribute recognition model is trained in the following manner:

[0036] The point data collected by the millimeter-wave radar is labeled, the grid is compared with the actual target, the corresponding grid attributes are marked according to the attributes of the target, and a training set is constructed;

[0037] The target attribute recognition model is trained using a training set; the input of the target attribute recognition model is the extracted point trace features of each grid and the overall features of the grid, and the output is the correlation between the attribute features of the grid and the target attributes of the grid.

[0038] The second technical solution adopted by the present invention is:

[0039] An electronic device comprises a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, the code set or the instruction set is loaded and executed by the processor to implement a millimeter-wave radar environment rasterized target attribute recognition method as described above.

[0040] The third technical solution adopted by the present invention is:

[0041] A computer-readable storage medium stores at least one instruction, at least one program, a code set or an instruction set, wherein the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by a processor to implement a method for identifying target attributes in a millimeter-wave radar environment by rasterization as described above.

[0042] The fourth technical solution adopted by the present invention is:

[0043] A computer program product or a computer program includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the above method.

[0044] The beneficial effects of the present invention are as follows: the present invention summarizes the overall characteristics of the points in the grid at the grid level, constructs the grid environment characteristics, fully extracts the characteristics in and around the grid, and uses deep learning methods to further mine the characteristics and establish the corresponding relationship between the grid characteristics and the target attributes, which can effectively identify the characteristics of the target and improve the prior information for subsequent target tracking, identification and calculation of the executable area. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the embodiments of the present invention or the drawings of related technical solutions in the prior art are introduced below. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0046] Figure 1 It is a flowchart of a method for rasterizing target attributes of a millimeter-wave radar based on deep learning in an embodiment of the present invention;

[0047] Figure 2 It is a flowchart of the steps of a method for identifying target attributes in a millimeter-wave radar environment rasterization according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limitations of the present invention. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0049] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0050] In the description of the present invention, "several" means one or more, "more" means more than two, "greater than", "less than", "exceed" etc. are understood as not including the number itself, and "above", "below", "within" etc. are understood as including the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0051] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.

[0052] In view of the existing technical problems, the present invention proposes a rasterized target attribute recognition method for millimeter-wave radar based on deep learning. First, the method processes the received millimeter-wave radar point trace data, and uses the dead reckoning method based on the vehicle driving information to compensate for the position deviation of the multi-frame historical point traces between the previous frame and the current frame. Secondly, at the point trace feature level, the discretization method and the multi-feature interaction method are combined to create multiple new features, fully explore the inherent laws of the point traces, and enhance the robustness of the point trace features. Then, at the grid level, the overall characteristics of the point traces in the grid are summarized based on a variety of statistical methods, and a variety of grid environmental features are constructed to fully extract the features in and around the grid. Finally, the deep learning method is used to further explore the features and establish the corresponding relationship between the grid features and the target attributes, so that the trained deep learning model can effectively identify the characteristics of the target, and improve the prior information for subsequent target tracking, recognition and calculation of the executable area.

[0053] Example 1

[0054] like Figure 1 and Figure 2 As shown, this embodiment provides a method for identifying target attributes in a millimeter-wave radar environment by rasterization, comprising the following steps:

[0055] S1. Obtain vehicle driving information and millimeter wave radar point data.

[0056] In some embodiments, existing sensors may be used to collect and obtain vehicle driving information, wherein the driving information includes the vehicle speed V car and yaw angular velocity Yaw.

[0057] In some embodiments, the vehicle-mounted millimeter-wave radar scans the environment and obtains millimeter-wave radar echo data. First, the target point information is obtained through millimeter-wave radar signal processing, and then the target point is projected in the direction of the vehicle's heading angle, and the current frame point is converted from the polar coordinate system to the Cartesian coordinate system to obtain the final point data, including the point azimuth θ, distance Range, pitch angle φ, radial speed rSpeed, horizontal ground speed abSpeed, point coordinates x, y, z, reflection cross section RCS, amplitude amp, signal-to-noise ratio SNR, a total of 11 dimensional features.

[0058] S2. According to the driving information, multiple frames of point traces in the point trace data are converted to the current vehicle body coordinate system.

[0059] After obtaining the vehicle's driving information and the millimeter-wave radar's point data, the multi-frame point information is converted to the current vehicle body coordinate system through coordinate conversion and DR dead reckoning. Specifically, the vehicle's motion displacement dx and dy between the previous frame and the current frame are calculated, and the multi-frame historical point data is compensated for the vehicle's displacement and angle transformation. The calculation of dx and dy is as follows:

[0060] dx=V car *T*cos(Yaw*T / 2)

[0061] dy=V car *T*sin(Yaw*T / 2)

[0062] Where T is the frame period.

[0063] S3. In the current vehicle body coordinate system, the measurement area is rasterized and raster features are extracted.

[0064] In some embodiments, step S3 specifically includes the following steps:

[0065] S31: Gridding. The entire measurement area is gridded, multiple frames of historical points are traversed, and the relationship between the points and the grid is established according to the point coordinates x and y.

[0066] S32: Grid feature extraction. Traverse all grids, and for all points in each grid, extract the overall features of the grid using the grid feature extraction method.

[0067] As an optional implementation, the grid features include point trace features and grid overall features, and the specific extraction method is as follows:

[0068] A1. Point feature extraction

[0069] First, the 11 features of each point in the grid are discretized into single variables, and the continuous variables are rounded down to obtain 11 new discretized features. This process can enhance the robustness of the original features.

[0070] Secondly, multi-factor feature extraction is performed based on the radar's illumination-reflection characteristics. First, M features to be interacted are screened out, and the following four features are selected (M=4): reflection cross section RCS, amplitude amp, signal-to-noise ratio SNR and distance Range. The combination of the above features can effectively mine the intrinsic feature logic of millimeter-wave radar detection points. The extraction method of multi-factor interactive features is as follows: Assuming that the dimension of the interactive feature is n-dimensional (n≤M), first select a variable from the M features to be interacted as the numerator, and then select n-1 variables from the M-1 features to be interacted and multiply them as the denominator, then there are a total of There are multiple permutations and combinations. The value of n ranges from 2 to M, so a total of MultiVarNum multi-factor features will be created:

[0071]

[0072] A2. Grid overall feature extraction

[0073] Based on the extraction of single point features in the grid, the overall features of the grid and the surrounding environment features of the grid are extracted.

[0074] First, the mean, variance, standard deviation, maximum, minimum, median and kurtosis of all features of multiple points in the grid are extracted. In addition, the total number of points in the grid, TotalNum, is calculated to obtain the overall characteristics of a single grid point.

[0075] Secondly, based on the above, the local environmental features around the grid are obtained, and the following environmental features are constructed:

[0076] 1) Grid connectivity, which is used to measure the degree of grid isolation; a smaller connectivity means that the grid is more likely to be an object edge, an isolated false target, or a small target. The expression is as follows:

[0077]

[0078] Where; N is the number of grids around the i-th grid; Bool(TotalNum j >0) indicates whether the total number of points in the surrounding j-th grid is greater than 0. If it is true, the result of the expression is 1, otherwise it is 0.

[0079] 2) Grid similarity, which is used to measure the similarity between a grid and its surrounding grids. The higher the similarity, the greater the probability of being the same target. The expression is as follows:

[0080]

[0081] Among them, Value i,k is the value of feature k of the i-th grid.

[0082] S4. Input the extracted features into the trained target attribute recognition model to obtain the predicted value of the raster target attribute.

[0083] In the embodiment, the point data collected in real time by the millimeter wave radar in each frame is rasterized and raster features are extracted. The extracted features are used as input to the trained deep learning model to obtain and output the predicted values ​​of the raster target attributes, providing prior information for subsequent target tracking and recognition.

[0084] As an optional implementation, the target attribute recognition model is a deep learning model, which is trained in the following way: First, the data collected by the millimeter wave radar is annotated, the grid is compared with the actual target, and the corresponding grid attributes are marked according to the attributes of the target. If the target of the grid is an object that affects the driving of the vehicle (such as a vehicle, a stone pier, a no-parking sign, a railing, etc.), it is marked as a "high" attribute. If the target does not affect the driving of the vehicle (such as a fallen ground lock, a manhole cover, etc.), it is marked as a "low" attribute. If there is no target at the position corresponding to the point trace, it is marked as a "false" attribute. Secondly, a deep learning model is used for training, and its input is the extracted grid features and environmental features of each grid, and the output is the attributes of the grid ("high", "low" and "false" attributes), so as to establish the association between the grid features and the grid target attributes.

[0085] Example 2

[0086] An embodiment of the present invention further provides an electronic device, the electronic device comprising a processor and a memory, the memory storing at least one instruction, at least one program, a code set or an instruction set, the at least one instruction, the at least one program, the code set or the instruction set being loaded and executed by the processor to implement the following Figure 1 and Figure 2 A method for rasterizing target attributes in a millimeter-wave radar environment is shown.

[0087] It is understood that the memory may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory includes a non-transitory computer-readable storage medium. The memory may be used to store instructions, programs, codes, code sets, or instruction sets. The memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function, instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data created according to the use of the server, etc.

[0088] The processor may include one or more processing cores. The processor uses various interfaces and lines to connect the various parts of the entire server, and executes various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory, and calling data stored in the memory. Optionally, the processor can be implemented in at least one hardware form of digital signal processing (DSP), field programmable gate array (FPGA), and programmable logic array (PLA). The processor can integrate one or a combination of a central processing unit (CPU) and a modem. Among them, the CPU mainly processes the operating system and application programs; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor, but implemented separately through a chip.

[0089] Since the electronic device is an electronic device corresponding to a millimeter-wave radar environment rasterized target attribute recognition method of an embodiment of the present invention, and the principle of solving the problem by the electronic device is similar to that of the method, the implementation of the electronic device can refer to the implementation process of the above-mentioned method embodiment, and the repeated parts will not be repeated.

[0090] Example 3

[0091] The embodiment of the present invention further provides a computer-readable storage medium, wherein the storage medium stores at least one instruction, at least one program, code set or instruction set, and the at least one instruction, the at least one program, the code set or instruction set is loaded and executed by a processor to implement the following Figure 1 and Figure 2 A method for rasterizing target attributes in a millimeter-wave radar environment is shown.

[0092] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0093] Since the storage medium is a storage medium corresponding to a millimeter-wave radar environment rasterized target attribute recognition method of an embodiment of the present invention, and the principle of solving the problem by the storage medium is similar to that of the method, the implementation of the storage medium can refer to the implementation process of the above-mentioned method embodiment, and the repeated parts will not be repeated.

[0094] Example 4

[0095] In some possible implementations, various aspects of the method of the embodiment of the present invention may also be implemented in the form of a program product, which includes a program code. When the program product is run on a computer device, the program code is used to enable the computer device to execute the steps of a method for identifying target attributes by rasterizing a millimeter-wave radar environment according to various exemplary embodiments of the present application described above in this specification. Among them, the executable computer program code or "code" for executing each embodiment may be written in a high-level programming language such as C, C++, C#, Smalltalk, Java, JavaScript, Visual Basic, structured query language (e.g., Transact-SQL), Perl, or in various other programming languages.

[0096] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0097] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0098] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made based on the essence of the content of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying target attributes in a millimeter-wave radar environment by rasterization, characterized in that: The following steps are involved: Obtain vehicle driving information and millimeter wave radar point data; According to the driving information, multiple frames of point traces in the point trace data are converted to the current vehicle body coordinate system; In the current vehicle body coordinate system, the measurement area is rasterized and raster features are extracted; The extracted features are input into the trained target attribute recognition model to obtain the predicted values ​​of the raster target attributes.

2. The method for identifying target attributes in a millimeter-wave radar environment by rasterization according to claim 1, characterized in that: The driving information includes the vehicle speed V car And the yaw rate Yaw: The point data includes 11-dimensional features, namely, azimuth angle θ, distance Range, pitch angle φ, radial speed rSpeed, horizontal ground speed abSpeed, point coordinates x, y, z, reflection cross section RCS, amplitude amp, and signal-to-noise ratio SNR.

3. A method for identifying target attributes by rasterization in a millimeter-wave radar environment according to claim 1 or 2, characterized in that: The method of converting multiple frames of point traces in the point trace data into the current vehicle body coordinate system according to the driving information includes: Calculate the vehicle's motion displacement dx and dy between the previous frame and the current frame based on the driving information, and compensate for the vehicle's displacement and angle transformation for multiple frames of historical point data; Among them, the motion displacement dx and dy are calculated as follows: dx=V car *T*cos(Yaw*T / 2) dy=V car *T*sin(You*T / 2) Where T is the frame period.

4. The method for identifying target attributes in a millimeter wave radar environment by rasterization according to claim 1, characterized in that: The process of rasterizing the measurement area and extracting raster features includes: The measurement area is rasterized, multiple frames of historical points are traversed, and the relationship between the points and the grid is established according to the coordinates x and y of the points; Traverse all grids, and perform point feature extraction and grid overall feature extraction for all points in each grid.

5. The method for identifying target attributes in a millimeter wave radar environment by rasterization according to claim 4, characterized in that: The point trace features are extracted in the following way: Perform univariate discretization processing on the features of each point in the grid to obtain new discretized features; According to the characteristics of radar illumination-reflection, multi-factor feature extraction is performed: M interactive features are selected, assuming that the dimension of the interactive features is n, where n≤M. First, a variable is selected from the M interactive features as the numerator, and then n-1 variables are selected from the M-1 interactive features and multiplied as the denominator, so there are a total of There are permutations and combinations; the value of n ranges from 2 to M, so a total of MultiVarNum multi-factor features will be created:

6. The method for identifying target attributes in a millimeter wave radar environment by rasterization according to claim 4, characterized in that: The overall grid features are extracted in the following way: After extracting the features of a single point in the grid, perform feature extraction of mean, variance, standard deviation, maximum, minimum, median, and kurtosis on all features of multiple points in the grid. The local environmental features around the grid are obtained, and the environmental features include grid connectivity and grid similarity; wherein the grid connectivity is used to measure the isolation degree of the grid, and the grid similarity is used to measure the similarity between the grid and the surrounding grids.

7. The method for identifying target attributes in a millimeter wave radar environment by rasterization according to claim 6, characterized in that: The calculation formula of the grid connectivity is as follows: Where N is the number of grids around the i-th grid; TotalNum is the total number of points in the grid; Bool(TotalNum j >0) indicates whether the total number of points in the surrounding j-th grid is greater than 0. If it is true, the result of the expression is 1, otherwise it is 0; The calculation formula of the grid similarity is as follows: Where, Value i，k is the value of feature k of the i-th grid, Value j，k is the value of feature k of the j-th grid.

8. The method for identifying target attributes in a millimeter wave radar environment by rasterization according to claim 1, characterized in that: The target attribute recognition model is a deep learning model, and the target attribute recognition model is trained in the following way: The point data collected by the millimeter-wave radar is labeled, the grid is compared with the actual target, the corresponding grid attributes are marked according to the attributes of the target, and a training set is constructed; The target attribute recognition model is trained using the training set; The input of the target attribute recognition model is the extracted point trace features of each grid and the overall features of the grid, and the output is the correlation between the attribute features of the grid and the target attributes of the grid.

9. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the method described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that: The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the method according to any one of claims 1 to 8.

Images