Target detection method and device based on vehicle-mounted radar, storage medium and vehicle

Abstract

The invention discloses a target detection method and device based on a vehicle-mounted radar, a storage medium and a vehicle, and the method comprises the steps: obtaining a point cloud data set of an environment where the vehicle is located through the vehicle-mounted radar; dividing the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; determining a static boundary model according to the static point cloud data subset; screening the dynamic point cloud data subset according to a static boundary model to determine a detection target; and determining the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target. Therefore, the target detection capability, the safety performance and the environmental adaptability of the vehicle can be remarkably improved, so that the driving experience of a user is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent driving control, and particularly relates to a target detection method based on an in-vehicle radar, a computer-readable storage medium, a target detection device based on an in-vehicle radar, and a vehicle. Background Art

[0002] In the rapid development of autonomous driving technology, the environmental perception system, as the core key technology for realizing vehicle autonomous navigation and safe driving, relies on a variety of high-precision sensors such as lidar, ultrasonic radar, cameras, and their fusion target detection technologies to comprehensively understand and respond to complex road traffic environments. These sensor technologies can capture and analyze the dynamic information of surrounding pedestrians, vehicles, and other traffic participants in real time, providing decision-making support for autonomous driving vehicles, aiming to ensure the safety of all road users.

[0003] However, in the face of the diversity and complexity of the actual road environment, there are still significant challenges. In particular, the target occlusion problem has become one of the key factors restricting the effectiveness of the environmental perception system. For example, in urban blocks with high-rise buildings or on highways with dense traffic, obstacles such as buildings and other moving vehicles frequently occlude target objects, resulting in the sensor being unable to continuously and effectively detect the position, speed, and movement trajectory of the occluded target, thus greatly increasing the collision risk and potential safety hazards. Summary of the Invention

[0004] The present invention aims to at least partly solve one of the technical problems in the related technologies. To this end, the first object of the present invention is to propose a target detection method based on an in-vehicle radar, which can significantly improve the vehicle's target detection ability, safety performance, and environmental adaptability, thereby enhancing the user's driving experience.

[0005] The second object of the present invention is to propose a computer-readable storage medium.

[0006] The third object of the present invention is to propose a target detection device based on an in-vehicle radar.

[0007] The fourth object of the present invention is to propose a vehicle.

[0008] To achieve the above object, the first aspect of the present invention provides a target detection method based on an in-vehicle radar, wherein the method includes: obtaining a point cloud data set of the vehicle's environment through the in-vehicle radar; dividing the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; determining a static boundary model according to the static point cloud data subset; screening the dynamic point cloud data subset according to the static boundary model to determine a detection target; and determining the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target.

[0009] According to the target detection method based on in-vehicle radar according to an embodiment of the present invention, a point cloud data set of the environment where the vehicle is located is obtained through the in-vehicle radar, and the point cloud data set is divided into a static point cloud data subset and a dynamic point cloud data subset, and a static boundary model is determined according to the static point cloud data subset, and then the dynamic point cloud data subset is screened according to the static boundary model to determine a detection target, and thus the motion state of the detection target is determined according to the dynamic point cloud data corresponding to the detection target. Thereby, the target detection ability, safety performance and environmental adaptability of the vehicle can be significantly improved, thereby improving the driving experience of users.

[0010] In addition, the target detection method based on in-vehicle radar according to the above embodiment of the present invention may further include the following additional technical features:

[0011] According to an embodiment of the present invention, the in-vehicle radar is a 4D millimeter-wave radar.

[0012] According to an embodiment of the present invention, each data point in the point cloud data set includes at least one of time information, speed information, position information, and signal strength information.

[0013] According to an embodiment of the present invention, the point cloud data set is divided into the static point cloud data subset and the dynamic point cloud data subset based on the speed information.

[0014] According to an embodiment of the present invention, the method further includes: determining the inlier data ratio of the static boundary model in the static point cloud data subset based on the position information; obtaining the fitting data amount and fitting accuracy of the static boundary model; determining the fitting iteration number of the static boundary model according to the inlier data ratio, the fitting data amount, and the fitting accuracy.

[0015] According to an embodiment of the present invention, screening the dynamic point cloud data subset according to the static boundary model includes: performing clustering processing on the dynamic point cloud data subset to obtain multiple clusters of clustering data; determining the position information of multiple targets to be confirmed based on the center points of the multiple clusters of clustering data; verifying the position information according to the static boundary model to filter out non-detection targets among the targets to be confirmed.

[0016] According to an embodiment of the present invention, the method further includes: obtaining the shape information of the multiple targets to be confirmed; filtering out non-detection targets among the multiple targets to be confirmed according to the shape information and the position information.

[0017] According to an embodiment of the present invention, determining the motion state of the detection target based on the dynamic point cloud data corresponding to the detection target includes: obtaining the position information, speed information, and signal strength information of the dynamic point cloud data corresponding to the detection target; and determining the motion state of the detection target according to the position information, the speed information, and the signal strength information.

[0018] To achieve the above object, an embodiment of the second aspect of the present invention provides a computer-readable storage medium, on which a target detection program based on an in-vehicle radar is stored. When the target detection program based on the in-vehicle radar is executed by a processor, the target detection method based on the in-vehicle radar in the foregoing embodiments of the present invention is implemented.

[0019] According to the computer-readable storage medium of the embodiment of the present invention, by executing the target detection program based on the in-vehicle radar through a processor, the target detection ability, safety performance, and environmental adaptability of the vehicle can be significantly improved, thereby enhancing the driving experience of users.

[0020] To achieve the above object, an embodiment of the third aspect of the present invention provides a target detection device based on an in-vehicle radar. The device includes: an acquisition module, configured to acquire a point cloud data set of the environment where the vehicle is located through the in-vehicle radar; a processing module, configured to divide the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; a determination module, configured to determine a static boundary model according to the static point cloud data subset; the determination module is further configured to screen the dynamic point cloud data subset according to the static boundary model to determine a detection target; and the determination module is further configured to determine the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target.

[0021] According to the target detection device based on the in-vehicle radar of the embodiment of the present invention, the acquisition module acquires a point cloud data set of the environment where the vehicle is located through the in-vehicle radar, and the processing module divides the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset, and the determination module determines a static boundary model according to the static point cloud data subset, and then the determination module screens the dynamic point cloud data subset according to the static boundary model to determine a detection target, so as to determine the motion state of the detection target by the determination module according to the dynamic point cloud data corresponding to the detection target, which can significantly improve the target detection ability, safety performance, and environmental adaptability of the vehicle, thereby enhancing the driving experience of users.

[0022] To achieve the above object, an embodiment of the fourth aspect of the present invention provides a vehicle, which includes the target detection device based on the in-vehicle radar in the foregoing embodiments of the present invention.

[0023] A vehicle according to an embodiment of the present invention, by adopting the on-vehicle radar-based target detection device of the above embodiment of the present invention, can significantly improve the vehicle's target detection ability, safety performance and environmental adaptability, thereby enhancing the user's driving and riding experience.

[0024] Additional aspects and advantages of the present invention will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flowchart of an on-vehicle radar-based target detection method according to an embodiment of the present invention;

[0026] Figure 2 is a block diagram of an on-vehicle radar-based target detection device according to an embodiment of the present invention;

[0027] Figure 3 is a block diagram of a vehicle according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] Embodiments of the present invention will be described in detail below. Examples of the embodiments are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention, and should not be construed as limiting the present invention.

[0029] The on-vehicle radar-based target detection method, computer-readable storage medium, on-vehicle radar-based target detection device and vehicle according to embodiments of the present invention will be described below with reference to the drawings.

[0030] Figure 1 is a flowchart of an on-vehicle radar-based target detection method according to an embodiment of the present invention.

[0031] Specifically, as Figure 1 shown, the on-vehicle radar-based target detection method includes:

[0032] S101, obtaining a point cloud data set of the environment where the vehicle is located through an on-vehicle radar.

[0033] Specifically, in this embodiment, the on-vehicle radar may preferably be a 4D millimeter-wave radar, and a point cloud data set of the environment where the vehicle is located can be obtained through the 4D millimeter-wave radar. Each data point in the point cloud data set includes at least one of time information, speed information, position information, and signal strength information. In addition, the present invention may not specifically limit the types of information included in each data point in the point cloud data set.

[0034] S102. Divide the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset.

[0035] Specifically, in this embodiment, the point cloud data set can be divided into a static point cloud data subset and a dynamic point cloud data subset based on speed information. Among them, the static point cloud data subset contains data points with speeds close to zero or almost unchanged, and these data points correspond to static obstacles in the road environment, such as buildings, trees, road signs, etc. The dynamic point cloud data subset contains data points with obvious speed characteristics, and these data points correspond to dynamic obstacles such as moving pedestrians, vehicles, animals, etc.

[0036] It should be noted that before dividing the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset, the original data of the vehicle's environment obtained by the vehicle-mounted radar is also converted from the vehicle-mounted radar's own coordinate system to the vehicle body coordinate system. And according to the timestamp of each frame of point cloud data, the system will select multiple consecutive frames of point cloud data for time synchronization and spatial registration to eliminate the point cloud misalignment caused by the vehicle's own movement or the radar sampling interval. Finally, it is integrated into a continuous, consistent and time-synchronized point cloud sequence, that is, the point cloud data set.

[0037] S103. Determine the static boundary model according to the static point cloud data subset.

[0038] Specifically, in this embodiment, the system first gives the model N to be fitted. The model N to be fitted is the target model that needs to be fitted based on the static point cloud data subset, and its form is diverse, including but not limited to planes, spheres, cylinders, etc. The specific selection depends on the requirements of the actual application scenario, and this embodiment does not specifically limit this. Then, the system obtains the static point cloud data subset. This subset is a set composed of multiple frames of static point cloud data. These static point cloud data usually come from sensors such as lidar and depth cameras, representing the three-dimensional position information of fixed objects in the environment. Then, the system randomly selects n data points from the static point cloud data subset as samples. The value of n is usually determined according to the complexity of the model N to be fitted and the required accuracy. Random sampling can avoid the excessive influence of outliers or noise in the data on the fitting result, and at the same time improve the robustness of the algorithm. Finally, the system uses the n randomly sampled data points to determine the static boundary model. The specific method depends on the type of the model N to be fitted. For example, if the model N to be fitted is a plane model, the static boundary model of the plane can be expressed as Ax + By + Cz + D = 0, where the parameter A, the parameter B, and the parameter C are the normal vector components of the plane, and the parameter D is the distance from the plane to the origin. The least squares method can be used to fit the plane, and by minimizing the sum of the squares of the perpendicular distances from the data points to the fitted plane, the parameters A, B, C, and D are solved, so as to obtain the static boundary model of the plane.

[0039] S104. Screen the dynamic point cloud data subset according to the static boundary model to determine the detection target.

[0040] Specifically, in this embodiment, the system first performs clustering processing on the dynamic point cloud data subset. Since the data points in the dynamic point cloud data subset usually correspond to dynamic obstacles such as moving pedestrians, vehicles, and animals, these data points will show a certain degree of aggregation in space. Therefore, through clustering processing, these data points with similar motion characteristics and spatial positions can be divided into the same cluster, thereby obtaining multiple clusters of clustered data. For each cluster of clustered data, calculate its center point, which can be regarded as the position information of the corresponding dynamic obstacle, and then the position information of multiple targets to be confirmed can be determined. The system verifies these position information according to the static boundary model. If the position information of a certain target to be confirmed meets the constraint conditions of the static boundary model (for example, located inside or near the boundary of the static obstacle), it is considered that this target may be a non-detection target (such as a stationary object or a misdetected noise point), and it is filtered out. On the contrary, if there is an obvious deviation between the position information of a certain target to be confirmed and the static boundary model, it is considered that this target is very likely to be a detection target (such as a moving pedestrian, vehicle, etc.). In addition, the shape information (such as size, contour, etc.) of multiple targets to be confirmed can also be obtained and screened together with the position information. For example, if the shape information of a certain target to be confirmed does not match the shape characteristics of common dynamic obstacles (such as being too large, too small, or having an irregular shape), even if its position information meets the constraint conditions of the static boundary model, it will be determined as a non-detection target and filtered out. Furthermore, the dynamic point cloud data subset can be screened according to the static boundary model to determine the detection target.

[0041] S105. Determine the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target.

[0042] Specifically, in this embodiment, the position information, velocity information, and signal strength information of the dynamic point cloud data corresponding to the detection target are obtained. For example, the three-dimensional coordinates (X, Y, Z) of each point are extracted from the dynamic point cloud data, and these coordinates represent the position information of the detection target in space. By using the change in the position of the detection target between adjacent frames and combining the time interval, the velocity information of the detection target is calculated. The velocity information includes linear velocity (i.e., the moving speed in a certain direction) and / or angular velocity (i.e., the rotational speed around a certain point), which specifically depends on the motion characteristics of the detection target. The signal strength information can be obtained through a signal strength sensor. Furthermore, the motion state of the detection target can be determined based on the position information, velocity information, and signal strength information. Among them, the motion state includes a stationary state, a uniform motion state, an accelerating / decelerating motion, a complex motion state, etc. For example, if the position information corresponding to the detection target remains almost unchanged in multiple consecutive frames and the velocity information is close to zero, it is determined that the motion state of the detection target is a stationary state. If the position information corresponding to the detection target changes linearly with time and the velocity information remains constant, it is determined that the motion state of the detection target is a uniform motion state.

[0043] Furthermore, in some embodiments of the present invention, the vehicle-mounted radar is a 4D millimeter-wave radar. Thus, the detection accuracy of the point cloud data set of the vehicle's environment can be significantly improved, and further, the vehicle's target detection ability, safety performance, and environmental adaptability can be improved, thereby enhancing the user's driving experience.

[0044] Furthermore, in some embodiments of the present invention, the method further includes: determining the proportion of inlier data of the static boundary model in the static point cloud data subset based on the position information; obtaining the fitting data volume and fitting accuracy of the static boundary model; and determining the fitting iteration times of the static boundary model according to the proportion of inlier data, the fitting data volume, and the fitting accuracy.

[0045] Specifically, in this embodiment, a threshold range (such as the X, Y, Z coordinate range) can be set. Among the data points in the static point cloud data subset, if the three-dimensional coordinates of the data point are within the threshold range, it is regarded as an inlier. Then, the static point cloud data subset can be traversed to calculate the number of data points that meet the threshold range (i.e., inliers), and based on this, the proportion of inlier data is calculated, that is, the proportion of inlier data in the total static point cloud data subset. Furthermore, the static boundary model can be fitted, and the number of point clouds participating in the fitting, that is, the fitting data volume, can be recorded. Among them, the fitting data volume can be set by the user, and the fitting accuracy is obtained. Among them, the user can set the fitting accuracy according to actual needs. After obtaining the proportion of inlier data, the fitting data volume, and the fitting accuracy, the fitting iteration times of the static boundary model can be determined according to the proportion of inlier data, the fitting data volume, and the fitting accuracy. The calculation formula is as follows:

[0046]

[0047] Among them, k represents the number of fitting iterations of the static boundary model, P represents the fitting accuracy of the static boundary model, t represents the proportion of inlier data of the static boundary model in the static point cloud data subset, and S represents the amount of fitting data. Furthermore, after fitting through the number of fitting iterations, the accuracy of the static boundary model can be improved.

[0048] In addition, in this embodiment, during each iteration, a specified number of points (i.e., the amount of fitting data) can be randomly selected from the static point cloud data subset to calculate the model, evaluate the fitting result of each iteration, calculate the fitting accuracy, and compare it with the fitting accuracy set by the user. According to the actual usage of the proportion of inlier data, the amount of fitting data, and the evaluation result of the fitting accuracy, the number of iterations can be dynamically adjusted. If the proportion of inlier data is high, the usage efficiency of the amount of fitting data is high, and the fitting accuracy is high, the number of iterations can be appropriately reduced; otherwise, the number of iterations needs to be increased. Thus, the accuracy and stability of the model can be improved.

[0049] Furthermore, in some embodiments of the present invention, screening the dynamic point cloud data subset according to the static boundary model includes: performing clustering processing on the dynamic point cloud data subset to obtain multiple clusters of clustering data; determining the position information of multiple targets to be confirmed based on the center points of the multiple clusters of clustering data; and verifying the position information according to the static boundary model to filter out non-detection targets among the targets to be confirmed.

[0050] Specifically, in this embodiment, since the data points in the dynamic point cloud data subset usually correspond to dynamic obstacles such as moving pedestrians, vehicles, and animals, these data points will show a certain degree of aggregation in space. Therefore, these data points with similar motion characteristics and spatial positions can be divided into the same cluster through clustering processing, thereby obtaining multiple clusters of clustering data. For each cluster of clustering data, calculate its center point, which can be regarded as the position information of the corresponding dynamic obstacle. Furthermore, the position information of multiple targets to be confirmed can be determined. Verify the position information according to the static boundary model. For example, the non-detection target can be judged according to the geometric relationship between the multipath target and the static boundary. Calculate the distance from the center point of the dynamic target (the position information of multiple targets to be confirmed) to the static boundary plane, and compare this distance with a preset threshold. If the distance is less than the preset threshold, it is considered that the target is a false target generated by multipath effect, that is, a non-detection target, so that the non-detection targets among the targets to be confirmed can be filtered out.

[0051] Furthermore, in some embodiments of the present invention, the method further includes: obtaining the shape information of multiple targets to be confirmed; and filtering out non-detection targets among the multiple targets to be confirmed according to the shape information and the position information.

[0052] Specifically, in this embodiment, the shape information of multiple targets to be confirmed can be obtained. The shape information includes information such as size and contour. Non-detection targets among the multiple targets to be confirmed can be filtered according to the shape information and position information. For example, if the shape information of a certain target to be confirmed does not match the shape characteristics of common dynamic obstacles (such as being too large, too small, or having an irregular shape), even if its position information meets the constraint conditions of the static boundary model, it will be identified as a non-detection target and filtered out. If the shape information of a certain target to be confirmed matches the shape characteristics of common dynamic obstacles, but according to the position information, it is determined that the target to be confirmed is outside the preset height range, then the target to be confirmed is identified as a non-detection module and filtered out. The preset height range can be a height range extending 3 meters vertically upward from the ground, and no specific limitation is made here.

[0053] Further, in some embodiments of the present invention, determining the motion state of a detection target according to the dynamic point cloud data corresponding to the detection target includes: obtaining the position information, speed information, and signal strength information of the dynamic point cloud data corresponding to the detection target; determining the motion state of the detection target according to the position information, speed information, and signal strength information.

[0054] Specifically, in this embodiment, the position information, speed information, and signal strength information of the dynamic point cloud data corresponding to the detection target are obtained. For example, the three-dimensional coordinates (X, Y, Z) of each point are extracted from the dynamic point cloud data, and these coordinates represent the position information of the detection target in space. Using the change in the position of the detection target between adjacent frames and combining the time interval, the speed information of the detection target is calculated. The speed information includes linear velocity (i.e., the moving speed along a certain direction) and / or angular velocity (i.e., the rotational speed around a certain point), specifically depending on the motion characteristics of the detection target. The signal strength information can be obtained through a signal strength sensor. Furthermore, the motion state of the detection target can be determined according to the position information, speed information, and signal strength information. The motion state includes a stationary state, a uniform motion state, an accelerating / decelerating motion state, a complex motion state, etc. For example, if the position information corresponding to the detection target is almost unchanged in multiple consecutive frames and the speed information is close to zero, then the motion state of the detection target is determined to be a stationary state. If the position information corresponding to the detection target changes linearly with time and the speed information remains constant, then the motion state of the detection target is determined to be a uniform motion state.

[0055] In summary, according to the object detection method based on vehicle-mounted radar according to an embodiment of the present invention, a point cloud data set of the environment where the vehicle is located is obtained through the vehicle-mounted radar, and the point cloud data set is divided into a static point cloud data subset and a dynamic point cloud data subset, and a static boundary model is determined according to the static point cloud data subset, and then the dynamic point cloud data subset is screened according to the static boundary model to determine the detection target, so as to determine the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target. Thus, the object detection ability, safety performance and environmental adaptability of the vehicle can be significantly improved, thereby improving the driving experience of users.

[0056] Based on the object detection method based on vehicle-mounted radar proposed in the foregoing embodiment of the present invention, an embodiment of the present invention also proposes a computer-readable storage medium, on which an object detection program based on vehicle-mounted radar is stored. When the object detection program based on vehicle-mounted radar is executed by a processor, the object detection method based on vehicle-mounted radar according to the foregoing embodiment of the present invention is implemented.

[0057] According to the computer-readable storage medium according to an embodiment of the present invention, by executing the object detection program based on vehicle-mounted radar by a processor, the object detection ability, safety performance and environmental adaptability of the vehicle can be significantly improved, thereby improving the driving experience of users.

[0058] Figure 2 It is a block diagram of an object detection device based on vehicle-mounted radar according to an embodiment of the present invention.

[0059] Specifically, as Figure 3 shown, the object detection device 100 based on vehicle-mounted radar includes an acquisition module 10, a processing module 20 and a determination module 30.

[0060] Among them, the acquisition module 10 is used to obtain a point cloud data set of the environment where the vehicle is located through the vehicle-mounted radar; the processing module 20 is used to divide the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; the determination module 30 is used to determine a static boundary model according to the static point cloud data subset; the determination module 30 is further used to screen the dynamic point cloud data subset according to the static boundary model to determine the detection target; the determination module 30 is further used to determine the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target.

[0061] In some embodiments of the present invention, the vehicle-mounted radar is a 4D millimeter-wave radar.

[0062] In some embodiments of the present invention, each data point in the point cloud data set includes at least one of time information, speed information, position information and signal strength information.

[0063] In some embodiments of the present invention, the point cloud data set is divided into a static point cloud data subset and a dynamic point cloud data subset based on the speed information.

[0064] In some embodiments of the present invention, the proportion of inlier data of the static boundary model in the static point cloud data subset is determined based on the position information; the amount of fitting data and the fitting accuracy of the static boundary model are obtained; the number of fitting iterations of the static boundary model is determined according to the proportion of inlier data, the amount of fitting data, and the fitting accuracy.

[0065] In some embodiments of the present invention, the determining module 30 is further configured to perform clustering processing on the dynamic point cloud data subset to obtain multi-cluster clustering data; determine the position information of multiple targets to be confirmed based on the center points of the multi-cluster clustering data; and verify the position information according to the static boundary model to filter out non-detection targets among the targets to be confirmed.

[0066] In some embodiments of the present invention, the determining module 30 is further configured to obtain the shape information of multiple targets to be confirmed; and filter out non-detection targets among the multiple targets to be confirmed according to the shape information and the position information.

[0067] In some embodiments of the present invention, the determining module 30 is further configured to obtain the position information, speed information, and signal strength information of the dynamic point cloud data corresponding to the detection target; and determine the motion state of the detection target according to the position information, speed information, and signal strength information.

[0068] It should be noted that for other specific embodiments of the target detection device based on vehicle-mounted radar proposed in the embodiments of the present invention, reference may be made to the specific embodiments of the target detection method based on vehicle-mounted radar in the foregoing embodiments of the present invention. To reduce redundancy, they will not be elaborated herein.

[0069] In summary, for the target detection device based on vehicle-mounted radar according to the embodiments of the present invention, the acquisition module obtains the point cloud data set of the vehicle's environment through the vehicle-mounted radar, and the processing module divides the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset, and the determining module determines the static boundary model according to the static point cloud data subset. Furthermore, the determining module filters the dynamic point cloud data subset according to the static boundary model to determine the detection target. Thus, the determining module determines the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target, which can significantly improve the vehicle's target detection ability, safety performance, and environmental adaptability, thereby enhancing the user's driving experience.

[0070] Figure 3 It is a schematic block diagram of a vehicle according to an embodiment of the present invention.

[0071] As Figure 3 shown, the vehicle 1000 includes the target detection device 100 based on vehicle-mounted radar according to the foregoing embodiments of the present invention.

[0072] A vehicle according to an embodiment of the present invention, by adopting the on-vehicle radar-based target detection device of the above embodiment of the present invention, can significantly improve the vehicle's target detection ability, safety performance and environmental adaptability, thereby enhancing the user's driving and riding experience.

[0073] In addition, the other constitutions and functions of the vehicle according to the embodiment of the present invention are known to those skilled in the art. To reduce redundancy, they are not described herein.

[0074] It should be noted that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus or device), or in combination with these instruction execution systems, apparatus or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus or device. More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other suitable processing as necessary, and then stored in a computer memory.

[0075] 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 embodiment, multiple 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, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0076] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0077] In the description of the present invention, it should be understood that the orientation or positional relationships indicated by terms such as "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. are based on the orientation or positional relationships shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation of the present invention.

[0078] In addition, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically and clearly defined.

[0079] In the present invention, unless otherwise clearly specified and limited, terms such as "mounted", "connected", "connected to", "fixed" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or integrated; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two elements or the interaction relationship between two elements, unless otherwise clearly limited. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0080] In the present invention, unless otherwise expressly specified or limited, a first feature being "on" or "under" a second feature may mean that the first and second features are in direct contact, or that the first and second features are in indirect contact via an intermediate medium. Further, a first feature being "above", "over" and "on top of" a second feature may mean that the first feature is directly above or obliquely above the second feature, or merely means that the first feature has a higher level of height than the second feature. A first feature being "under", "beneath" and "underneath" a second feature may mean that the first feature is directly below or obliquely below the second feature, or merely means that the first feature has a lower level of height than the second feature.

[0081] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and those of ordinary skill in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A target detection method based on vehicle-mounted radar, characterized in that: The method comprises: Acquire a point cloud data set of the environment in which the vehicle is located by using the vehicle-mounted radar; Dividing the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; Determining a static boundary model based on the static point cloud data subset; Screening the dynamic point cloud data subset according to the static boundary model to determine a detection target; The motion state of the detection target is determined according to the dynamic point cloud data corresponding to the detection target.

2. The target detection method based on vehicle-mounted radar according to claim 1, characterized in that: The vehicle-mounted radar is a 4D millimeter-wave radar.

3. The target detection method based on vehicle-mounted radar according to claim 1, characterized in that: Each data point in the point cloud data set includes at least one of time information, speed information, location information and signal strength information.

4. The target detection method based on vehicle-mounted radar according to claim 3 is characterized in that: The point cloud data set is divided into the static point cloud data subset and the dynamic point cloud data subset based on the speed information.

5. The target detection method based on vehicle-mounted radar according to claim 3 is characterized in that: The method further comprises: Determine the proportion of inlier data of the static boundary model in the static point cloud data subset based on the position information; Obtaining the amount of fitting data and fitting accuracy of the static boundary model; The number of fitting iterations of the static boundary model is determined according to the proportion of the interior point data, the amount of fitting data and the fitting accuracy.

6. The target detection method based on vehicle-mounted radar according to claim 1, characterized in that: Screening the dynamic point cloud data subset according to the static boundary model includes: Performing clustering processing on the dynamic point cloud data subset to obtain multi-cluster cluster data; Determine the location information of multiple targets to be confirmed by using the center points of the multiple clustered data; The position information is verified according to the static boundary model to filter out non-detected targets in the targets to be confirmed.

7. The target detection method based on vehicle-mounted radar according to claim 6, characterized in that: The method further comprises: Acquiring shape information of the multiple objects to be confirmed; Non-detection targets among the multiple targets to be confirmed are filtered out according to the shape information and the position information.

8. The target detection method based on vehicle-mounted radar according to claim 3 is characterized in that: Determining the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target includes: Obtaining position information, speed information, and signal strength information of dynamic point cloud data corresponding to the detection target; The motion state of the detection target is determined according to the position information, the speed information and the signal strength information.

9. A computer-readable storage medium, characterized in that: A target detection program based on a vehicle-mounted radar is stored thereon, and when the target detection program based on a vehicle-mounted radar is executed by a processor, a target detection method based on a vehicle-mounted radar according to any one of claims 1-8 is implemented.

10. A target detection device based on vehicle-mounted radar, characterized in that: The device comprises: An acquisition module, used to acquire a point cloud data set of the environment in which the vehicle is located through the vehicle-mounted radar; A processing module, used for dividing the point cloud data set into a static point cloud data subset and a dynamic point cloud data subset; A determination module, configured to determine a static boundary model according to the static point cloud data subset; The determination module is further used to screen the dynamic point cloud data subset according to the static boundary model to determine the detection target; The determination module is further used to determine the motion state of the detection target according to the dynamic point cloud data corresponding to the detection target.

11. A vehicle, characterized in that: Including the target detection device based on vehicle-mounted radar as described in claim 10.

Images