Target identification method and system for vehicle, vehicle and equipment

Abstract

The invention discloses a target identification method and system for a vehicle, the vehicle and equipment. The target recognition method for the vehicle comprises the steps that point cloud data of a radar and perception data of a camera are obtained, and the radar and the camera are installed on the vehicle; determining an identification area according to the sensing data, and filtering the point cloud data according to the identification area to obtain filtered point cloud data; performing target clustering on the filtered point cloud data to obtain a clustering target; and obtaining a drivable area according to the sensing data, and filtering the clustering target according to the drivable area and the parameters of the clustering target to obtain a final target recognition result. According to the embodiment of the invention, the accuracy and reliability of target recognition can be effectively improved, and the recognition efficiency is high.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and specifically to a vehicle target recognition method, system, vehicle and equipment. Background Art

[0002] Autonomous driving is the trend of automobile development, and the realization of autonomous driving includes many key technologies such as environmental perception, positioning, path planning and wire control. Among them, environmental perception is the basis of other technologies, and these technologies need to use the data obtained from environmental perception to be realized.

[0003] Environmental perception technology mainly includes target recognition and state perception, which is the result obtained by processing the data received by the vehicle-mounted sensors. Considering the complexity of the environment and the limitations of a single sensor, multiple types of sensors are usually used for detection to obtain better perception effects. At present, most of them use a combination of laser radar, camera and millimeter wave radar. However, the cost of laser radar is high and it is easily affected by weather. Traditional millimeter wave radar lacks the ability to measure height. 4D millimeter wave radar is a new sensor that has emerged in recent years. It increases the amount of point cloud data and adds height information, so that the state attributes of the target can be measured more accurately. It is also low-cost and highly adaptable. However, in the target clustering method of 4D millimeter wave radar, due to the working principle of millimeter wave radar, targets such as manhole covers on the road that do not affect the driving of vehicles are often detected, resulting in false detection, resulting in relatively low accuracy and reliability of target recognition. Summary of the invention

[0004] In view of the above-mentioned defects or deficiencies in the prior art, it is desirable to provide a multi-vehicle target recognition method, system, vehicle and device, which have the advantages of high target recognition accuracy and reliability, and have high recognition efficiency.

[0005] In a first aspect, an embodiment of the present application provides a method for identifying a target of a vehicle, comprising:

[0006] Obtaining point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle;

[0007] Determine a recognition area according to the perception data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data;

[0008] Performing target clustering on the filtered point cloud data to obtain cluster targets;

[0009] A drivable area is obtained according to the perception data, and the cluster targets are filtered according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

[0010] Further, determining a recognition area according to the perception data, and filtering the point cloud data according to the recognition area to obtain filtered point cloud data, includes:

[0011] Obtaining road information on a horizontal plane in a vehicle body coordinate system according to the perception data;

[0012] According to the road information, obtaining an identification area on the horizontal plane;

[0013] Converting the point cloud data into the vehicle body coordinate system and mapping the conversion result onto the horizontal plane;

[0014] The point cloud data corresponding to the mapping result outside the recognition area on the horizontal plane is eliminated to obtain the filtered point cloud data.

[0015] Further, obtaining the identification area on the horizontal plane according to the road information includes:

[0016] If a curb is detected, obtaining the identification area according to the curb;

[0017] If the roadside is not detected but the lane line is detected, the recognition area is obtained according to the lane line;

[0018] If the roadside and lane line are not detected, the recognition area is obtained according to a preset width. Further, after converting the point cloud data to the vehicle body coordinate system, the method further includes:

[0019] The point cloud data corresponding to the conversion results below the horizontal plane are eliminated.

[0020] Furthermore, the target clustering is performed on the filtered point cloud data to obtain cluster targets, including:

[0021] Determine a search radius correction coefficient of the point cloud according to the hyperbolic tangent function and the horizontal distance of the point cloud in the filtered point cloud data in the vehicle body coordinate system;

[0022] Obtaining an adaptive search radius of the point cloud according to an initial search radius and the search radius correction coefficient;

[0023] Obtain the Mahalanobis distance between point clouds in the filtered point cloud data;

[0024] A clustering algorithm based on the adaptive search radius is applied to perform target clustering on the point clouds according to the Mahalanobis distance between the point clouds in the filtered point cloud data to obtain clustering targets.

[0025] Further, obtaining a drivable area according to the perception data, and filtering the cluster targets according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result includes:

[0026] According to the sensing data, a drivable area is identified, where the drivable area is an area formed by a plurality of points in a horizontal plane under a vehicle body coordinate system;

[0027] Get the center point height and absolute speed of the clustered target;

[0028] For the same clustering target, when the following conditions are met at the same time, the clustering target is eliminated, and the conditions are:

[0029] Cluster targets whose projections fall into the drivable area; cluster targets whose center point height is less than a predetermined height; and cluster targets whose absolute speed and the predetermined speed have a difference less than a preset difference.

[0030] When the elimination is completed, the final target recognition result is obtained.

[0031] Furthermore, after obtaining the final target recognition result, the method further includes:

[0032] The vehicle is controlled to perform intelligent driving according to the final target recognition result.

[0033] In a second aspect, an embodiment of the present application provides a vehicle target recognition system, including:

[0034] An acquisition module, used to obtain point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle;

[0035] A first filtering module, used to determine a recognition area according to the perception data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data;

[0036] A clustering module, used for performing target clustering on the filtered point cloud data to obtain cluster targets;

[0037] The second filtering module is used to obtain a drivable area according to the perception data, and filter the cluster targets according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

[0038] In a third aspect, an embodiment of the present application provides a vehicle, comprising: a target recognition system for the vehicle according to the embodiment of the second aspect.

[0039] In a fourth aspect, an embodiment of the present application provides a computing device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the vehicle target recognition method described in the embodiment of the first aspect of the present application is implemented.

[0040] The vehicle target recognition method, system, vehicle and equipment proposed in the embodiments of the present application first filter the radar point cloud data according to the camera's perception data to filter out some interference points and point clouds that do not affect the vehicle's driving to obtain filtered point cloud data, then cluster the filtered point cloud data to effectively improve the efficiency and accuracy of clustering, and finally, combine the drivable area detected by the camera to perform a secondary filtering on the clustered targets of the point cloud data. In this way, the radar point cloud data is prevented from detecting targets on the road that do not affect the vehicle's driving, such as manhole covers, which may cause false detection of targets, thereby effectively ensuring the accuracy and reliability of target recognition.

[0041] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0043] Figure 1 A schematic diagram of a process flow of a vehicle target recognition method according to an embodiment of the present application;

[0044] Figure 2 This is an application flow chart of a vehicle target recognition method according to an embodiment of the present application;

[0045] Figure 3 A structural block diagram of a vehicle target recognition system according to an embodiment of the present application;

[0046] Figure 4 A schematic diagram of the structure of a computing device suitable for implementing an embodiment of the present application is shown. DETAILED DESCRIPTION

[0047] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant application, rather than to limit the application. It is also necessary to explain that, for ease of description, only the parts related to the application are shown in the accompanying drawings.

[0048] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0049] The following is combined with Figure 1 A target recognition method for a vehicle according to an embodiment of the present application is described to achieve high target recognition accuracy and recognition efficiency.

[0050] Figure 1 FIG. 1 is a flow chart of a method for identifying a vehicle target according to an embodiment of the present application. Figure 1 As shown, a vehicle target recognition method according to an embodiment of the present application includes the following steps:

[0051] S101: Obtaining point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle.

[0052] The radar is a vehicle-mounted radar, which is used to scan point cloud data such as the environment in front of the vehicle and its surroundings. The camera is a vehicle-mounted camera, which is used to collect image information such as the environment in front of the vehicle and its surroundings. The image information collected by the camera is used as perception data.

[0053] In one embodiment of the present application, the radar is but not limited to a 4D millimeter wave radar. A 4D millimeter wave radar can obtain more three-dimensional and more point cloud data.

[0054] It should be noted that before obtaining point cloud data through radar and perception data through camera, the 4D millimeter-wave radar and camera should be calibrated and synchronized in time and space. Specifically, the 4D millimeter-wave radar and camera installed on the vehicle should be jointly calibrated to ensure their synchronization in time and space.

[0055] After completing the calibration of the 4D millimeter-wave radar and camera and the synchronization processing of time and space, point cloud data is obtained through the 4D millimeter-wave radar and perception data is obtained through the camera.

[0056] S102: Determine a recognition area according to the perception data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data.

[0057] The identification area can be understood as a region of interest (ROI), that is, an area that needs to be paid attention to, such as an area that may generally have an impact on the driving of the vehicle.

[0058] In one embodiment of the present application, a recognition area is determined based on perception data, and point cloud data is filtered based on the recognition area to obtain filtered point cloud data, which may specifically include: obtaining road information on a horizontal plane under a vehicle body coordinate system based on the perception data; obtaining a recognition area on the horizontal plane based on the road information; converting the point cloud data to a vehicle body coordinate system, and mapping the conversion result to the horizontal plane; and eliminating point cloud data corresponding to the mapping result outside the recognition area on the horizontal plane to obtain filtered point cloud data.

[0059] In this example, based on the road information, an identification area on the horizontal plane is obtained, for example, including: if a curb is detected, the identification area is obtained based on the curb; if a lane line is detected but a curb is not detected, the identification area is obtained based on the lane line; if neither the curb nor the lane line is detected, the identification area is obtained based on a preset width.

[0060] That is to say, based on the camera's detection results of the roadside and lane lines, the recognition area, namely the region of interest (ROI), is determined, and the point cloud data outside the region of interest (ROI) is removed, thereby retaining the point cloud data that may affect the autonomous driving function. In this way, the amount of point cloud data is reduced, thereby reducing the subsequent calculation amount related to the point cloud data, which can effectively improve processing efficiency.

[0061] Specifically, in the point cloud data of the 4D millimeter wave radar, there are some points that are not important for the current driving state of the vehicle, that is, points outside the region of interest ROI, for example, point cloud data far away from the sides of the vehicle body. In order to improve the accuracy and efficiency of subsequent clustering of point cloud data, the point cloud data of the 4D millimeter wave radar is filtered for the first time. After filtering, the point cloud data that may affect the driving of the vehicle is retained. The filtering method is as follows:

[0062] Convert the point cloud data of the 4D millimeter wave radar to the vehicle body coordinate system. Consider the horizontal direction (i.e., in the xOy plane). Get the results of the curb and lane line detected by the camera. The result is the curve equation in the horizontal plane (i.e., xOy plane) in the vehicle body coordinate system, so as to determine the ROI (region of interest) in the horizontal direction:

[0063] 1. If the curb is currently detected, the ROI is the area within the curb;

[0064] 2. If the curb is not detected, but two or more lane lines are detected, the two outermost lane lines are selected, and the ROI is the area within these two lane lines;

[0065] 3. If the roadside is not detected and a sufficient number of lane lines are not detected, the ROI can be determined according to a preset width, wherein the preset width can be determined based on experience, for example, the width is set to 6 meters. It should be noted that the setting value of 6 meters is only exemplary, and in addition, the preset width can be set to be large enough, for example, larger than the width of the entire horizontal detection area of ​​the radar, in which case the entire horizontal detection area of ​​the radar can be used as the ROI.

[0066] In addition, there are usually some interference points in the point cloud data, such as points below the horizontal plane. In this method, after converting the point cloud data to the vehicle body coordinate system, it also includes: eliminating the point cloud data corresponding to the conversion result below the horizontal plane. Specifically, consider the vertical direction (ie, the z-axis direction). Combined with the installation height of the 4D millimeter-wave radar, the points below the ground are regarded as interference points, and the point cloud data of this part of the radar is eliminated from the point cloud data, thereby further improving the accuracy and efficiency of the subsequent clustering of the point cloud data.

[0067] S103: Performing target clustering on the filtered point cloud data to obtain cluster targets.

[0068] By filtering the point cloud data in step S102, some point cloud data that will not affect vehicle driving and some interference points are filtered out, thereby greatly reducing the amount of point cloud data. In this way, when the point cloud data is clustered by a clustering algorithm, the clustering efficiency can be effectively improved and the amount of clustering calculation can be reduced.

[0069] The clustering algorithm may adopt some existing clustering algorithms.

[0070] In one embodiment of the present application, in order to improve the clustering effect, the characteristics of the point cloud data of the 4D millimeter wave radar are fully considered in the clustering of the point cloud data, that is, the farther the distance of the point cloud data collected by the 4D millimeter wave radar, the sparser the point cloud. In order to obtain a good clustering effect at both close and long distances, an adaptive search radius is used in the embodiment of the present application to achieve clustering. For example, target clustering is performed on the filtered point cloud data to obtain a clustering target, including: determining a search radius correction coefficient of the point cloud according to a hyperbolic tangent function (tanh function) and the horizontal distance of the point cloud in the filtered point cloud data in the vehicle body coordinate system; obtaining an adaptive search radius of the point cloud according to the initial search radius and the search radius correction coefficient; obtaining the Mahalanobis distance between the point clouds in the filtered point cloud data (such as calculating the Mahalanobis distance between the point clouds according to the coordinates, speed, etc. of the point clouds); applying a clustering algorithm based on the adaptive search radius, according to the Mahalanobis distance between the point clouds in the filtered point cloud data, target clustering is performed on the point cloud to obtain a clustering target.

[0071] That is, the Mahalanobis distance between two points is calculated, and the improved DBSCAN clustering algorithm is used to cluster the point cloud data to obtain the clustering target. Specifically, the point cloud of the 4D millimeter wave radar obtained after filtering in step 3 is clustered. The clustering method can be density clustering based on the Mahalanobis distance, such as the improved DBSCAN algorithm.

[0072] For example: use p(x,y,z,v x ,v y ) is used to describe the points in the point cloud data of the 4D millimeter-wave radar, where x, y, and z are the coordinates of point p, and v x 、v y are the velocities in the x- and y-axis directions.

[0073] Calculate two points p i 、p j The Mahalanobis distance between them is:

[0074]

[0075] Since the point cloud data collected by the 4D millimeter wave radar is sparser as the distance increases, an adaptive search radius is used to obtain good clustering effects at both short and long distances. The tanh function has a good match with it due to its characteristics, and its mathematical formula is as follows:

[0076]

[0077] Therefore, the correction coefficient of the search radius is determined based on the tanh function:

[0078]

[0079] in, is the horizontal distance of point p. The parameters k, a, and b can be determined through actual tests.

[0080] Set the initial search radius to eps 0 , calculate the adaptive search radius eps = eps for each point 0 *f(r), then, based on the calculated adaptive search radius of each point, the point cloud data of the 4D millimeter-wave radar is clustered to obtain the cluster target result, i.e., the cluster target.

[0081] S104: Obtain a drivable area according to the perception data, and filter the cluster targets according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

[0082] Specifically, after clustering is completed, the clustering targets can be filtered again according to the perception data, the parameters of the clustering targets, etc. Among them, the parameters of the clustering targets include but are not limited to the center point height and absolute speed of the clustering targets, etc. The implementation method of obtaining the final target recognition result may specifically include: identifying the drivable area according to the perception data, the drivable area is the area composed of multiple points in the horizontal plane under the vehicle body coordinate system; obtaining the center point height and absolute speed of the clustering target; for the same clustering target, when the following conditions are met at the same time, the clustering target is eliminated, the conditions are: the projection of the clustering target falls into the drivable area; the center point height of the clustering target is less than the predetermined height of the clustering target; the difference between the absolute speed of the clustering target and the predetermined speed is less than the preset difference; when the elimination is completed, the final target recognition result is obtained.

[0083] Specifically, due to the working principle of millimeter wave radar, targets on the road that do not affect vehicle driving, such as manhole covers, are often detected, resulting in false detection. Therefore, in order to further improve the accuracy of target recognition, it is necessary to filter the clustered targets obtained in step S103 again to achieve more accurate and reliable target detection accuracy. For example:

[0084] Get the drivable area (freespace) result detected by the camera, which is the area surrounded by several points in the horizontal plane (i.e., xOy plane) under the vehicle body coordinate system, denoted as S. For example: the road in front of the vehicle.

[0085] Select cluster targets that meet the following filtering conditions: 1. The horizontal projection of the target falls within the area S; 2. The center point height of the target is less than z 0 ; 3. The absolute speed of the target is less than v 0 Among them, z 0 、v 0 The thresholds are small and close to 0, which can be determined through actual testing. The clustering targets that meet the above 1, 2 and 3 are eliminated.

[0086] According to the vehicle target recognition method of the embodiment of the present application, the radar point cloud data is first filtered according to the perception data of the camera to filter out some interference points and point clouds that do not affect the vehicle's driving to obtain filtered point cloud data. Then, the filtered point cloud data is clustered to effectively improve the efficiency and accuracy of clustering. Finally, the clustered targets of the point cloud data are filtered again in combination with the drivable area detected by the camera. In this way, the radar point cloud data is prevented from detecting targets on the road that do not affect the vehicle's driving, such as manhole covers, which may cause false detection of targets. This effectively ensures the accuracy and reliability of target recognition.

[0087] like Figure 2As shown, it is an application flow chart of the target recognition method of a vehicle in an embodiment of the present application, such as Figure 2 As shown, the 4D millimeter-wave radar and camera are calibrated. After the calibration is completed, the point cloud data of the 4D millimeter-wave radar and the perception data of the camera are received; then, the point cloud data is filtered once, and the improved DBSCAN algorithm is used for point cloud clustering. Finally, the clustering result target is filtered twice to obtain and publish the final clustering result, that is, the target recognition result. In this way, the point cloud of the 4D millimeter-wave radar is filtered, and the data that may affect the autonomous driving function is retained, which not only improves the effectiveness of the clustering results, but also reduces the amount of calculation. In addition, in the clustering method, compared with the clustering algorithm in the prior art, an adaptive search radius is used to improve the accuracy of point cloud clustering at a long distance. At the same time, targets such as manhole covers that may exist on the road and do not affect the driving of the vehicle are further filtered, thereby reducing the false detection rate and improving the accuracy and reliability of target recognition.

[0088] In one embodiment of the present application, the vehicle target recognition method, after obtaining the final target recognition result, further includes: controlling the vehicle to perform intelligent driving according to the final target recognition result. That is, the result of step S104 is organized into the final 4D millimeter wave radar target detection result, and then the result can be published to other modules such as data fusion or path planning to support the realization of relevant functions of the vehicle.

[0089] Figure 3 FIG. 1 is a block diagram of a vehicle target recognition system according to an embodiment of the present application. Figure 3 As shown, the vehicle target recognition system according to the embodiment of the present application includes: an acquisition module 310, a first filtering module 320, a clustering module 330 and a second filtering module 340, wherein:

[0090] An acquisition module 310 is used to obtain point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle;

[0091] A first filtering module 320, configured to determine a recognition area according to the sensing data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data;

[0092] A clustering module 330 is used to perform target clustering on the filtered point cloud data to obtain cluster targets;

[0093] The second filtering module 340 is used to obtain a drivable area according to the perception data, and filter the cluster targets according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

[0094] According to the target recognition system of the vehicle in the embodiment of the present application, the point cloud data of the radar is first filtered according to the perception data of the camera to filter out some interference points and point clouds that do not affect the driving of the vehicle to obtain filtered point cloud data. Then, the filtered point cloud data is clustered to effectively improve the efficiency and accuracy of clustering. Finally, the clustered targets of the point cloud data are filtered for a second time in combination with the drivable area detected by the camera. In this way, the radar point cloud data is prevented from detecting targets on the road that do not affect the driving of the vehicle, such as manhole covers, which may cause false detection of targets, thereby effectively ensuring the accuracy and reliability of target recognition.

[0095] It should be noted that the specific implementation method of the target recognition system of the vehicle in the embodiment of the present application is similar to the specific implementation method of the target recognition method of the vehicle in the embodiment of the present application. Please refer to the description of the method part for details and will not be repeated here.

[0096] Furthermore, an embodiment of the present application provides a vehicle, including: a target recognition system for a vehicle according to the above embodiment. The vehicle can filter the point cloud data of the radar once according to the perception data of the camera, filter out some interference points and point clouds that do not affect the driving of the vehicle, and obtain filtered point cloud data. Then, the filtered point cloud data is subjected to target clustering, so that the efficiency and accuracy of clustering are effectively improved. Finally, the clustered targets of the point cloud data are filtered again in combination with the drivable area detected by the camera. In this way, the radar point cloud data is prevented from detecting targets on the road that do not affect the driving of the vehicle, such as manhole covers, and causing misdetection of the target, thereby effectively ensuring the accuracy and reliability of target recognition.

[0097] Reference below Figure 4 , Figure 4 A schematic diagram of the structure of a computing device suitable for implementing an embodiment of the present application is shown.

[0098] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1008 into the random access memory (RAM) 1003. In the RAM 1003, various programs and data required for the operation instructions of the system are also stored. The CPU 1001, the ROM 1002, and the RAM 1003 are connected to each other through a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0099] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, etc.; an output section 1007 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 1008 including a hard disk, etc.; and a communication section 1009 including a network interface card such as a LAN card, a modem, etc. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. A removable medium 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 1010 as needed, so that a computer program read therefrom is installed into the storage section 1008 as needed.

[0100] In particular, according to an embodiment of the present application, the above reference flow chart Figure 1 The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 1009, and / or installed from the removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, the above-mentioned functions defined in the system of the present application are executed.

[0101] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0102] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.

[0103] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.

[0104] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the computing device described in the above embodiment, or may exist independently without being assembled into the computing device. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the vehicle target recognition method described in the present application.

[0105] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. A vehicle target recognition method, characterized in that: include: Obtaining point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle; Determine a recognition area according to the perception data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data; Performing target clustering on the filtered point cloud data to obtain cluster targets; A drivable area is obtained according to the perception data, and the cluster targets are filtered according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

2. The vehicle target recognition method according to claim 1, characterized in that: The determining of the recognition area according to the perception data, and filtering the point cloud data according to the recognition area to obtain filtered point cloud data, includes: Obtaining road information on a horizontal plane in a vehicle body coordinate system according to the perception data; According to the road information, obtaining an identification area on the horizontal plane; Converting the point cloud data into the vehicle body coordinate system and mapping the conversion result onto the horizontal plane; The point cloud data corresponding to the mapping result outside the recognition area on the horizontal plane is eliminated to obtain the filtered point cloud data.

3. The vehicle target recognition method according to claim 2, characterized in that: The step of obtaining the identification area on the horizontal plane according to the road information includes: If a curb is detected, obtaining the identification area according to the curb; If the roadside is not detected but the lane line is detected, the recognition area is obtained according to the lane line; If the roadside and the lane line are not detected, the recognition area is obtained according to a preset width.

4. The vehicle target recognition method according to claim 2 or 3, characterized in that: After converting the point cloud data into the vehicle body coordinate system, the method further includes: The point cloud data corresponding to the conversion results below the horizontal plane are eliminated.

5. The vehicle target recognition method according to claim 1, characterized in that: The performing target clustering on the filtered point cloud data to obtain a cluster target includes: Determine a search radius correction coefficient of the point cloud according to the hyperbolic tangent function and the horizontal distance of the point cloud in the filtered point cloud data in the vehicle body coordinate system; Obtaining an adaptive search radius of the point cloud according to an initial search radius and the search radius correction coefficient; Obtain the Mahalanobis distance between point clouds in the filtered point cloud data; A clustering algorithm based on the adaptive search radius is applied to perform target clustering on the point clouds according to the Mahalanobis distance between the point clouds in the filtered point cloud data to obtain clustering targets.

6. The vehicle target recognition method according to claim 1, characterized in that: The obtaining of the drivable area according to the sensing data, and filtering the cluster targets according to the drivable area and the parameters of the cluster targets to obtain the final target recognition result includes: According to the sensing data, a drivable area is identified, where the drivable area is an area formed by a plurality of points in a horizontal plane under a vehicle body coordinate system; Get the center point height and absolute speed of the clustered target; For the same clustering target, when the following conditions are met at the same time, the clustering target is eliminated, and the conditions are: Cluster targets whose projections fall into the drivable area; cluster targets whose center point height is less than a predetermined height; and cluster targets whose absolute speed and the predetermined speed have a difference less than a preset difference. When the elimination is completed, the final target recognition result is obtained.

7. The vehicle target recognition method according to claim 1, characterized in that: After obtaining the final target recognition result, the method further includes: The vehicle is controlled to perform intelligent driving according to the final target recognition result.

8. A vehicle target recognition system, characterized in that: include: An acquisition module, used to obtain point cloud data of a radar and perception data of a camera, wherein the radar and the camera are installed on a vehicle; A first filtering module, used to determine a recognition area according to the perception data, and filter the point cloud data according to the recognition area to obtain filtered point cloud data; A clustering module, used for performing target clustering on the filtered point cloud data to obtain cluster targets; The second filtering module is used to obtain a drivable area according to the perception data, and filter the cluster targets according to the drivable area and the parameters of the cluster targets to obtain a final target recognition result.

9. A vehicle, characterized in that: include: The object recognition system for a vehicle according to claim 8.

10. A computing device, characterized in that The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for target recognition of a vehicle according to any one of claims 1 to 7 is implemented.

Images