Point cloud map freshness updating early warning method and device, electronic equipment and medium

Abstract

The invention relates to a point cloud map freshness updating early warning method and device, electronic equipment and a medium, and is applied to the technical field of automatic driving, and the method comprises the steps: matching real-time point cloud data with a point cloud map, and determining a track segment corresponding to the point cloud data of which the matching confidence coefficient is smaller than a preset confidence coefficient as a target track segment; constructing a normal distribution transformation map according to the point cloud data of the target track segment in the point cloud map; matching the features of the real-time point cloud data in each point cloud frame of the target track segment with the features in the normal distribution transformation map, and determining a matching result; determining the pose deviation of the point cloud frame according to the fusion positioning pose and the point cloud matching positioning pose; and if N continuous point cloud frames of which the corresponding matching results are mismatched and the corresponding pose deviations are smaller than the preset pose deviation exist, determining that the map freshness of the target trajectory segment is abnormal, and performing early warning. The accuracy and efficiency of the freshness early warning of the point cloud map can be improved.

Description

Technical Field

[0001] The present application relates to the field of autonomous driving technology, and in particular to a point cloud map freshness update warning method, device, electronic equipment and medium. Background Art

[0002] As the technology of the autonomous driving industry matures, L4 autonomous driving has brought revolutionary impacts on people's lives. It can improve the operating efficiency and safety of autonomous vehicles, reduce labor costs, and improve service quality and coverage. For autonomous vehicles, accurate and reliable point cloud maps are crucial to the positioning module. In related technologies, on-site operators are required to observe the degree of change in the surrounding environment to determine whether the point cloud map needs to be updated. However, this method is unstable and inefficient. Summary of the invention

[0003] In order to solve the above technical problems, the present application provides a point cloud map freshness update warning method, device, electronic device and medium.

[0004] According to a first aspect of the present application, a point cloud map freshness update warning method is provided, comprising:

[0005] Acquire real-time point cloud data collected by the autonomous driving vehicle during the driving trajectory, match the real-time point cloud data with the point cloud map, and determine the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment;

[0006] Constructing a normal distribution transformation map according to the point cloud data of the target trajectory segment in the point cloud map;

[0007] For each point cloud frame of the target trajectory segment, matching features of the real-time point cloud data in the point cloud frame with features in the normal distribution transformation map to determine a matching result of the point cloud frame;

[0008] If there are no consecutive N point cloud frames whose corresponding matching results are all unmatched, it is determined that there is no abnormality in the map freshness of the target trajectory segment; N is a positive integer;

[0009] If there are N consecutive point cloud frames whose corresponding matching results are all unmatched, multi-sensor data of the target trajectory segment is acquired, and fusion positioning is performed according to the multi-sensor data to obtain a fusion positioning posture;

[0010] For each point cloud frame of the target trajectory segment, matching and positioning are performed according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching and positioning posture of the point cloud frame;

[0011] Determining a posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame;

[0012] If the posture deviations corresponding to the N consecutive point cloud frames are all smaller than the preset posture deviations, it is determined that the map freshness of the target trajectory segment is abnormal, and an early warning is issued.

[0013] Optionally, matching the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map to determine the matching result of the point cloud frame includes:

[0014] Dividing the real-time point cloud data in the point cloud frame into a plurality of different voxels, and calculating Gaussian distribution characteristics of the point cloud data in different voxels;

[0015] Matching the Gaussian distribution features of the point cloud data in each voxel with the Gaussian distribution features of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel;

[0016] If the number of voxels in the point cloud frame corresponding to the matching result is greater than or equal to M, then the matching result of the point cloud frame is determined to be a match; M is a positive integer;

[0017] If the number of voxels in the point cloud frame corresponding to the matching result is less than M, it is determined that the matching result of the point cloud frame is not matched.

[0018] Optionally, the Gaussian distribution features of the point cloud data within each voxel include: a feature vector of the point cloud data within each voxel and position coordinates of the centroid of the point cloud data within each voxel;

[0019] The step of matching the Gaussian distribution features of the point cloud data in each voxel with the Gaussian distribution features of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel includes:

[0020] For each voxel, calculating the angle between the feature vector of the point cloud data in the voxel and the feature vector of the corresponding point cloud data in the voxel in the normal distribution transformation map; and,

[0021] Calculating the distance between the centroid of the point cloud data within the voxel and the centroid of the corresponding point cloud data within the voxel in the normal distribution transformation map;

[0022] If the angle is less than or equal to a preset angle, and the distance is less than or equal to a preset distance, determining that the matching result corresponding to the voxel is a match;

[0023] If the angle is greater than a preset angle, or the distance is greater than a preset distance, it is determined that the matching result corresponding to the voxel is not matched.

[0024] Optionally, the pose deviation of each point cloud frame includes: a six-degree-of-freedom pose deviation;

[0025] Determine whether the pose deviation of each point cloud frame is less than the preset pose deviation, including:

[0026] If the six-degree-of-freedom posture deviation of each point cloud frame is respectively less than the corresponding deviation threshold, it is determined that the posture deviation of each point cloud frame is less than the preset posture deviation;

[0027] Otherwise, it is determined that the pose deviation of each point cloud frame is greater than or equal to the preset pose deviation.

[0028] According to a second aspect of the present application, a point cloud map freshness update warning device is provided, comprising:

[0029] A real-time point cloud data acquisition module is used to acquire real-time point cloud data collected by the autonomous driving vehicle during its driving trajectory;

[0030] a target trajectory segment determination module, used to match the real-time point cloud data with the point cloud map, and determine the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment;

[0031] A normal distribution transformation map construction module, used to construct a normal distribution transformation map according to the point cloud data of the target trajectory segment in the point cloud map;

[0032] A point cloud matching module, for each point cloud frame of the target trajectory segment, matches the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map, and determines the matching result of the point cloud frame;

[0033] A map freshness non-abnormal determination module is used to determine that the map freshness of the target trajectory segment does not have abnormality if there are no consecutive N point cloud frames whose corresponding matching results are all unmatched; N is a positive integer;

[0034] A fusion positioning posture determination module is used to obtain multi-sensor data of the target trajectory segment if there are N consecutive point cloud frames whose corresponding matching results are all unmatched, and perform fusion positioning according to the multi-sensor data to obtain a fusion positioning posture;

[0035] A point cloud matching positioning posture determination module is used to match and locate each point cloud frame of the target trajectory segment according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching positioning posture of the point cloud frame;

[0036] A posture deviation determination module, used to determine the posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame;

[0037] The point cloud map freshness warning module is used to determine that the map freshness of the target trajectory segment is abnormal and issue a warning if the posture deviations corresponding to the N consecutive point cloud frames are all less than the preset posture deviations.

[0038] Optionally, the point cloud matching module is specifically used to divide the real-time point cloud data in each point cloud frame of the target trajectory segment into multiple different voxels, and calculate the Gaussian distribution characteristics of the point cloud data in different voxels; match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel; if the corresponding matching result in the point cloud frame is that the number of matched voxels is greater than or equal to M, then the matching result of the point cloud frame is determined to be a match; if the corresponding matching result in the point cloud frame is that the number of matched voxels is less than M, then the matching result of the point cloud frame is determined to be a mismatch, and M is a positive integer.

[0039] Optionally, the Gaussian distribution features of the point cloud data within each voxel include: a feature vector of the point cloud data within each voxel and position coordinates of the centroid of the point cloud data within each voxel;

[0040] The point cloud matching module is specifically used to match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map through the following steps to obtain a matching result corresponding to each voxel:

[0041] For each voxel, calculating the angle between the feature vector of the point cloud data in the voxel and the feature vector of the corresponding point cloud data in the voxel in the normal distribution transformation map; and,

[0042] Calculating the distance between the centroid of the point cloud data within the voxel and the centroid of the corresponding point cloud data within the voxel in the normal distribution transformation map;

[0043] If the angle is less than or equal to a preset angle, and the distance is less than or equal to a preset distance, determining that the matching result corresponding to the voxel is a match;

[0044] If the angle is greater than a preset angle, or the distance is greater than a preset distance, it is determined that the matching result corresponding to the voxel is not matched.

[0045] Optionally, the pose deviation of each point cloud frame includes: a six-degree-of-freedom pose deviation;

[0046] The device also includes: a posture deviation judgment module;

[0047] The posture deviation judgment module is used to determine that the posture deviation of each point cloud frame is less than a preset posture deviation if the six-degree-of-freedom posture deviation of each point cloud frame is less than the corresponding deviation threshold; otherwise, determine that the posture deviation of each point cloud frame is greater than or equal to the preset posture deviation.

[0048] According to a third aspect of the present application, an electronic device is provided, comprising: a processor, wherein the processor is configured to execute a computer program stored in a memory, wherein the computer program implements the method described in the first aspect when executed by the processor.

[0049] According to a fourth aspect of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in the first aspect is implemented.

[0050] According to a fifth aspect of the present application, a computer program product is provided. When the computer program product is run on a computer, the computer is enabled to execute the method described in the first aspect.

[0051] Compared with the prior art, the technical solution provided by the embodiments of the present application has the following advantages:

[0052] By matching the real-time point cloud data collected by the autonomous driving vehicle in the driving trajectory with the point cloud data in the point cloud map, the target trajectory segment with a smaller matching confidence is determined. According to the point cloud data of the target trajectory segment in the point cloud map, a normal distribution transformation map is constructed. For each point cloud frame of the target trajectory segment, the features of the real-time point cloud data in the point cloud frame are matched with the features in the normal distribution transformation map to determine the matching result of the point cloud frame. In addition, if there are N consecutive point cloud frames whose corresponding matching results are all unmatched, the posture deviation between the two is calculated by calculating the point cloud matching positioning pose and the fusion positioning pose. If the posture deviations corresponding to the N consecutive point cloud frames are all less than the preset posture deviation, it is determined that the map freshness of the target trajectory segment is abnormal, and an early warning is issued. In the embodiment of the present application, the accuracy of the point cloud map freshness update early warning can be improved by cross-validation of the matching results and posture deviations of the point cloud frame. At the same time, there is no need for on-site operators to judge whether the point cloud map needs to be updated through experience, which can improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0054] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0055] Figure 1 A flow chart of a point cloud map freshness update warning method in an embodiment of the present application;

[0056] Figure 2 A schematic diagram of a structure of a point cloud map freshness update warning device in an embodiment of the present application;

[0057] Figure 3 A schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0058] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the scheme of the present application will be further described below. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present application, but the present application may also be implemented in other ways different from those described herein; obviously, the embodiments in the specification are only part of the embodiments of the present application, rather than all of the embodiments.

[0060] See also Figure 1 , Figure 1 This is a flow chart of a method for early warning of point cloud map freshness update in an embodiment of the present application, which may include the following steps:

[0061] Step S102, obtaining real-time point cloud data collected by the autonomous driving vehicle during the driving trajectory, matching the real-time point cloud data with the point cloud map, and determining the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment.

[0062] When driving, an autonomous vehicle can collect nearby real-time point cloud data through a laser radar. The pre-built point cloud map also stores point cloud data for each area. By matching the real-time point cloud data of multiple different track segments in the driving trajectory with the point cloud map, the matching confidences corresponding to the multiple different track segments can be obtained. The higher the matching confidence corresponding to a track segment, the more consistent the real-time point cloud data of the track segment is with the point cloud data in the point cloud map; the lower the matching confidence corresponding to a track segment, the more inconsistent the real-time point cloud data of the track segment is with the point cloud data in the point cloud map.

[0063] If the matching confidence of the trajectory segment is greater than or equal to the preset confidence, it can be considered that the surrounding environment of the trajectory segment has not changed; if the matching confidence is less than the preset confidence, it is considered that the surrounding environment of the trajectory segment may have changed. Therefore, the target trajectory segment with a matching confidence less than the preset confidence is selected for further judgment.

[0064] Step S104: construct a normal distribution transformation map based on the point cloud data of the target trajectory segment in the point cloud map.

[0065] By performing a normal distribution transformation on the point cloud data of the target trajectory segment in the point cloud map, a normal distribution transformation map can be obtained. That is, the space where the point cloud data of the target trajectory segment in the point cloud map is located is divided into multiple grids, namely voxels, and the Gaussian distribution characteristics of the point cloud data in each voxel are calculated. Voxels are pixels in 3D space, quantized point clouds of fixed size.

[0066] Step S106, for each point cloud frame of the target trajectory segment, matching the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map to determine the matching result of the point cloud frame.

[0067] In some embodiments, the real-time point cloud data in the point cloud frame can be divided into a plurality of different voxels, and the Gaussian distribution features of the point cloud data in different voxels are calculated. The Gaussian distribution features of the point cloud data in each voxel are matched with the Gaussian distribution features of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain the matching result corresponding to each voxel.

[0068] Optionally, the Gaussian distribution features of the point cloud data within each voxel may include: a feature vector of the point cloud data within each voxel and the position coordinates of the center of mass of the point cloud data within each voxel. The feature vector of the point cloud data within each voxel may include feature vectors in three directions of XYZ. The position coordinates of the center of mass are the average of the position coordinates of the point cloud data contained in the voxel. For each voxel, the angle between the feature vector of the point cloud data within the voxel and the feature vector of the corresponding point cloud data within the voxel in the normal distribution transformation map is calculated; and the distance between the center of mass of the point cloud data within the voxel and the center of mass of the corresponding point cloud data within the voxel in the normal distribution transformation map is calculated. If the angle is less than or equal to the preset angle, and the distance is less than or equal to the preset distance, it is considered that the matching effect is good, and the matching result corresponding to the voxel is determined to be a match; if the angle is greater than the preset angle, or the distance is greater than the preset distance, it is considered that the matching effect is poor, and the matching result corresponding to the voxel is determined to be a mismatch.

[0069] If the number of voxels corresponding to the matching result in the point cloud frame is greater than or equal to M, the matching result corresponding to the larger number of voxels in the point cloud frame is considered to be a match, and the matching result of the point cloud frame is determined to be a match; if the number of voxels corresponding to the matching result in the point cloud frame is less than M, the matching result corresponding to the smaller number of voxels in the point cloud frame is considered to be a match, and the matching result of the point cloud frame is determined to be a mismatch, and M is a positive integer.

[0070] Step S108, determining whether there are N consecutive point cloud frames whose corresponding matching results are all unmatched.

[0071] If there are N consecutive point cloud frames whose corresponding matching results are all unmatched, indicating that the surrounding environment of the target trajectory segment may have changed, step S110 is executed for further judgment; if there are no N consecutive point cloud frames whose corresponding matching results are all unmatched, indicating that the surrounding environment of the target trajectory segment has not changed, step S120 is executed. Wherein, N is a positive integer, and the trajectory segment corresponding to the N consecutive point cloud frames is greater than or equal to the preset distance.

[0072] Step S110, acquiring multi-sensor data of the target trajectory segment, and performing fusion positioning based on the multi-sensor data to obtain a fusion positioning posture.

[0073] Multi-sensor data may include image data and IMU (inertial measurement unit) data, etc. Based on the image data and IMU data, the autonomous driving vehicle can be fused and positioned to obtain a fused positioning posture.

[0074] Step S112, for each point cloud frame of the target trajectory segment, matching and positioning are performed according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching and positioning posture of the point cloud frame.

[0075] Step S114, determining the posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame.

[0076] Step S116, determining whether there are N consecutive point cloud frames whose corresponding matching results are all mismatched and whose corresponding posture deviations are all smaller than a preset posture deviation.

[0077] If the corresponding matching results are all unmatched, and the corresponding pose deviations of N consecutive point cloud frames are all less than the preset pose deviation, it means that the surrounding environment of the target trajectory segment has changed, and step S118 is executed; if the corresponding matching results are all unmatched, and there are point cloud frames in N consecutive point cloud frames whose corresponding pose deviations are greater than or equal to the preset pose deviation, it is not sufficient to indicate that the surrounding environment of the target trajectory segment has not changed.

[0078] Optionally, the pose deviation of each point cloud frame includes: a six-degree-of-freedom pose deviation. The method for determining whether the pose deviation of each point cloud frame is less than a preset pose deviation can be: if the six-degree-of-freedom pose deviation of each point cloud frame is less than the corresponding deviation threshold, then the pose deviation of each point cloud frame is determined to be less than the preset pose deviation; otherwise, the pose deviation of each point cloud frame is determined to be greater than or equal to the preset pose deviation. The pose deviation of each point cloud frame can be determined more accurately based on the six-degree-of-freedom pose deviation.

[0079] Step S118, determining whether the map freshness of the target trajectory segment is abnormal and issuing an early warning.

[0080] The point cloud map freshness update warning method of the embodiment of the present application is to match the real-time point cloud data collected by the autonomous driving vehicle in the driving trajectory with the point cloud data in the point cloud map to determine the target trajectory segment with a smaller matching confidence. According to the point cloud data of the target trajectory segment in the point cloud map, a normal distribution transformation map is constructed. For each point cloud frame of the target trajectory segment, the features of the real-time point cloud data in the point cloud frame are matched with the features in the normal distribution transformation map to determine the matching result of the point cloud frame. If there are N consecutive point cloud frames whose corresponding matching results are all unmatched, the posture deviation between the two is calculated by calculating the point cloud matching positioning pose and the fusion positioning pose. If the posture deviations corresponding to the N consecutive point cloud frames are all less than the preset posture deviation, it is determined that the map freshness of the target trajectory segment is abnormal, and an early warning is issued. In the embodiment of the present application, the accuracy of the point cloud map freshness update warning can be improved by cross-validation of the matching results and posture deviations of the point cloud frames. At the same time, there is no need for on-site operators to judge whether the point cloud map needs to be updated through experience, which can improve work efficiency.

[0081] It should be noted that although the steps of the method in the present application are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0082] Corresponding to the above method embodiment, the present application embodiment also provides a point cloud map freshness update warning device, see Figure 2 The point cloud map freshness update warning device 200 includes:

[0083] A real-time point cloud data acquisition module 202 is used to acquire real-time point cloud data collected by the autonomous driving vehicle during its driving trajectory;

[0084] The target trajectory segment determination module 204 is used to match the real-time point cloud data with the point cloud map, and determine the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment;

[0085] A normal distribution transformation map construction module 206 is used to construct a normal distribution transformation map according to the point cloud data of the target trajectory segment in the point cloud map;

[0086] The point cloud matching module 208 is used to match the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map for each point cloud frame of the target trajectory segment, and determine the matching result of the point cloud frame;

[0087] The map freshness non-abnormal determination module 210 is used to determine that the map freshness of the target trajectory segment does not have abnormality if there are no consecutive N point cloud frames whose corresponding matching results are all unmatched; N is a positive integer;

[0088] The fusion positioning posture determination module 212 is used to obtain multi-sensor data of the target trajectory segment if there are N consecutive point cloud frames whose corresponding matching results are all unmatched, and perform fusion positioning based on the multi-sensor data to obtain a fusion positioning posture;

[0089] The point cloud matching positioning posture determination module 214 is used to match and locate each point cloud frame of the target trajectory segment according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching positioning posture of the point cloud frame;

[0090] A posture deviation determination module 216, for determining the posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame;

[0091] The point cloud map freshness warning module 218 is used to determine that the map freshness of the target trajectory segment is abnormal and issue a warning if the posture deviations corresponding to N consecutive point cloud frames are all smaller than a preset posture deviation.

[0092] Optionally, the point cloud matching module 208 is specifically used to divide the real-time point cloud data in the point cloud frame into multiple different voxels for each point cloud frame of the target trajectory segment, and calculate the Gaussian distribution characteristics of the point cloud data in different voxels; match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel; if the corresponding matching result in the point cloud frame is that the number of matched voxels is greater than or equal to M, then the matching result of the point cloud frame is determined to be a match; if the corresponding matching result in the point cloud frame is that the number of matched voxels is less than M, then the matching result of the point cloud frame is determined to be a mismatch, and M is a positive integer.

[0093] Optionally, the Gaussian distribution features of the point cloud data within each voxel include: a feature vector of the point cloud data within each voxel and position coordinates of the centroid of the point cloud data within each voxel;

[0094] The point cloud matching module 208 is specifically used to match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map through the following steps to obtain the matching result corresponding to each voxel:

[0095] For each voxel, calculating the angle between the eigenvector of the point cloud data within the voxel and the eigenvector of the point cloud data within the corresponding voxel in the normal distribution transformation map; and,

[0096] Calculate the distance between the centroid of the point cloud data within the voxel and the centroid of the point cloud data within the corresponding voxel in the normal distribution transformation map;

[0097] If the angle is less than or equal to the preset angle, and the distance is less than or equal to the preset distance, then the matching result corresponding to the voxel is determined to be a match;

[0098] If the angle is greater than the preset angle, or the distance is greater than the preset distance, it is determined that the matching result corresponding to the voxel is not matched.

[0099] Optionally, the pose deviation of each point cloud frame includes: a six-degree-of-freedom pose deviation;

[0100] The point cloud map freshness update warning device 200 further includes: a posture deviation judgment module;

[0101] The posture deviation judgment module is used to determine that the posture deviation of each point cloud frame is less than the preset posture deviation if the six-degree-of-freedom posture deviation of each point cloud frame is less than the corresponding deviation threshold; otherwise, determine that the posture deviation of each point cloud frame is greater than or equal to the preset posture deviation.

[0102] The specific details of each module or unit in the above device have been described in detail in the corresponding method, so they will not be repeated here.

[0103] It should be noted that, although several modules or units of the equipment for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present application, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into being embodied by multiple modules or units.

[0104] In an exemplary embodiment of the present application, an electronic device is also provided, comprising: a processor; a memory for storing processor executable instructions; wherein the processor is configured to execute the above-mentioned point cloud map freshness update warning method in this example implementation.

[0105] Figure 3 This is a schematic diagram of the structure of an electronic device in an embodiment of the present application. It should be noted that: Figure 3 The electronic device 300 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0106] like Figure 3 As shown, electronic device 300 includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage part 308 into a random access memory (RAM) 303. In RAM 303, various programs and data required for system operation are also stored. Central processing unit 301, ROM 302 and RAM 303 are connected to each other via a bus 304. Input / output (I / O) interface 305 is also connected to bus 304.

[0107] The following components are connected to the I / O interface 305: an input section 306 including a keyboard, a mouse, etc.; an output section 307 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a local area network (LAN) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0108] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart 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 contains program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication part 309, and / or installed from a removable medium 311. When the computer program is executed by the central processing unit 301, various functions defined in the device of the present application are executed.

[0109] In an embodiment of the present application, a computer-readable storage medium is further provided, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned point cloud map freshness update warning method is implemented.

[0110] It should be noted that the computer-readable storage medium shown in the present application 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, a read-only memory, 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. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, radio frequency, etc., or any suitable combination of the above.

[0111] In an embodiment of the present application, a computer program product is also provided. When the computer program product is run on a computer, the computer executes the above-mentioned point cloud map freshness update warning method.

[0112] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0113] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A point cloud map freshness update warning method, characterized in that: include: Acquire real-time point cloud data collected by the autonomous driving vehicle during the driving trajectory, match the real-time point cloud data with the point cloud map, and determine the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment; Constructing a normal distribution transformation map according to the point cloud data of the target trajectory segment in the point cloud map; For each point cloud frame of the target trajectory segment, matching features of the real-time point cloud data in the point cloud frame with features in the normal distribution transformation map to determine a matching result of the point cloud frame; If there are no consecutive N point cloud frames whose corresponding matching results are all unmatched, it is determined that there is no abnormality in the map freshness of the target trajectory segment; N is a positive integer; If there are N consecutive point cloud frames whose corresponding matching results are all unmatched, multi-sensor data of the target trajectory segment is acquired, and fusion positioning is performed according to the multi-sensor data to obtain a fusion positioning posture; For each point cloud frame of the target trajectory segment, matching and positioning are performed according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching and positioning posture of the point cloud frame; Determining a posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame; If the posture deviations corresponding to the N consecutive point cloud frames are all smaller than the preset posture deviations, it is determined that the map freshness of the target trajectory segment is abnormal, and an early warning is issued.

2. The method according to claim 1, characterized in that The matching of the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map to determine the matching result of the point cloud frame includes: Dividing the real-time point cloud data in the point cloud frame into a plurality of different voxels, and calculating Gaussian distribution characteristics of the point cloud data in different voxels; Matching the Gaussian distribution features of the point cloud data in each voxel with the Gaussian distribution features of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel; If the number of voxels in the point cloud frame corresponding to the matching result is greater than or equal to M, then the matching result of the point cloud frame is determined to be a match; M is a positive integer; If the number of voxels in the point cloud frame corresponding to the matching result is less than M, it is determined that the matching result of the point cloud frame is not matched.

3. The method according to claim 2, characterized in that The Gaussian distribution features of the point cloud data in each voxel include: the feature vector of the point cloud data in each voxel and the position coordinates of the centroid of the point cloud data in each voxel; The step of matching the Gaussian distribution features of the point cloud data in each voxel with the Gaussian distribution features of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel includes: For each voxel, calculating the angle between the feature vector of the point cloud data in the voxel and the feature vector of the corresponding point cloud data in the voxel in the normal distribution transformation map; and, Calculating the distance between the centroid of the point cloud data within the voxel and the centroid of the corresponding point cloud data within the voxel in the normal distribution transformation map; If the angle is less than or equal to a preset angle, and the distance is less than or equal to a preset distance, determining that the matching result corresponding to the voxel is a match; If the angle is greater than a preset angle, or the distance is greater than a preset distance, it is determined that the matching result corresponding to the voxel is not matched.

4. The method according to claim 1, characterized in that: The pose deviation of each point cloud frame includes: six-degree-of-freedom pose deviation; Determine whether the pose deviation of each point cloud frame is less than the preset pose deviation, including: If the six-degree-of-freedom posture deviation of each point cloud frame is respectively less than the corresponding deviation threshold, it is determined that the posture deviation of each point cloud frame is less than the preset posture deviation; Otherwise, it is determined that the pose deviation of each point cloud frame is greater than or equal to the preset pose deviation.

5. A point cloud map freshness update warning device, characterized in that: The device comprises: A real-time point cloud data acquisition module is used to acquire real-time point cloud data collected by the autonomous driving vehicle during its driving trajectory; a target trajectory segment determination module, used to match the real-time point cloud data with the point cloud map, and determine the trajectory segment corresponding to the point cloud data with a matching confidence less than a preset confidence as the target trajectory segment; A normal distribution transformation map construction module, used to construct a normal distribution transformation map according to the point cloud data of the target trajectory segment in the point cloud map; A point cloud matching module, for each point cloud frame of the target trajectory segment, matches the features of the real-time point cloud data in the point cloud frame with the features in the normal distribution transformation map, and determines the matching result of the point cloud frame; A map freshness non-abnormal determination module is used to determine that the map freshness of the target trajectory segment does not have abnormality if there are no consecutive N point cloud frames whose corresponding matching results are all unmatched; N is a positive integer; A fusion positioning posture determination module is used to obtain multi-sensor data of the target trajectory segment if there are N consecutive point cloud frames whose corresponding matching results are all unmatched, and perform fusion positioning according to the multi-sensor data to obtain a fusion positioning posture; A point cloud matching positioning posture determination module is used to match and locate each point cloud frame of the target trajectory segment according to the real-time point cloud data of the point cloud frame and the point cloud data in the point cloud map to obtain the point cloud matching positioning posture of the point cloud frame; A posture deviation determination module, used to determine the posture deviation of the point cloud frame according to the fused positioning posture and the point cloud matching positioning posture of the point cloud frame; The point cloud map freshness warning module is used to determine that the map freshness of the target trajectory segment is abnormal and issue a warning if the posture deviations corresponding to the N consecutive point cloud frames are all less than the preset posture deviations.

6. The device according to claim 5, characterized in that The point cloud matching module is specifically used to divide the real-time point cloud data in each point cloud frame of the target trajectory segment into a plurality of different voxels, calculate the Gaussian distribution characteristics of the point cloud data in different voxels; match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map to obtain a matching result corresponding to each voxel; if the number of voxels corresponding to the matching result in the point cloud frame is greater than or equal to M, then determine that the matching result of the point cloud frame is a match; If the number of voxels in the corresponding matching result in the point cloud frame is less than M, it is determined that the matching result of the point cloud frame is not matched, and M is a positive integer.

7. The device according to claim 6, characterized in that The Gaussian distribution features of the point cloud data in each voxel include: the feature vector of the point cloud data in each voxel and the position coordinates of the centroid of the point cloud data in each voxel; The point cloud matching module is specifically used to match the Gaussian distribution characteristics of the point cloud data in each voxel with the Gaussian distribution characteristics of the point cloud data in the corresponding voxel in the normal distribution transformation map through the following steps to obtain a matching result corresponding to each voxel: For each voxel, calculating the angle between the feature vector of the point cloud data in the voxel and the feature vector of the corresponding point cloud data in the voxel in the normal distribution transformation map; and, Calculating the distance between the centroid of the point cloud data within the voxel and the centroid of the corresponding point cloud data within the voxel in the normal distribution transformation map; If the angle is less than or equal to a preset angle, and the distance is less than or equal to a preset distance, determining that the matching result corresponding to the voxel is a match; If the angle is greater than a preset angle, or the distance is greater than a preset distance, it is determined that the matching result corresponding to the voxel is not matched.

8. The device according to claim 5, characterized in that The pose deviation of each point cloud frame includes: six-degree-of-freedom pose deviation; The device also includes: a posture deviation judgment module; The posture deviation judgment module is used to determine that the posture deviation of each point cloud frame is less than a preset posture deviation if the six-degree-of-freedom posture deviation of each point cloud frame is less than the corresponding deviation threshold; otherwise, determine that the posture deviation of each point cloud frame is greater than or equal to the preset posture deviation.

9. An electronic device, characterized in that: include: A processor, wherein the processor is used to execute a computer program stored in a memory, wherein the computer program, when executed by the processor, implements the method according to any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images