High-precision map generation method, device, electronic device and storage medium

Abstract

The present disclosure provides a map generation method, which relates to the field of artificial intelligence technology, and in particular to high-precision maps, autonomous driving, and intelligent transportation technology. The specific implementation scheme is: based on the multi-frame initial point cloud data of each of the multiple laser radars and the initial trajectory information of the inertial measurement unit, a target point cloud map is obtained; based on the target point cloud map and the multi-frame initial point cloud data of each of the multiple laser radars, multiple association relationship information is obtained; based on the initial trajectory information, multiple association relationship information, and the initial external parameters of each of the multiple laser radars, multiple distance information is respectively determined; based on the multiple distance information, the initial trajectory information and multiple initial external parameters are adjusted to obtain target trajectory information and multiple target external parameters; and based on the target trajectory information and multiple target external parameters, a target map is generated. The present disclosure also provides a map generation device, an electronic device, and a storage medium.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, particularly to high-precision mapping, autonomous driving, and intelligent transportation. More specifically, the present disclosure provides a map generation method, device, electronic device, and storage medium. Background Art

[0002] High-precision maps, also known as high-accuracy maps, can be used by autonomous vehicles. These maps contain precise vehicle location information and rich road element data, helping vehicles predict complex road conditions, such as slope, curvature, and heading, to better mitigate potential risks. With the advancement of artificial intelligence and high-precision mapping technologies, the application scenarios for autonomous driving and assisted driving technologies are continuously expanding. In both autonomous and assisted driving modes, HD maps can be used to determine the location of vehicles and other obstacles, enabling vehicle control. Summary of the Invention

[0003] The present disclosure provides a map generation method, apparatus, device, and storage medium.

[0004] According to one aspect of the present disclosure, a map generation method is provided, the method comprising: obtaining a target point cloud map based on multi-frame initial point cloud data of each of a plurality of laser radars and initial trajectory information of an inertial measurement unit; obtaining a plurality of association relationship information based on the target point cloud map and the multi-frame initial point cloud data of each of the plurality of laser radars, wherein the association relationship information is used to indicate: the association relationship between a plurality of initial points in the multi-frame initial point cloud data of the laser radar and a plurality of surface features of the target point cloud map; determining a plurality of distance information based on the initial trajectory information, the plurality of association relationship information and the initial external parameters of each of the plurality of laser radars, wherein the distance information includes a plurality of distance sub-information, and the distance sub-information is used to indicate the distance between the surface feature corresponding to the initial point and the initial point; adjusting the initial trajectory information and the plurality of initial external parameters based on the plurality of distance information to obtain target trajectory information and a plurality of target external parameters; and generating a target map based on the target trajectory information and the plurality of target external parameters.

[0005] According to another aspect of the present disclosure, a map generation device is provided, which includes: a first acquisition module for obtaining a target point cloud map based on multi-frame initial point cloud data of each of a plurality of laser radars and initial trajectory information of an inertial measurement unit; a second acquisition module for obtaining multiple association relationship information based on the target point cloud map and the multi-frame initial point cloud data of each of the plurality of laser radars, wherein the association relationship information is used to indicate: the association relationship between multiple initial points in the multi-frame initial point cloud data of the laser radar and multiple surface features of the target point cloud map; a determination module for determining multiple distance information based on the initial trajectory information, the multiple association relationship information and the initial external parameters of each of the plurality of laser radars, wherein the distance information includes multiple distance sub-information, and the distance sub-information is used to indicate the distance between the surface feature corresponding to the initial point and the initial point; an adjustment module for adjusting the initial trajectory information and the multiple initial external parameters based on the multiple distance information to obtain target trajectory information and multiple target external parameters; and a generation module for generating a target map based on the target trajectory information and the multiple target external parameters.

[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method provided according to the present disclosure.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided. The computer instructions are used to cause a computer to execute the method provided according to the present disclosure.

[0008] According to another aspect of the present disclosure, a computer program product is provided, including a computer program, which implements the method provided according to the present disclosure when executed by a processor.

[0009] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure.

[0011] Figure 1 is a schematic diagram of an exemplary system architecture to which a map generation method and apparatus according to an embodiment of the present disclosure may be applied;

[0012] Figure 2 is a flowchart of a map generation method according to an embodiment of the present disclosure;

[0013] Figure 3 is a flowchart of a map generating method according to another embodiment of the present disclosure;

[0014] Figure 4A is a schematic diagram of a point cloud stitching effect according to another embodiment of the present disclosure;

[0015] Figure 4B is a schematic diagram of a point cloud stitching effect according to an embodiment of the present disclosure;

[0016] Figure 5 is a block diagram of a map generating apparatus according to an embodiment of the present disclosure; and

[0017] Figure 6 is a block diagram of an electronic device to which a map generating method can be applied according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0019] Laser Radar (LiDAR) can collect data about a target area, pre-process it, and send it to a server. For example, a LiDAR emits a scanning laser beam. When the laser beam hits an object, it is reflected and received by the LiDAR, completing the transmission and reception of the laser beam. This allows for the continuous collection of large amounts of point cloud data.

[0020] An inertial measurement unit (IMU) can determine the trajectory of a target object, such as a vehicle or other device equipped with a lidar and an IMU.

[0021] The extrinsic parameters (Extrinsics Parameters) between the LiDAR and IMU can include the spatial-temporal relationship between them. Spatial parameters can be the transformation between the LiDAR's point cloud coordinate system and the IMU's inertial coordinate system. Temporal parameters can be the synchronization error between the sensor's measurement time. This error is caused by signal processing and communication delays between the LiDAR and IMU. For example, at the same moment, the timestamps of the LiDAR's point cloud data and the inertial measurement data output by the IMU may not match.

[0022] In some embodiments, a data collection vehicle can be equipped with multiple LiDARs. For example, a first-target LiDAR can be mounted on the roof of the data collection vehicle, enabling 360-degree scanning and scanning objects around the vehicle. A second-target LiDAR can be mounted on the rear of the data collection vehicle, enabling ground scanning. For another example, a data collection vehicle can be equipped with four or three LiDARs. When using multiple LiDARs for joint mapping, the external parameters of each LiDAR can be used to fuse the point cloud data from different LiDARs. The accuracy of these external parameters can have a direct impact on the mapping results.

[0023] In the case where multiple lidars are deployed on the data collection vehicle, the external parameters of multiple lidars can be calibrated separately. External parameters may include time extrinsic parameters and space extrinsic parameters. Spatial extrinsic parameters may include 6 degrees of freedom components, namely translation x, translation y, translation z, rotation x, rotation y and rotation z. When calibrating multiple lidars separately, the translation z component of each lidar is not observable, and it can be considered that the component is unchanged and does not need to be optimized. "Unobservable" is a cybernetic term that indicates whether the data can be mined or estimated from existing data. In the embodiment of the present disclosure, the true value change of the translation z component does not lead to changes in other components, that is, other components are insensitive to the true value change of translation z. In this case, the extrinsic elevation error between different lidars is difficult to eliminate, which may cause ground stratification in the spliced point cloud.

[0024] In addition, if the point cloud data collected by the lidar is relatively sparse or there are few identical areas between frames, it may be difficult to splice the multi-frame point cloud data of the lidar, which will lead to certain errors in the motion trajectory of the lidar obtained based on the spliced point cloud data, thereby affecting the external parameter calibration results.

[0025] Figure 1 is a schematic diagram of an exemplary system architecture to which a map generation method and apparatus can be applied according to an embodiment of the present disclosure. It should be noted that: Figure 1 The examples shown are merely examples of system architectures to which the embodiments of the present disclosure may be applied, to help those skilled in the art understand the technical content of the present disclosure, but do not mean that the embodiments of the present disclosure may not be used in other devices, systems, environments or scenarios.

[0026] like Figure 1 As shown, the system architecture 100 according to this embodiment may include sensors 101, 102, and 103, a network 120, a server 130, and a roadside unit (RSU) 140. The network 120 is used as a medium for providing communication links between the sensors 101, 102, and 103 and the server 130. The network 120 may include various connection types, such as wired and / or wireless communication links.

[0027] The sensors 101 , 102 , 103 may interact with the server 130 via the network 120 to receive or send messages, etc.

[0028] Sensors 101, 102, and 103 may be functional components integrated into vehicle 110, such as infrared sensors, ultrasonic sensors, millimeter-wave radars, information acquisition devices, lidars, inertial measurement units, and the like. Sensors 101, 102, and 103 may be used to collect status data of perceived objects (e.g., pedestrians, vehicles, obstacles, etc.) around vehicle 110, as well as surrounding road data.

[0029] The vehicle 110 can communicate with the roadside unit 140, receive information from the roadside unit 140, or send information to the roadside unit.

[0030] The roadside unit 140 may be deployed on a traffic light, for example, to adjust the duration or frequency of the traffic light.

[0031] The server 130 may be located at a remote location capable of establishing communication with the vehicle-mounted terminal, and may be implemented as a distributed server cluster consisting of multiple servers, or as a single server.

[0032] Server 130 can be a server that provides various services. For example, mapping applications and data processing applications can be installed on server 130. For example, server 130 running a data processing application can receive obstacle status data and map data transmitted from sensors 101, 102, and 103 via network 120. One or more of the obstacle status data and map data can be used as data to be processed. The data to be processed is then processed to obtain target data.

[0033] It should be noted that the map generation method provided in the embodiments of the present disclosure can generally be executed by server 130. Accordingly, the map generation device provided in the embodiments of the present disclosure can also be located in server 130. However, this is not a limitation. The map generation method provided in the embodiments of the present disclosure can also generally be executed by sensor 101, 102, or 103. Accordingly, the map generation device provided in the embodiments of the present disclosure can also be located in sensor 101, 102, or 103.

[0034] I understand. Figure 1 The number of sensors, networks, and servers in the embodiment is only illustrative. Any number of sensors, networks, and servers may be used depending on the implementation requirements.

[0035] It should be noted that the sequence numbers of the operations in the following method are only used to indicate the operation for the purpose of description, and should not be regarded as indicating the order in which the operations should be performed. Unless explicitly stated, the method does not need to be performed in the order shown.

[0036] Figure 2 is a flowchart of a map generation method according to an embodiment of the present disclosure.

[0037] like Figure 2 As shown, the method 200 may include operations S210 to S250.

[0038] In operation S210 , a target point cloud map is obtained based on multiple frames of initial point cloud data of each of the multiple laser radars and initial trajectory information of the inertial measurement unit.

[0039] For example, the initial trajectory information can be used to adjust multiple frames of initial point cloud data from multiple lidars to obtain multiple frames of adjusted point cloud data from each of the multiple lidars. Based on the multiple frames of adjusted point cloud data from at least one lidar, a point cloud map can be generated as the target point cloud map.

[0040] In operation S220 , a plurality of association relationship information is obtained based on the target point cloud map and the multiple frames of initial point cloud data of the plurality of laser radars.

[0041] In the embodiment of the present disclosure, the association relationship information may indicate: the association relationship between multiple initial points in multiple frames of initial point cloud data of the laser radar and multiple surface features of the target point cloud map.

[0042] In the disclosed embodiments, a surface feature is a feature of a target plane in a target point cloud map. There can be at least one target plane. For example, multiple target planes can be perpendicular to each other. In one example, there can be three target planes. These three target planes can also be referred to as orthogonal planes.

[0043] In the embodiments of the present disclosure, various methods can be used to extract the surface features of the target plane. For example, the surface features of the target plane can be extracted based on the Lidar Odometry and Mapping in real time (LOAM) algorithm.

[0044] In operation S230 , a plurality of distance information is determined based on the initial trajectory information, the plurality of association relationship information, and the initial external parameters of the plurality of laser radars.

[0045] In the embodiment of the present disclosure, the distance information includes a plurality of distance sub-information, and the distance sub-information may indicate the distance between the surface feature corresponding to the initial point and the initial point.

[0046] In the disclosed embodiment, a coordinate system conversion can be performed based on the initial external parameters so that the initial trajectory information and the association relationship information are in the same coordinate system. For example, the association relationship information can be in the coordinate system of the target point cloud map. The initial trajectory information can come from an inertial coordinate system. Based on the initial external parameters, the initial trajectory information can be converted to the coordinate system of the target point cloud map so that the distance sub-information between the initial point and the surface feature can be determined in the same coordinate system. Multiple distance sub-information related to the lidar can be used as distance information.

[0047] In operation S240 , the initial trajectory information and the plurality of initial external parameters are adjusted according to the plurality of distance information to obtain target trajectory information and a plurality of target external parameters.

[0048] In the disclosed embodiments, the initial external parameters and initial trajectory information can be adjusted in various ways based on the distance information. For example, the adjustment goal can be to reduce the distance indicated by the distance sub-information. In another example, the adjusted initial trajectory information and initial external parameters can be used as the target trajectory information and target external parameters, respectively.

[0049] In operation S250 , a target map is generated based on the target trajectory information and a plurality of target external parameters.

[0050] For example, the target trajectory information can be used to adjust the point cloud data. Then, a point cloud map can be generated based on the point cloud data adjusted with the target trajectory information. Based on the point cloud map, the target map can be obtained.

[0051] Through the disclosed embodiments, multiple external parameters are jointly optimized based on the point cloud data of multiple lidars, with a high success rate for external parameter optimization. The multiple external parameters obtained after optimization can improve the accuracy of joint mapping by multiple lidars.

[0052] In addition, through the embodiments of the present disclosure, when the external parameters of multiple lidars are jointly optimized, only one lidar's translation z component is not significant, while the translation z components of other lidars are significant, which can eliminate the external parameter elevation error to a certain extent and improve the quality of the target map.

[0053] In addition, through the embodiments of the present disclosure, the initial trajectory information and initial external parameters are adjusted according to the distance information, which can also reduce the accuracy requirements for the initial trajectory information or initial external parameters, helping to improve the efficiency of map generation and reduce the cost required for map generation.

[0054] The following describes in detail some implementation methods for generating a target point cloud map in conjunction with relevant embodiments.

[0055] In some embodiments of the aforementioned operation S210, obtaining a target point cloud map based on multiple frames of initial point cloud data from multiple laser radars and initial trajectory information from an inertial measurement unit may include determining point cloud compensation information for each of the multiple laser radars based on the initial trajectory information and initial external parameters of each of the multiple laser radars. Using the multiple point cloud compensation information, respectively, adjust the multiple frames of initial point cloud data from each of the multiple laser radars to obtain multiple frames of target point cloud data from each of the multiple laser radars. Obtaining a target point cloud map based on the multiple frames of target point cloud data from each of the multiple laser radars.

[0056] For example, point cloud compensation information can be determined based on initial external parameters and the initial trajectory information of the inertial measurement unit. Using this point cloud compensation information, motion compensation can be performed on multiple frames of initial point cloud data to obtain multiple frames of target point cloud data. Based on these multiple frames of target point cloud map data, a target point cloud map can be obtained.

[0057] It can be understood that some implementation methods of point cloud compensation are described above, and some implementation methods of obtaining the association relationship will be described in detail below in conjunction with relevant embodiments.

[0058] In some embodiments, in some implementations of the above-mentioned operation S220, obtaining multiple association relationship information based on the target point cloud map and the multiple frames of initial point cloud data of each of the multiple lidars may include: obtaining multiple association relationship information based on the target point cloud map and the multiple frames of target point cloud data of each of the multiple lidars.

[0059] In an embodiment of the present disclosure, at least one target plane may be determined in a target point cloud map. Multiple surface features of the target plane may be extracted. Correlation sub-relationship information between an initial point and the surface feature may be determined. Based on the multiple correlation sub-relationship information, correlation relationship information may be determined.

[0060] For example, the associated sub-relationship information can indicate the surface feature corresponding to the initial point. There can be three target planes. The three target planes can be perpendicular to each other. These three target planes can also be referred to as three orthogonal planes. Based on the lidar odometry and real-time mapping algorithm, surface features in the target planes can be extracted (for example, greater than or equal to 80 surface features can be extracted from each target plane). The associated sub-relationship information between the initial point and the surface feature can be determined so that the initial point corresponds to one surface feature. It is understood that one surface feature can correspond to multiple initial points. The associated sub-relationship information of multiple initial points can be used as the associated relationship information.

[0061] It will be appreciated that the above describes some methods for generating a target point cloud map and determining association relationship information. In the disclosed embodiments, the multiple lidars may include a first target lidar as a master lidar and a second target lidar as a slave lidar. Based on this, the following will further explain some methods for obtaining a target point cloud map and determining association relationship information in conjunction with relevant implementations.

[0062] In some embodiments, obtaining a target point cloud map based on multiple frames of target point cloud data from multiple laser radars may include obtaining first motion trajectory information of the first target laser radar based on the multiple frames of target point cloud data from the first target laser radar, and obtaining the target point cloud map based on the first motion trajectory information.

[0063] For example, for a primary target lidar radar, multiple frames of target point cloud data can be stitched together to obtain first motion trajectory information. Based on this first motion trajectory information, at least one frame of key point cloud data can be identified from the multiple frames of target point cloud data. By stitching together the at least one frame of key point cloud data, a target point cloud map can be obtained.

[0064] Next, after obtaining a point cloud map using the multi-frame target point cloud data of the master radar, the respective correlation relationship information of the master radar and the slave radar can be determined respectively.

[0065] In some embodiments, obtaining multiple association relationship information based on the target point cloud map and the multi-frame target point cloud data of each of the multiple laser radars may include: obtaining first splicing relationship information based on the first motion trajectory information. The first association relationship information is obtained based on the first splicing relationship information. For example, the first splicing relationship between the multi-frame target point cloud data of the first target laser radar and the target point cloud map can be obtained based on the first motion trajectory information. Next, the first association relationship information can be determined based on the splicing relationship information. The first association relationship information may indicate: the association relationship between multiple initial points in the multi-frame target point cloud data of the first target laser radar and multiple surface features of the target point cloud map.

[0066] In some embodiments, obtaining multiple association relationship information based on the target point cloud map and the multi-frame target point cloud data of each of the multiple laser radars may include: splicing the multi-frame target point cloud data of each of the at least one second target laser radar with the target point cloud map to obtain the second splicing relationship information of each of the at least one second target laser radar. At least one second association relationship information is obtained based on the at least one second splicing relationship information. For example, the multi-frame target point cloud data of the second target laser radar and the target point cloud map may be subjected to point cloud splicing to obtain the second splicing relationship information. The second association relationship information may be obtained based on the second splicing relationship information. The second association relationship information may indicate: the association relationship between multiple initial points in the multi-frame target point cloud data of the second target laser radar and multiple surface features of the target point cloud map. Through the embodiments of the present disclosure, the point cloud data of the main radar with a wide viewing angle or multiple light beams is fully utilized, the point cloud map can be quickly established, and the efficiency of target map generation is improved.

[0067] It will be appreciated that the target point cloud map was obtained above based on the point cloud data from the master radar. In the disclosed embodiments, the target point cloud map can also be obtained using the point cloud data from the slave radar. Based on this, the following will further describe other methods for generating the target point cloud map and determining association information, in conjunction with related implementations.

[0068] In some embodiments, the multiple frames of target point cloud data of the first target lidar may include: at least one frame of first target point cloud data and at least one frame of second target point cloud data.

[0069] In some embodiments, obtaining a target point cloud map based on multiple frames of target point cloud data from multiple laser radars includes: obtaining a spliced point cloud map based on at least one frame of first target point cloud data from a first target laser radar. Splicing the spliced point cloud map with at least one frame of second target point cloud data from the first target laser radar and multiple frames of target point cloud data from at least one second target laser radar to obtain first motion trajectory information from the first target laser radar and second motion trajectory information from at least one second target laser radar. The target point cloud map is obtained based on the first motion trajectory information and the at least one second motion trajectory information.

[0070] For example, at least one frame of first target point cloud data can be stitched together to obtain a stitched point cloud map. After determining that the target point cloud data from the slave radar can be stitched together with the stitched point cloud map, it can be determined whether the coverage and point density of the stitched point cloud map are sufficient. The stage of obtaining the stitched point cloud map can also be referred to as the boosting stage. Next, at least one frame of second target point cloud data not used to generate the stitched point cloud map can be stitched together with the stitched point cloud map. Multiple frames of target point cloud data from at least one second target lidar can also be stitched together with the stitched point cloud map to obtain first motion trajectory information and at least one second motion trajectory information. Based on the first motion trajectory information and the at least one second motion trajectory information, at least one frame of key point cloud data can be determined from the multiple frames of target point cloud data from each of the master radar and the slave radar. By stitching together the at least one frame of key point cloud data, a target point cloud map can be obtained.

[0071] Next, after obtaining a point cloud map using multi-frame target point cloud data from the master radar and the slave radar, the respective association relationship information of the master radar and the slave radar can be determined respectively.

[0072] In some embodiments, obtaining multiple association relationship information based on the target point cloud map and multiple frames of target point cloud data from multiple laser radars may include: obtaining first association relationship information based on first motion trajectory information, and obtaining at least one second association relationship information based on at least one second motion trajectory information.

[0073] For example, a first splicing relationship between the multi-frame target point cloud data of the first target lidar and the target point cloud map can be obtained based on the first motion trajectory information. Next, first association relationship information can be determined based on the splicing relationship information. The first association relationship information can indicate the association relationship between multiple initial points in the multi-frame target point cloud data of the first target lidar and multiple surface features in the target point cloud map.

[0074] For example, the second splicing relationship between the multi-frame target point cloud data of the second target laser radar and the target point cloud map can be obtained based on the second motion trajectory information. Next, the second association relationship information can be determined based on the splicing relationship information. The second association relationship information can indicate: the association relationship between multiple initial points in the multi-frame target point cloud data of the second target laser radar and multiple surface features of the target point cloud map. Through the embodiment of the present disclosure, different laser radars have different viewing angles. When establishing the target point cloud map, the point cloud data of the master radar and the slave radar are fully utilized, which reduces the difference in point cloud density in different areas of the map, improves the distribution of surface features, makes the surface features more uniform, and also helps to improve the quality of the target map.

[0075] It will be appreciated that the above description uses multiple lidars, including a master and slave radars, as an example to illustrate some methods for obtaining a target point cloud map and determining association relationships. However, the present disclosure is not limited to this, and the multiple lidars can be equivalent. The following further describes other methods for generating a target point cloud map and determining association relationship information, in conjunction with relevant implementations.

[0076] In some embodiments, obtaining a target point cloud map based on multiple frames of target point cloud data from multiple laser radars may include obtaining motion trajectory information of the multiple laser radars based on the multiple frames of target point cloud data from the multiple laser radars, and obtaining multiple target point cloud maps based on the multiple motion trajectory information.

[0077] For example, for each LiDAR, multiple frames of target point cloud data can be stitched together to obtain motion trajectory information. Based on this motion trajectory information, at least one frame of key point cloud data can be identified from the multiple frames of corresponding target point cloud data. By stitching together the at least one frame of key point cloud data, a target point cloud map corresponding to each LiDAR can be obtained.

[0078] Next, after obtaining multiple target point cloud maps, the respective association relationship information of each lidar can be determined.

[0079] In some embodiments, obtaining multiple association relationship information based on the target point cloud map and multiple frames of target point cloud data from multiple lidars may include: for any lidar, using the target point cloud map obtained based on the motion trajectory of the lidar as a first target point cloud map of the lidar, and obtaining third association relationship information based on the motion trajectory information of the lidar.

[0080] For example, multiple laser radars may include laser radar Lidar_A, laser radar Lidar_B and laser radar Lidar_C. Based on the motion trajectory information of laser radar Lidar_A, the target point cloud map Map_A is obtained. Based on the motion trajectory information of laser radar Lidar_B, the target point cloud map Map_B is obtained. Based on the motion trajectory information of laser radar Lidar_C, the target point cloud map Map_C is obtained. The target point cloud map Map_A is used as the first target point cloud map of laser radar Lidar_A. Based on the motion trajectory information of laser radar Lidar_A, third association relationship information can be obtained. The third association relationship information can indicate: the association relationship between multiple initial points in the multi-frame target point cloud data of the laser radar and multiple surface features of the first target point cloud map.

[0081] In some embodiments, obtaining multiple association relationship information based on the target point cloud map and multiple frames of target point cloud data from multiple lidars may further include: for any lidar, using target point cloud maps other than the first target point cloud map as second target point cloud maps; splicing the multiple frames of target point cloud data from the lidar with the at least one second target point cloud map to obtain at least one piece of splicing relationship information; and obtaining at least one fourth piece of association relationship information based on the at least one piece of splicing relationship information.

[0082] For example, the target point cloud map Map_A is used as the first target point cloud map of the laser radar Lidar_A. The target point cloud map Map_C and the target point cloud map Map_B are respectively used as the second target point cloud maps of the laser radar Lidar_A. The multi-frame target point cloud data of the laser radar Lidar_A are spliced with the target point cloud map Map_B and the target point cloud map Map_C respectively, and two splicing relationship information can be obtained. According to these splicing relationship information, two fourth association relationship information can be obtained. The fourth association relationship information can indicate: the association relationship between multiple initial points in the multi-frame target point cloud data of the laser radar and multiple surface features of the second target point cloud map. Through the embodiment of the present disclosure, when there is no obvious advantage in the viewing angles of different laser radars, the point cloud data of multiple laser radars can be used to respectively establish point cloud maps, and the point cloud data of different laser radars can be fully utilized to improve the quality of the target map.

[0083] It is understood that the above describes various implementation methods for obtaining association information. After obtaining the association information, distance information can be determined. Some implementation methods for determining distance information will be described in detail below in conjunction with relevant embodiments.

[0084] In some embodiments of the aforementioned operation S230, determining the plurality of distance information based on the initial trajectory information, the plurality of association relationship information, and the initial external parameters of the plurality of lidars may include: determining initial rotation estimation information and initial displacement estimation information based on the initial trajectory information; obtaining target point cloud map coordinate data of the initial point based on the initial external parameters, the point cloud coordinate data of the initial point, the initial rotation estimation information, and the initial displacement estimation information; and determining the distance information based on the target point cloud map coordinate data and the association relationship information.

[0085] In the embodiment of the present disclosure, the initial rotation estimation information may include an initial rotation estimation value. The initial displacement estimation information may include an initial displacement estimation value. For example, according to the initial trajectory information , the initial rotation estimate can be determined and the initial displacement estimate .

[0086] In an embodiment of the present disclosure, the point cloud coordinate data is used to indicate the position of the initial point in the coordinate system of the initial point cloud data, and the target point cloud map coordinate data is used to indicate the position of the initial point in the coordinate system of the target point cloud map.

[0087] In an embodiment of the present disclosure, obtaining target point cloud map coordinate data of the initial point based on initial external parameters, point cloud coordinate data of the initial point, initial rotation estimation information, and initial displacement estimation information may include: obtaining inertial coordinate data of the initial point based on the initial external parameters and the point cloud coordinate data; determining transformation information based on the initial rotation estimation information and the initial displacement estimation information; and processing the inertial coordinate data using the transformation information to obtain target point cloud map coordinate data.

[0088] For example, the inertial coordinate data is used to indicate the position of the initial point in the inertial coordinate system. For the i-th laser radar of the multiple laser radars, according to the initial external parameters And the point cloud data of the initial point k , we can get the inertial coordinate data of the initial point k The inertial coordinate data may indicate the position of the initial point k in the inertial coordinate system. For example, according to the initial rotation estimate and the initial displacement estimate , the transformation matrix can be determined The transformation matrix can be used as transformation information. For example, by multiplying the transformation matrix and the inertial coordinate data, the target point cloud map coordinate data of the initial point k can be obtained. For another example, i may be an integer greater than 1.

[0089] In the embodiment of the present disclosure, the surface feature corresponding to the initial point is determined based on the associated sub-relationship information included in the associated relationship information. For example, based on the target point cloud map coordinate data of the initial point k and the characteristic parameters of the surface feature corresponding to the initial point k , the distance measurement residual can be determined , as the distance sub-information. The distance sub-information of the multiple initial points is used as the distance information of the corresponding lidar. In one example, based on the above-mentioned lidar odometer measurement and real-time mapping algorithm, the distance measurement residual can be determined.

[0090] It is understood that some embodiments for determining distance information have been described in detail above. In order to adjust the initial external parameters and initial trajectory information, acceleration information and angular velocity information may also be obtained, which will be described in detail below.

[0091] In some embodiments, in some implementations of operation S240, adjusting the initial trajectory information and the multiple initial external parameters based on the multiple distance information to obtain the target trajectory information and the multiple target external parameters may include: adjusting the multiple initial external parameters and the initial trajectory information based on the acceleration information, the angular velocity information, and the multiple distance information.

[0092] In some embodiments, the inertial measurement data collected by the inertial measurement unit may include inertial rotation data and inertial translation data.

[0093] In some embodiments, acceleration information may be determined based on the inertial translation data and the initial trajectory information.

[0094] In the embodiment of the present disclosure, the initial rotation estimation information and the initial acceleration estimation information are determined based on the initial trajectory information. For example, the initial acceleration estimation information may include the initial acceleration estimation value .

[0095] In the embodiment of the present disclosure, initial acceleration measurement information can be determined based on the inertial translation data. For example, the initial acceleration measurement information can include the initial acceleration measurement value .

[0096] In the embodiment of the present disclosure, acceleration information may be determined according to the initial rotation estimation information, the initial acceleration estimation information, and the initial acceleration measurement information.

[0097] In the embodiment of the present disclosure, the first acceleration estimation information is processed by performing a first operation on the initial acceleration estimation information using the first preset acceleration information to obtain the first acceleration estimation information. For example, the first preset acceleration information may include the gravity acceleration value For another example, the first operation process may include a subtraction operation process. In one example, the gravity acceleration value and the initial acceleration estimate The difference between them can be used as the first acceleration estimation information.

[0098] In the embodiment of the present disclosure, the first acceleration estimation value is subjected to a second operation process using the initial rotation estimation information to obtain the second acceleration estimation information. For example, the second operation process may include a multiplication operation process. For another example, the initial acceleration estimation value may be converted to obtain the initial rotation estimation value. The reciprocal of The product obtained by multiplying the above difference can be used as the second acceleration estimation information.

[0099] In the embodiment of the present disclosure, the acceleration difference information between the initial acceleration measurement information and the second acceleration estimation information can be determined. For example, the initial acceleration measurement value can be used to determine the acceleration difference information between the initial acceleration measurement information and the second acceleration estimation information. Subtracting the above product, the acceleration difference value is obtained as the acceleration difference information.

[0100] In the embodiment of the present disclosure, acceleration information can be obtained by adjusting the acceleration difference information using the preset translation deviation. For example, the above acceleration difference value and the preset translation deviation are Adding them together, we can get the acceleration measurement residual , as acceleration information.

[0101] In one example, the acceleration measurement residual can be determined by the following formula :

[0102] (Formula 1)

[0103] It can be the first acceleration estimated value mentioned above. It can be the second acceleration estimate. Can be an initial rotation estimate. It can be an estimate of the initial acceleration. This can be the initial acceleration measurement. The translation deviation can be preset. It can be the gravity acceleration value.

[0104] In some embodiments, angular velocity information may be determined based on the inertial rotation data and the initial trajectory information.

[0105] In the embodiment of the present disclosure, the initial angular velocity estimation information can be determined based on the initial trajectory information. For example, the initial angular velocity estimation information may include the initial angular velocity estimation value .

[0106] In the embodiment of the present disclosure, the initial angular velocity measurement information can be determined based on the inertial rotation data. For example, the initial angular velocity measurement information may include the initial angular velocity measurement value .

[0107] In the embodiment of the present disclosure, angular velocity information may be determined according to the initial angular velocity estimation information and the initial angular velocity measurement information.

[0108] In the embodiment of the present disclosure, the angular velocity difference information between the initial angular velocity measurement information and the initial angular velocity estimation information can be determined. For example, the initial angular velocity measurement value Subtract the initial angular velocity estimate , and obtain the angular velocity difference value as the angular velocity difference information.

[0109] In the embodiment of the present disclosure, the angular velocity difference information is adjusted using the preset rotation deviation to obtain the angular velocity information. For example, the angular velocity measurement residual can be obtained by subtracting the preset rotation deviation from the angular velocity difference value. , as angular velocity information.

[0110] In one example, the angular velocity measurement residual can be determined by the following formula :

[0111] (Formula 2)

[0112] It can be an angular velocity difference value. It can be an estimate of the initial angular velocity. It can be the initial angular velocity measurement value. A rotation deviation can be preset.

[0113] It is understood that the above describes in detail the method for obtaining acceleration information, angular velocity information, and distance information. The following describes in detail some implementation methods for adjusting initial external parameters and initial trajectory information.

[0114] In some embodiments, adjusting initial external parameters and initial trajectory information based on acceleration information, angular velocity information, and multiple distance information to obtain target trajectory information and target external parameters may include determining fusion information based on the acceleration information, angular velocity information, and distance information. Adjusting the initial trajectory information and initial external parameters so that the fusion information converges to obtain the target trajectory information and target external parameters.

[0115] For example, the fusion information of the i-th lidar can be implemented as the objective function shown in the following formula:

[0116] (Formula 3)

[0117] It can be used to find the variable value that minimizes the function value of the objective function. It can be the target trajectory information, It can be the target external parameter of the i-th lidar. 、 、 They can be respectively the target gravity acceleration value, target translation deviation, target rotation deviation, and can be respectively compared with the gravity acceleration value , preset translation deviation , preset rotation deviation Consistent, can be considered as an unchanging value. For example, the initial trajectory information and initial external parameters are adjusted to obtain the target trajectory information and target external parameters. The goal of the adjustment is to reduce or converge the function value of the objective function shown in Formula 3 calculated based on the target trajectory information and target external parameters.

[0118] It will be appreciated that the above describes in detail the method for adjusting the initial trajectory information and initial external parameters. In the embodiment of the present disclosure, the original trajectory information and original external parameters can be iterated N times to achieve convergence of the fusion information. The above-mentioned initial trajectory information and initial external parameters can be used as the initial trajectory information of the nth iteration and the initial external parameters of the nth iteration, respectively. The above-mentioned target trajectory information and target external parameters can be used as the target trajectory information of the nth iteration and the target external parameters of the nth iteration, respectively. This will be described in detail below in conjunction with relevant embodiments.

[0119] Figure 3 is a flowchart of a map generation method according to an embodiment of the present disclosure.

[0120] like Figure 3 As shown, each operation in method 300 is used to perform the nth iteration, where n is an integer greater than or equal to 1. n is an integer less than or equal to N. N is an integer greater than 1. In an embodiment of the present disclosure, before performing the first iteration, an initialization operation may be performed to determine the original trajectory information and the original external parameters. For example, the original trajectory information may be used as the initial trajectory information for the first iteration. The original external parameters may also be determined as the initial external parameters for the first iteration. For another example, initialization may be performed in various ways to obtain the original trajectory information and the original external parameters. In one example, the local trajectory output by the LiDAR Inertial Odometry (LIO) or the trajectory output by the vehicle-side positioning module may be used as the original trajectory information.

[0121] In operation S310 , a target point cloud map of the nth iteration is obtained based on the initial trajectory information of the nth iteration and multiple frames of initial point cloud data of each of the multiple laser radars.

[0122] For example, based on the initial extrinsic parameters and initial trajectory information of the nth iteration, the nth-level point cloud compensation information can be determined. The nth-level point cloud compensation information is then used to perform motion compensation on the initial point cloud data to obtain the nth-level target point cloud data. Point cloud stitching is then performed on multiple frames of the nth-level target point cloud data to obtain the target point cloud map of the nth iteration.

[0123] In operation S320 , a plurality of n-th level association relationship information is obtained based on the target point cloud map of the n-th iteration and the multi-frame n-th level target point cloud data of each of the plurality of laser radars.

[0124] For example, for each lidar, three target planes can be determined in the target point cloud map of the nth iteration, the surface features of the target planes can be extracted, and the nth-level association sub-relationship information between the initial points and the surface features in the nth-level target point cloud data of each lidar can be determined. The nth-level association sub-relationship information of multiple initial points of each lidar is used as the nth-level association relationship information.

[0125] In operation S330 , n-th level acceleration information, n-th level angular velocity information, and a plurality of n-th level distance information are determined based on the inertial measurement data, the initial external parameters of the n-th iteration, the initial trajectory information of the n-th iteration, and a plurality of n-th level association relationship information.

[0126] For example, based on the inertial translation data and the initial trajectory information of the nth iteration, the nth-level acceleration measurement residual can be determined as the nth-level acceleration information.

[0127] For example, based on the inertial rotation data and the initial trajectory information of the n-th iteration, the n-th level angular velocity measurement residual can be determined as the n-th level angular velocity information.

[0128] For example, for each lidar, based on the initial external parameters of the nth iteration, the initial trajectory information of the nth iteration, and multiple nth-level associated sub-relationship information, multiple distance measurement residuals can be determined respectively as multiple nth-level distance sub-information. Multiple nth-level distance sub-information can be used as the nth-level distance information of a lidar.

[0129] In operation S341 , n-th level fusion information is determined based on n-th level acceleration information, n-th level angular velocity information, and n-th level distance information.

[0130] For example, the objective function shown in Formula 3 is used to process the n-th level acceleration measurement residual, the n-th level angular velocity measurement residual, and the n-th level distance measurement residual to obtain the n-th level function value as the n-th level fusion information.

[0131] In operation S342 , it is determined whether the n-th level fusion information is converged.

[0132] For example, it can be determined whether the difference between the nth-level function value and the n-1th-level function value is less than a preset difference threshold. If so, it can be determined that the nth-level fusion information has converged. It will be appreciated that other methods can also be used to determine whether the fusion information has converged. For another example, when n=1, it can be determined that the first-level fusion information has not converged.

[0133] In the embodiment of the present disclosure, in response to determining that the n-th level fusion information has converged, operation S350 is performed. For example, operation S350 may be performed when the n-th level fusion information of each of the plurality of lidars has converged.

[0134] In the embodiment of the present disclosure, in response to determining that the n-th level fusion information has not converged, operation S343 is performed. For example, operation S343 may be performed when the n-th level fusion information of any one of the multiple laser radars has not converged.

[0135] In operation S343 , the initial external parameters of the n-th iteration and the initial trajectory information of the n-th iteration are adjusted to obtain the target external parameters of the n-th iteration and the target trajectory information of the n-th iteration.

[0136] For example, for each lidar, various methods can be used to adjust the initial extrinsic parameters and initial trajectory information of the nth iteration to obtain the target extrinsic parameters and target trajectory sub-information of the nth iteration. The target trajectory sub-information of the nth iteration from multiple lidars can be fused to obtain the target trajectory information of the nth iteration. It is understood that the target trajectory sub-information can be fused by averaging.

[0137] In operation S344 , the target external parameters of the nth iteration and the target trajectory information of the nth iteration are used as the initial external parameters of the (n+1)th iteration and the initial trajectory information of the (n+1)th iteration, respectively, and the process returns to operation S310 .

[0138] For example, the n+1th iteration is performed according to the initial trajectory information of the n+1th iteration and the initial external parameters of the n+1th iteration of each of the multiple lidars.

[0139] In operation S350 , a target map is generated based on trajectory information and external parameters corresponding to the converged fusion information.

[0140] For example, when the n-th level fusion information converges, the target trajectory information of the n-1th iteration and the target external parameters of the n-1th iteration can be used to generate the target map. For another example, when the fusion information converges, the target trajectory information of the n-th iteration and the target external parameters of the n-1th iteration can be used to generate the target map. It can be understood that in the two iterations before and after the fusion information converges, the difference between the two related trajectory information is small, and at least one of the trajectory information can be used to generate the target map. It can also be understood that in the two iterations before and after the fusion information converges, the difference between the two related external parameters is small and can be used to generate the target map.

[0141] In the disclosed embodiment, if the Nth-level fusion information converges after N iterations, the Nth-level trajectory information can be determined. Based on this Nth-level trajectory information, the Nth-level point cloud compensation information can be determined. The Nth-level point cloud compensation information is used to perform motion compensation on the initial point cloud data to obtain the Nth-level target point cloud data. Multiple frames of the Nth-level target point cloud data are then spliced to obtain a target map.

[0142] It is understood that some embodiments of generating a map have been described in detail above. In order to improve the accuracy of the generated map, some information or data may be verified during the above-mentioned operations, which will be described in detail below.

[0143] In some embodiments, determining the associated sub-relationship information between the initial point and the surface feature includes: in response to determining that the surface feature data is greater than or equal to a preset surface feature number threshold, determining the associated sub-relationship information between the initial point and the surface feature. For example, the preset surface feature number threshold may be 80. If the number of surface features is less than 80, the previous iteration may be re-executed to readjust the initial trajectory information and / or initial external parameters of the previous iteration.

[0144] The following will further illustrate the effect of joint adjustment of the external parameters of multiple lidars in conjunction with relevant embodiments.

[0145] Figure 4A 3 is a schematic diagram of a point cloud stitching effect according to another embodiment of the present disclosure.

[0146] In other embodiments, the external parameters of multiple laser radars can be adjusted independently. After the adjustment is completed, the point cloud data of multiple laser radars can be spliced to obtain the following: Figure 4A The splicing results are shown in Figure 2. Figure 4A As shown in the figure, when two external parameters are optimized independently, the point cloud stitching result may have obvious faults and large errors, resulting in an unreasonable increase in the thickness of the object.

[0147] Figure 4B 3 is a schematic diagram of a point cloud stitching effect according to an embodiment of the present disclosure.

[0148] In some embodiments, the external parameters of multiple laser radars can be adjusted jointly according to method 200 or 300. After the adjustment is completed, the point cloud data of the multiple laser radars can be spliced to obtain the following: Figure 4B The splicing results are shown in Figure 2. Figure 4B As shown in the figure, when the two external parameters are jointly optimized, the error of the point cloud stitching result is smaller, the thickness of the object is more reasonable, and it is closer to the real object.

[0149] Figure 5 is a block diagram of a map generating apparatus according to an embodiment of the present disclosure.

[0150] like Figure 5 As shown, the apparatus 500 may include a first obtaining module 510 , a second obtaining module 520 , a determining module 530 , an adjusting module 540 and a generating module 550 .

[0151] The first acquisition module 510 is used to obtain a target point cloud map based on the multi-frame initial point cloud data of each of the multiple laser radars and the initial trajectory information of the inertial measurement unit.

[0152] The second obtaining module 520 is configured to obtain a plurality of association relationship information based on the target point cloud map and the multiple frames of initial point cloud data from the multiple lidars. For example, the association relationship information may indicate association relationships between multiple initial points in the multiple frames of initial point cloud data from the lidars and multiple surface features in the target point cloud map.

[0153] Determination module 530 is configured to determine multiple distance information based on the initial trajectory information, the multiple association relationship information, and the initial external parameters of the multiple lidars. For example, the distance information includes multiple distance sub-information, each of which indicates the distance between the surface feature corresponding to the initial point and the initial point.

[0154] The adjustment module 540 is configured to adjust the initial trajectory information and the multiple initial external parameters according to the multiple distance information to obtain the target trajectory information and the multiple target external parameters.

[0155] The generation module 550 is used to generate a target map according to the target trajectory information and multiple target external parameters.

[0156] In some embodiments, the first acquisition module includes: a first determination submodule for determining point cloud compensation information for each of the multiple lidars based on initial trajectory information and initial external parameters of each of the multiple lidars; a first adjustment submodule for adjusting multiple frames of initial point cloud data for each of the multiple lidars using the multiple point cloud compensation information to obtain multiple frames of target point cloud data for each of the multiple lidars; and a first acquisition submodule for obtaining a target point cloud map based on the multiple frames of target point cloud data for each of the multiple lidars.

[0157] In some embodiments, the second acquisition module is further used to obtain multiple association relationship information based on the target point cloud map and the multi-frame target point cloud data of multiple laser radars.

[0158] In some embodiments, the plurality of laser radars includes a first target laser radar. The first acquisition submodule includes: a first acquisition unit configured to obtain first motion trajectory information of the first target laser radar based on multiple frames of target point cloud data from the first target laser radar; and a second acquisition unit configured to obtain a target point cloud map based on the first motion trajectory information.

[0159] In some embodiments, the second obtaining module includes: a second obtaining submodule configured to obtain first splicing relationship information based on the first motion trajectory information; and a third obtaining submodule configured to obtain first association relationship information based on the first splicing relationship information.

[0160] In some embodiments, the plurality of laser radars further includes at least one second target laser radar. The second acquisition module includes: a first splicing submodule configured to splice multiple frames of target point cloud data from each of the at least one second target laser radar with the target point cloud map to obtain second splicing relationship information for each of the at least one second target laser radar; and a fourth acquisition submodule configured to obtain at least one second association relationship information based on the at least one second splicing relationship information.

[0161] In some embodiments, the multiple laser radars include a first target laser radar and at least one second target laser radar. The multi-frame target point cloud data of the first target laser radar includes: at least one frame of first target point cloud data and at least one frame of second target point cloud data. The first acquisition submodule includes: a third acquisition unit, which is used to obtain a spliced point cloud map based on at least one frame of first target point cloud data of the first target laser radar. The first splicing unit is used to splice the at least one frame of second target point cloud data of the first target laser radar and the multi-frame target point cloud data of at least one second target laser radar with the spliced point cloud map to obtain the first motion trajectory information of the first target laser radar and the second motion trajectory information of at least one second target laser radar. The fourth acquisition unit is used to obtain a target point cloud map based on the first motion trajectory information and at least one second motion trajectory information.

[0162] In some embodiments, the second obtaining module includes: a fifth obtaining submodule configured to obtain first association relationship information based on the first motion trajectory information; and a sixth obtaining submodule configured to obtain at least one second association relationship information based on at least one second motion trajectory information.

[0163] In some embodiments, the first obtaining submodule includes: a fifth obtaining unit for obtaining motion trajectory information of the plurality of laser radars based on the multi-frame target point cloud data of the plurality of laser radars; and a sixth obtaining unit for obtaining a plurality of target point cloud maps based on the plurality of motion trajectory information.

[0164] In some embodiments, the second acquisition module includes: a seventh acquisition submodule for, for any laser radar, using the target point cloud map corresponding to the laser radar as the first target point cloud map of the laser radar. An eighth acquisition submodule for obtaining third association relationship information based on the motion trajectory information of the laser radar. For example, the third association relationship information may indicate the association relationship between multiple initial points in multiple frames of target point cloud data of the laser radar and multiple surface features of the first target point cloud map.

[0165] In some embodiments, the second acquisition module further includes: a ninth acquisition submodule, for taking, for any laser radar, target point cloud maps other than the first target point cloud map as second target point cloud maps. A second splicing submodule, for splicing the multi-frame target point cloud data of the laser radar with at least one second target point cloud map, respectively, to obtain at least one splicing relationship information. A tenth acquisition submodule, for obtaining at least one fourth association relationship information based on at least one splicing relationship information. For example, the fourth association relationship information is used to indicate the association relationship between multiple initial points in the multi-frame target point cloud data of the laser radar and multiple surface features of the second target point cloud map.

[0166] In some embodiments, the determination module includes: a second determination submodule, configured to determine initial rotation estimation information and initial displacement estimation information based on the initial trajectory information. An eleventh acquisition submodule, configured to obtain target point cloud map coordinate data of the initial point based on the initial external parameters, the point cloud coordinate data of the initial point, the initial rotation estimation information, and the initial displacement estimation information. For example, the point cloud coordinate data is used to indicate the position of the initial point in the coordinate system of the initial point cloud data, and the target point cloud map coordinate data is used to indicate the position of the initial point in the coordinate system of the target point cloud map. A third determination submodule is configured to determine distance information based on the target point cloud map coordinate data and the association relationship information.

[0167] In some embodiments, the eleventh obtaining submodule includes: a seventh obtaining unit configured to obtain inertial coordinate data of the initial point based on the initial external parameters and the point cloud coordinate data. For example, the inertial coordinate data indicates the position of the initial point in the inertial coordinate system; a first determining unit configured to determine transformation information based on the initial rotation estimation information and the initial displacement estimation information; and a first processing unit configured to process the inertial coordinate data using the transformation information to obtain target point cloud map coordinate data.

[0168] In some embodiments, the adjustment module includes: a fourth determination submodule for determining acceleration information based on the inertial translation data and the initial trajectory information; a fifth determination submodule for determining angular velocity information based on the inertial rotation data and the initial trajectory information; and a second adjustment submodule for adjusting the initial extrinsic parameters and the initial trajectory information based on the acceleration information, the angular velocity information, and the distance information to obtain target trajectory information and target extrinsic parameters.

[0169] In some embodiments, the fourth determination submodule includes: a second determination unit configured to determine initial rotation estimation information and initial acceleration estimation information based on the initial trajectory information; a third determination unit configured to determine initial acceleration measurement information based on the inertial translation data; and a fourth determination unit configured to determine acceleration information based on the initial rotation estimation information, the initial acceleration estimation information, and the initial acceleration measurement information.

[0170] In some embodiments, the fourth determination unit includes: a first processing subunit configured to perform a first processing operation on the initial acceleration estimation information using first preset acceleration information to obtain first acceleration estimation information; a second processing subunit configured to perform a second processing operation on the first acceleration estimation information using initial rotation estimation information to obtain second acceleration estimation information; a first determination subunit configured to determine acceleration difference information between the initial acceleration measurement information and the second acceleration estimation information; and a first adjustment subunit configured to adjust the acceleration difference information using a preset translation deviation to obtain acceleration information.

[0171] In some embodiments, the fifth determination submodule includes: a fifth determination unit configured to determine initial angular velocity estimation information based on the initial trajectory information; a sixth determination unit configured to determine initial angular velocity measurement information based on the inertial rotation data; and a seventh determination unit configured to determine angular velocity information based on the initial angular velocity estimation information and the initial angular velocity measurement information.

[0172] In some embodiments, the seventh determining unit includes: a second determining subunit configured to determine angular velocity difference information between the initial angular velocity measurement information and the initial angular velocity estimation information; and a second adjusting subunit configured to adjust the angular velocity difference information using a preset rotation deviation to obtain angular velocity information.

[0173] In some embodiments, the adjustment module includes: a sixth determination submodule for determining fusion information based on the acceleration information, the angular velocity information, and the distance information; and a third adjustment submodule for adjusting the initial trajectory information and the initial external parameters to converge the fusion information and obtain the target trajectory information and the target external parameters.

[0174] In some embodiments, the second acquisition module includes: a seventh determination submodule for determining at least one target plane in the target point cloud map; an extraction submodule for extracting multiple surface features of the target plane; an eighth determination submodule for determining association subrelationship information between the initial point and the surface features. For example, the association subrelationship indicates the surface feature corresponding to the initial point; and a ninth determination submodule for determining association relationship information based on the multiple association subrelationship information.

[0175] In some embodiments, the eighth determining submodule is further configured to: in response to determining that the data of the surface feature is greater than or equal to a preset surface feature quantity threshold, determine the associated sub-relationship information between the initial point and the surface feature.

[0176] In some embodiments, the initial point cloud data and the initial trajectory information come from a target object traveling along a preset driving route, where the preset driving route includes at least one preset driving sub-route whose arc is greater than or equal to a preset arc threshold.

[0177] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0178] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0179] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0180] like Figure 6 As shown, device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. RAM 603 may also store various programs and data required for the operation of device 600. Computing unit 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to bus 604.

[0181] Various components in device 600 are connected to I / O interface 605, including an input unit 606, such as a keyboard, mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a magnetic disk, optical disk, etc.; and a communication unit 609, such as a network card, modem, wireless communication transceiver, etc. The communication unit 609 allows device 600 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0182] The computing unit 601 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as the map generation method. For example, in some embodiments, the map generation method may be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed onto the device 600 via the ROM 602 and / or the communication unit 609. When the computer program is loaded into the RAM 603 and executed by the computing unit 601, one or more steps of the map generation method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the map generation method via any other suitable means (e.g., via firmware).

[0183] Various embodiments of the systems and techniques described above can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-a-chip systems (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0184] The program code for implementing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device so that when the program code is executed by the processor or controller, the functions / operations specified in the flow chart and / or block diagram are implemented. The program code can be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0185] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on 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), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0186] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) display or an LCD (liquid crystal display)) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0187] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0188] Computer systems may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The client and server relationship arises through computer programs running on the respective computers and having a client-server relationship to each other.

[0189] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not limited herein.

[0190] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A map generation method, comprising: Obtain the target point cloud map based on the multi-frame initial point cloud data of multiple lidars and the initial trajectory information of the inertial measurement unit; Obtaining a plurality of association relationship information based on the target point cloud map and the multi-frame initial point cloud data of each of the plurality of laser radars, wherein the association relationship information is used to indicate an association relationship between a plurality of initial points in the multi-frame initial point cloud data of the laser radar and a plurality of surface features of the target point cloud map; Determining, based on the initial trajectory information, the plurality of association relationship information, and initial external parameters of each of the plurality of laser radars, a plurality of distance information, wherein the distance information includes a plurality of distance sub-information, each of the distance sub-information being used to indicate a distance between the surface feature corresponding to the initial point and the initial point; Adjusting the initial trajectory information and the initial external parameters according to the plurality of distance information to obtain target trajectory information and a plurality of target external parameters; and A target map is generated according to the target trajectory information and a plurality of target external parameters.

2. The method according to claim 1, wherein The step of obtaining a target point cloud map based on the multi-frame initial point cloud data of each of the plurality of laser radars and the initial trajectory information of the inertial measurement unit includes: Determining point cloud compensation information of each of the plurality of laser radars according to the initial trajectory information and the initial external parameters of each of the plurality of laser radars; Using the plurality of point cloud compensation information, respectively adjusting the multi-frame initial point cloud data of the plurality of laser radars to obtain the multi-frame target point cloud data of the plurality of laser radars; Obtaining the target point cloud map according to the multi-frame target point cloud data of each of the plurality of laser radars; The multiple association relationship information obtained based on the target point cloud map and the multiple frames of initial point cloud data of the multiple laser radars includes: A plurality of the association relationship information is obtained based on the target point cloud map and the multi-frame target point cloud data of each of the plurality of the laser radars.

3. The method according to claim 2, wherein: The plurality of laser radars include a first target laser radar, The obtaining of the target point cloud map according to the multi-frame target point cloud data of each of the plurality of laser radars comprises: Obtaining first motion trajectory information of the first target laser radar according to the multi-frame target point cloud data of the first target laser radar; The target point cloud map is obtained according to the first motion trajectory information.

4. The method according to claim 3, wherein: The obtaining of the plurality of association relationship information according to the target point cloud map and the multi-frame target point cloud data of the plurality of laser radars includes: Obtaining first splicing relationship information according to the first motion trajectory information; and First association relationship information is obtained according to the first splicing relationship information.

5. The method according to claim 3, wherein The plurality of laser radars further includes at least one second target laser radar, The obtaining of the plurality of association relationship information according to the target point cloud map and the multi-frame target point cloud data of the plurality of laser radars includes: splicing the multi-frame target point cloud data of at least one second target laser radar with the target point cloud map to obtain second splicing relationship information of at least one second target laser radar; At least one second association relationship information is obtained according to at least one second splicing relationship information.

6. The method according to claim 2, wherein: The plurality of laser radars include a first target laser radar and at least one second target laser radar, wherein the multi-frame target point cloud data of the first target laser radar includes: at least one frame of first target point cloud data and at least one frame of second target point cloud data, The obtaining of the target point cloud map according to the multi-frame target point cloud data of each of the plurality of laser radars comprises: Obtaining a spliced point cloud map based on at least one frame of first target point cloud data of the first target laser radar; Splicing at least one frame of the second target point cloud data of the first target laser radar and multiple frames of target point cloud data of at least one second target laser radar with the stitched point cloud map to obtain first motion trajectory information of the first target laser radar and second motion trajectory information of at least one second target laser radar; The target point cloud map is obtained according to the first motion trajectory information and at least one second motion trajectory information.

7. The method according to claim 6, wherein: The obtaining of the plurality of association relationship information according to the target point cloud map and the multi-frame target point cloud data of the plurality of laser radars includes: Obtaining first association relationship information according to the first motion trajectory information; At least one second association relationship information is obtained according to at least one second motion trajectory information.

8. The method according to claim 2, wherein: The obtaining of the target point cloud map according to the multi-frame target point cloud data of each of the plurality of laser radars comprises: Obtaining motion trajectory information of each of the plurality of laser radars according to the multi-frame target point cloud data of each of the plurality of laser radars; Based on the plurality of motion trajectory information, a plurality of target point cloud maps are obtained.

9. The method according to claim 8, wherein The obtaining of the plurality of association relationship information according to the target point cloud map and the multi-frame target point cloud data of the plurality of laser radars includes: For any of the laser radars, using the target point cloud map corresponding to the laser radar as the first target point cloud map of the laser radar; According to the motion trajectory information of the laser radar, third association relationship information is obtained, wherein the third association relationship information is used to indicate the association relationship between multiple initial points in the multi-frame target point cloud data of the laser radar and multiple surface features of the first target point cloud map.

10. The method according to claim 9, wherein: The obtaining of the plurality of association relationship information according to the target point cloud map and the multi-frame target point cloud data of the plurality of laser radars further includes: For any of the laser radars, use other target point cloud maps except the first target point cloud map as second target point cloud maps; splicing the multi-frame target point cloud data of the laser radar with at least one of the second target point cloud maps to obtain at least one piece of splicing relationship information; At least one fourth association relationship information is obtained based on at least one of the splicing relationship information, wherein the fourth association relationship information is used to indicate the association relationship between multiple initial points in the multi-frame target point cloud data of the laser radar and multiple surface features of the second target point cloud map.

11. The method according to claim 1, wherein Determining a plurality of distance information respectively according to the initial trajectory information, the plurality of association relationship information, and the respective initial external parameters of the plurality of laser radars includes: determining initial rotation estimation information and initial displacement estimation information based on the initial trajectory information; Obtaining target point cloud map coordinate data of the initial point based on the initial external parameters, the point cloud coordinate data of the initial point, the initial rotation estimation information, and the initial displacement estimation information, wherein the point cloud coordinate data is used to indicate the position of the initial point in the coordinate system of the initial point cloud data, and the target point cloud map coordinate data is used to indicate the position of the initial point in the coordinate system of the target point cloud map; and The distance information is determined according to the target point cloud map coordinate data and the association relationship information.

12. The method according to claim 11, wherein The obtaining, according to the initial external parameters, the point cloud coordinate data of the initial point, the initial rotation estimation information, and the initial displacement estimation information, of the target point cloud map coordinate data of the initial point comprises: Obtaining inertial coordinate data of the initial point according to the initial external parameters and the point cloud coordinate data, wherein the inertial coordinate data is used to indicate a position of the initial point in an inertial coordinate system; determining conversion information based on the initial rotation estimation information and the initial displacement estimation information; and The inertial coordinate data is processed using the conversion information to obtain the target point cloud map coordinate data.

13. The method according to claim 1, wherein The adjusting the initial trajectory information and the initial external parameters according to the plurality of distance information to obtain the target trajectory information and the plurality of target external parameters includes: determining acceleration information according to the inertial translation data and the initial trajectory information; determining angular velocity information based on the inertial rotation data and the initial trajectory information; and The initial external parameters and the initial trajectory information are adjusted according to the acceleration information, the angular velocity information, and the distance information to obtain the target trajectory information and the target external parameters.

14. The method according to claim 13, wherein Determining acceleration information according to the inertial translation data and the initial trajectory information includes: determining initial rotation estimation information and initial acceleration estimation information based on the initial trajectory information; determining initial acceleration measurement information based on the inertial translation data; and The acceleration information is determined according to the initial rotation estimation information, the initial acceleration estimation information, and the initial acceleration measurement information.

15. The method according to claim 14, wherein The determining the acceleration information according to the initial rotation estimation information, the initial acceleration estimation information, and the initial acceleration measurement information includes: performing a first operation on the initial acceleration estimation information using the first preset acceleration information to obtain first acceleration estimation information; performing a second operation on the first acceleration estimation information using the initial rotation estimation information to obtain second acceleration estimation information; determining acceleration difference information between the initial acceleration measurement information and the second acceleration estimate information; and The acceleration difference information is adjusted using a preset translation deviation to obtain the acceleration information.

16. The method according to claim 13, wherein: The determining of angular velocity information according to the inertial rotation data and the initial trajectory information includes: determining initial angular velocity estimation information based on the initial trajectory information; determining initial angular velocity measurement information based on the inertial rotation data; and The angular velocity information is determined according to the initial angular velocity estimation information and the initial angular velocity measurement information.

17. The method according to claim 16, wherein: The determining the angular velocity information according to the initial angular velocity estimation information and the initial angular velocity measurement information includes: determining angular velocity difference information between the initial angular velocity measurement information and the initial angular velocity estimation information; The angular velocity difference information is adjusted using a preset rotation deviation to obtain the angular velocity information.

18. The method according to claim 13, wherein The adjusting the initial external parameters and the initial trajectory information according to the acceleration information, the angular velocity information, and the distance information to obtain the target trajectory information and the target external parameters includes: determining fusion information according to the acceleration information, the angular velocity information, and the distance information; and The initial trajectory information and the initial external parameters are adjusted to converge the fusion information, thereby obtaining the target trajectory information and the target external parameters.

19. The method according to claim 1, wherein The multiple association relationship information obtained based on the target point cloud map and the multiple frames of initial point cloud data of the multiple laser radars includes: Determining at least one target plane in the target point cloud map; extracting a plurality of surface features of the target plane; Determining association sub-relationship information between the initial point and the surface feature, wherein the association sub-relationship is used to indicate the surface feature corresponding to the initial point; and The association relationship information is determined according to the plurality of association sub-relationship information.

20. The method according to claim 19, wherein The determining of the associated sub-relationship information between the initial point and the surface feature includes: In response to determining that the data of the surface feature is greater than or equal to a preset surface feature quantity threshold, the associated sub-relationship information between the initial point and the surface feature is determined.

21. The method according to claim 1, wherein The initial point cloud data and the initial trajectory information come from a target object traveling along a preset driving route, where the preset driving route includes at least one preset driving sub-route whose arc is greater than or equal to a preset arc threshold.

22. A map generating device, comprising: The first acquisition module is used to obtain a target point cloud map based on the multi-frame initial point cloud data of the multiple laser radars and the initial trajectory information of the inertial measurement unit; A second obtaining module is configured to obtain a plurality of association relationship information based on the target point cloud map and the multi-frame initial point cloud data of each of the plurality of laser radars, wherein the association relationship information is used to indicate the association relationship between the plurality of initial points in the multi-frame initial point cloud data of the laser radar and the plurality of surface features of the target point cloud map; a determination module, configured to determine a plurality of distance information based on the initial trajectory information, the plurality of association relationship information, and the initial external parameters of each of the plurality of laser radars, wherein the distance information includes a plurality of distance sub-information, and the distance sub-information is used to indicate the distance between the surface feature corresponding to the initial point and the initial point; an adjustment module, configured to adjust the initial trajectory information and the initial external parameters according to the plurality of distance information to obtain target trajectory information and a plurality of target external parameters; and A generation module is used to generate a target map according to the target trajectory information and a plurality of target external parameters.

23. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 21.

24. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 21.

25. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 21.

Images