Point cloud data processing method, device and equipment, and readable storage medium

Abstract

Embodiments of the present application provide a point cloud data processing method, device and equipment, and a readable storage medium. The method comprises: acquiring a plurality of sets of scanning data, the plurality of sets of scanning data being laser point cloud data of an intersection environment collected by a vehicle in a plurality of scanning passes; determining a predicted scanning trajectory and position information of each set of scanning data corresponding to a scanning pass; filtering at least part of the scanning data based on a maximum group algorithm according to the correlation of the plurality of predicted scanning trajectories in the geographical and / or time dimensions, to obtain a plurality of maximum groups, each maximum group including a plurality of sets of target scanning data; tracking a target feature in each set of target scanning data to determine pose information corresponding to the plurality of sets of target scanning data in each scanning data set, the target feature being a high-reflectivity marker in the intersection environment; and splicing the plurality of sets of target scanning data in the plurality of maximum groups according to the pose information corresponding to each set of target scanning data, to obtain a three-dimensional map of the intersection environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a point cloud data processing method, device and equipment, and a readable storage medium. BACKGROUND

[0002] Light Laser Detection and Ranging (LiDAR) has high accuracy and the ability to work under various lighting and weather conditions, and is increasingly applied to geographic information systems and autonomous vehicles.

[0003] When a vehicle equipped with a laser radar passes through an intersection, the laser radar will collect point cloud data of the surrounding environment based on the principle of laser reflection and upload it to a cloud server. The cloud server receives point cloud data collected by different vehicles at different times, aligns the trajectories, detects inter-scan loops, and optimizes the poses to splice the point cloud data into a seamless high-precision three-dimensional map.

[0004] In the above manner, the quality and relevance of the captured data are not considered, and all captured data is aggregated, which increases the complexity of data processing and significantly increases the consumption of computing resources. At the same time, the processing of multi-scan data relies on the LOAM algorithm of surface and line features, which can perform real-time positioning and map construction in a single scan. However, when multiple scans involve different environmental conditions or different time periods, traditional surface and line features may lose stability and significance due to environmental changes. The instability of such features makes it difficult to achieve high-precision alignment of inter-scan data, thereby affecting the quality and reliability of the overall map. SUMMARY

[0005] The present application provides a point cloud data processing method, device and equipment, and a readable storage medium. The aspects related to the embodiments of the present application are introduced below.

[0006] In a first aspect, a point cloud data processing method is provided. The method comprises: obtaining a plurality of sets of scanning data, the plurality of sets of scanning data being laser point cloud data of an intersection environment collected by a vehicle in a plurality of scanning passes; determining a predicted scanning trajectory of each scanning pass corresponding to each set of scanning data and position information of each scanning pass; filtering at least part of the scanning data in the plurality of sets of scanning data based on a maximum clique algorithm according to the correlation of the plurality of predicted scanning trajectories in the geographical and / or time dimensions, to obtain a plurality of maximum cliques, wherein each maximum clique includes a plurality of sets of target scanning data; tracking a target feature in each set of target scanning data to determine pose information corresponding to the plurality of sets of target scanning data in each scanning data set, wherein the target feature is a high-reflectivity marker in the intersection environment; and splicing the plurality of sets of target scanning data in the plurality of maximum cliques according to the pose information corresponding to each set of target scanning data to obtain a three-dimensional map of the intersection environment.

[0007] According to the above technical means, when synthesizing a three-dimensional map using point cloud data of multiple passes, the correlation of the vehicle trajectory is analyzed to filter the plurality of sets of scanning data, thereby reducing redundant data. When performing feature matching, the high-reflectivity marker in the intersection environment is used to achieve stable cross-pass feature matching. This technical solution can reduce the amount of calculation while improving the accuracy of the three-dimensional map.

[0008] In some embodiments, determining the predicted scanning trajectory of each scanning pass corresponding to each set of scanning data and the position information of each scanning pass comprises: processing each set of scanning data using a laser inertial odometry (LIO) algorithm to generate the predicted scanning trajectory of each scanning pass; and using the coordinate information of the starting position of each scanning pass determined by a satellite positioning system as the position information.

[0009] According to the above technical means, the vehicle trajectory and position information of each scanning pass are determined based on the LIO algorithm and the satellite positioning system, which can improve the accuracy of the prediction of the vehicle scanning trajectory and provide a geographical reference for the subsequent processing process.

[0010] In some embodiments, before filtering the plurality of sets of scanning data based on the maximum clique algorithm according to the correlation of the plurality of predicted trajectories in the geographical and / or time dimensions, the method further comprises: determining a first predicted driving trajectory of the vehicle in each scanning pass based on an inertial measurement unit of the vehicle; determining a second predicted driving trajectory of the vehicle in each scanning pass based on a satellite positioning system; and filtering the scanning data with a trajectory correlation less than a first correlation threshold based on the trajectory correlation of the first predicted driving trajectory, the second predicted driving trajectory, and the predicted scanning trajectory.

[0011] According to the above technical means, by performing quality analysis on each single-pass data, and according to factors such as data integrity, accuracy and environmental variables, data with low quality is removed, so that the data for subsequent processing is reliable and useful, and interference in the subsequent map stitching process can be avoided, and the quality and reliability of the map are improved.

[0012] In some embodiments, the screening of at least part of the scanning data in the plurality of groups of scanning data based on the maximum group algorithm comprises: constructing an undirected graph based on the correlation of the plurality of predicted scanning trajectories corresponding to the at least part of the scanning data in the geographical and / or time dimensions; wherein each node in the undirected graph represents a predicted scanning trajectory, and the edge in the undirected graph represents that the correlation of the predicted scanning trajectories corresponding to the two nodes in the geographical and / or time dimensions is greater than a second correlation threshold; determining a plurality of maximum groups in the undirected graph based on the maximum group algorithm, wherein each maximum group includes a plurality of groups of target scanning data.

[0013] According to the above technical means, based on the correlation of the predicted trajectories of the scanning data, the maximum group algorithm in graph theory is used to identify and aggregate single-pass data with high correlation. In this process, each maximum group identified contains trajectories that are closely related to each other, and the maximum group found represents the trajectory set that should be aggregated the most. Such aggregation can significantly reduce the need to process redundant data, thereby improving the processing efficiency of the overall data and system performance.

[0014] In some embodiments, the target feature in each group of target scanning data is tracked, and the pose information corresponding to the plurality of groups of target scanning data in each scanning data set is determined, comprising: tracking a plurality of feature points corresponding to the target feature in each group of target scanning data, and determining the translation and rotation relationship of each group of target scanning data relative to the target feature.

[0015] According to the above technical means, by identifying the target feature in the intersection environment, a detailed local map is constructed for each pass using the extracted features to provide sufficient details to support accurate matching.

[0016] In some embodiments, the target feature in each group of target scanning data is tracked, and the pose information corresponding to the plurality of groups of target scanning data in each scanning data set is determined, further comprising: tracking a plurality of feature points corresponding to the target feature in each group of target scanning data, and determining the positional deviation and angular deviation between different target data according to the shape, size and angle of the target feature in different target scanning data; wherein the target feature is a road sign in the intersection environment.

[0017] According to the above technical means, the high-reflective road sign and the like are used as the target feature, and the strong reflection signal emitted thereby is captured by the laser scanner and can be recognized as a stable feature. By recognizing the same feature points in the data of different scans, it is ensured that the multi-scan data can be correctly aligned in space, overcoming the instability problem of the traditional method based on surface and line features in multi-scan.

[0018] In some embodiments, the splicing of the multiple groups of target scan data in the multiple maximum groups according to the pose information corresponding to each group of target scan data comprises: constructing a pose graph according to the pose information corresponding to each group of target scan data; adjusting the pose of each target scan data based on a pose graph optimization algorithm; and splicing the three-dimensional map of the intersection environment according to the adjusted pose.

[0019] According to the above technical means, the pose of each single scan is accurately adjusted based on the graph optimization technology, and then spliced into a continuous and seamless point cloud map according to the optimized pose, which can ensure the overall consistency and accuracy of the map.

[0020] In a second aspect, a point cloud data processing apparatus is provided, comprising: an acquisition unit configured to acquire multiple groups of scan data, the multiple groups of scan data being laser point cloud data of an intersection environment collected by a vehicle in multiple scan passes; a first determination unit configured to determine a predicted scan trajectory of each scan pass corresponding to each group of scan data and position information of each scan pass; a screening unit configured to screen at least part of the scan data in the multiple groups of scan data based on a maximum group algorithm according to the correlation of the multiple predicted scan trajectories in the geographical and / or time dimensions, to obtain multiple maximum groups; wherein each maximum group comprises multiple groups of target scan data; a second determination unit configured to track a target feature in each group of target scan data and determine pose information corresponding to the multiple groups of target scan data in each scan data set, wherein the target feature is a high-reflectivity marker in the intersection environment; and a splicing unit configured to splice the multiple groups of target scan data in the multiple maximum groups according to the pose information corresponding to each group of target scan data, to obtain a three-dimensional map of the intersection environment.

[0021] In a third aspect, a point cloud data processing device is provided, comprising a memory and a processor, the memory being configured to store a computer program, and the processor being configured to execute the computer program to implement the method described in the first aspect.

[0022] In a fourth aspect, a computer readable storage medium is provided for storing a computer program, the computer program being executed to implement the method described in the first aspect. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 a schematic flowchart of a method for generating an intersection three-dimensional map in the related art;

[0024] Figure 2 A schematic flow chart of a point cloud data processing method provided by an embodiment of the present application is shown in FIG. 1.

[0025] Figure 3 A schematic flow chart of a point cloud data processing method provided by another embodiment of the present application is shown in FIG. 2.

[0026] Figure 4 A schematic flow chart of a method for screening scanning data in a point cloud data processing method provided by an embodiment of the present application is shown in FIG. 3.

[0027] Figure 5 A schematic flow chart of a method for splicing target scanning data in a point cloud data processing method provided by an embodiment of the present application is shown in FIG. 4.

[0028] Figure 6 A schematic structural diagram of a point cloud data processing device provided by an embodiment of the present application is shown in FIG. 5.

[0029] Figure 7 A schematic structural diagram of a point cloud data processing device provided by an embodiment of the present application is shown in FIG. 6. DETAILED DESCRIPTION

[0030] The present application will be described in detail by reference to the attached drawings and preferred embodiments. The skilled person can easily understand other advantages and effects of the present application from the disclosure in the present specification. The present application can also be implemented or applied by different specific embodiments, and the details in the present specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for illustrating the present application, but not for limiting the protection scope of the present application.

[0031] It should be noted that the diagrams provided in the following embodiments only schematically illustrate the basic concept of the present application, and only the components related to the present application are shown in the diagrams, but not the number, shape and size of the components in actual implementation. The shape, number and proportion of the components in actual implementation can be arbitrarily changed, and the layout pattern of the components can be more complex.

[0032] Before introducing the technical solutions of the embodiments of the present application, the point cloud data processing method in the related art and the problems thereof will be described in detail.

[0033] With the continuous development of road traffic, the number of various traffic participants on the road, such as vehicles, pedestrians, bicycles, and the like, is also increasing. Road traffic contains many complex structures (for example, roundabouts, overpasses, and the like), and it is a challenge for an intelligent assisted driving system to accurately obtain the geographic location, road signs, traffic signals, and various traffic objects in road traffic. The safety of the intelligent driving system in road traffic is crucial, and not only does the system itself need to have high stability and robustness to cope with various complex road conditions, but the intelligent driving system also needs to understand and comply with the rules and information (such as traffic lights, traffic signs, and the like) in road traffic to ensure the safe passage of users.

[0034] In the environment of highways, urban expressways, bridges, and the like, the intelligent driving system of a vehicle can use the signs or markings painted on the road surface as a reference and accurately perceive surrounding vehicles and obstacles to intelligently assist driving of the vehicle. However, in urban roads, such environments usually have complex traffic flow and diversified road signs, which puts higher requirements on the perception and decision-making ability of intelligent assisted driving. The vehicle needs to be able to identify and understand various traffic participants such as traffic signals, pedestrians, and vehicles, and make appropriate decisions according to real-time traffic conditions.

[0035] In related technologies, one feasible way to improve the intelligent assisted driving capability of a vehicle is to use a high-precision map to assist driving, which first requires constructing a three-dimensional high-precision map of the road. In the field of map construction, various technologies are widely used, including satellite positioning systems (such as the Global Positioning System, Galileo positioning system, and Beidou positioning system), radar detection, and optical camera systems. These technologies support the generation of various maps from simple road maps to complex city models.

[0036] However, although the above methods have achieved widespread application, they still face many challenges when constructing high-precision maps in complex environments such as intersections. In particular, in multi-lane, traffic-dense urban environments, the above methods often fail to capture complete and accurate road surface information.

[0037] Existing map construction methods mainly rely on global positioning systems and camera systems, which can provide reasonable accuracy under good environmental conditions, but their effectiveness is greatly reduced in poor visibility or unfavorable weather conditions. In addition, the poor coverage of positioning signals in urban canyons and signal reflection problems further limit their application in complex intersection map construction.

[0038] In recent years, Light Laser Detection and Ranging (LiDAR) systems are increasingly applied in geographic information systems and autonomous vehicles due to their high accuracy and ability to work under various lighting and weather conditions.

[0039] Laser scanning is independent of external light sources and generates precise three-dimensional coordinates of objects based on the principle of laser reflection, which is particularly important for map construction at complex intersections. Figure 1 A schematic flowchart of a method for generating a three-dimensional map of an intersection based on laser scanning technology in the related art is shown. The method includes steps S110-S140.

[0040] Step S110, trajectory coarse alignment. The trajectory coarse alignment is performed using GPS data. The trajectory here refers to the movement trajectory of a collection device such as a vehicle when obtaining a scanning data set using a laser radar. Each laser scanning data set is associated with a GPS coordinate, which is used to preliminarily convert the data set into a unified coordinate system. The GPS coordinate is usually the coordinate of the starting position of the trajectory.

[0041] Step S120, inter-pass loop detection. Loop detection is performed by comparing features in each scanning data set with a recorded feature library to determine whether the same location has been repeatedly passed through. The recorded feature library can be fixed objects in the intersection environment, such as curbs, fixed traffic lights, etc.

[0042] Step S130, pose optimization. According to the results of loop detection, a pose graph is constructed. The nodes represent the positions of each scanning pass, and the edges represent the relative poses between passes. The pose graph is optimized to accurately correct the positional and directional deviations between passes.

[0043] Step S140, point cloud stitching. According to the optimized pose data, the point cloud data of each pass is adjusted and stitched according to the pose to construct a continuous and seamless high-precision three-dimensional map.

[0044] The above method can improve the overall accuracy of the map and enhance the continuity and completeness of the map in details. However, it still has the following problems:

[0045] In the above method, the quality and relevance of the captured data are not considered, and all the captured data is aggregated, which increases the complexity of data processing and significantly increases the consumption of computing resources. At the same time, the processing of multi-scan data relies on the LOAM algorithm of surface and line features, which can show good real-time positioning and map construction capability in a single scan. However, when multiple scans involve different environmental conditions or different time periods, traditional surface and line features may lose stability and significance due to environmental changes. The instability of such features makes it difficult to achieve high-precision alignment in inter-scan data fusion, thereby affecting the quality and reliability of the overall map. Therefore, how to improve the data processing efficiency while ensuring the fusion accuracy has become a problem to be solved.

[0046] In view of the above problems, the embodiments of the present application provide a point cloud data processing method, device and equipment, and a readable storage medium. The technical solutions provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0047] Figure 2 is a schematic flowchart of the point cloud data processing method provided by the embodiments of the present application. The method can be executed by a host computer or a server. The server can be a server deployed locally or a server deployed in the cloud.

[0048] Figure 2 The method in S210-S250 includes steps.

[0049] In step S210, a plurality of sets of scanning data are acquired. The plurality of sets of scanning data are laser point cloud data of an intersection environment collected by a vehicle in a plurality of scanning passes.

[0050] The laser point cloud data is a data set of spatial points scanned by a laser radar installed on the vehicle body during the vehicle's travel. The laser radar emits a laser signal, then collects the laser signal reflected by objects in the vehicle's surrounding environment, and combines the speed of light, the time from emission to return of the laser signal, and other information to measure the distance information from the target object. The three-dimensional coordinate information of each spatial point is calculated in combination with other information.

[0051] It can be understood that the plurality of sets of scanning data can be laser point cloud data scanned by the same vehicle in different scanning passes, or can be collected by different vehicles, and the embodiments of the present application do not make specific limitations.

[0052] It can also be understood that the performance parameters of the laser radar configured for different vehicles are usually not the same. For example, the number of scanning lines and / or the scanning frequency are different, which will make the accuracy of different scanning data and the range of intersection environment covered by the scanning data not the same, and the embodiments of the present application also do not make limitations.

[0053] The method of obtaining the plurality of sets of scanning data can be various, for example, the vehicle can automatically send the scanned laser point cloud data to the server in real time through a wireless network during driving; or the vehicle can also store the scanned laser point cloud data in the storage unit of the vehicle during driving, and upload it to the server by the operator of the vehicle.

[0054] In step S220, the predicted scanning trajectory of each scanning pass corresponding to each set of scanning data and the position information of each scanning pass are determined.

[0055] It can be understood that the coordinates of the point cloud data collected by the laser radar of the vehicle are relative to the coordinate system with the laser radar itself as the origin, which can be a Cartesian coordinate system or a polar coordinate system. In order to facilitate the fusion of different data, it is necessary to convert different point cloud data to the same coordinate system. The purpose of this step is to prepare and optimize the data of each pass for more complex processing and analysis.

[0056] The method of determining the predicted scanning trajectory of each scanning pass can be various, for example, the vehicle's inertial measurement unit can be used to predict the scanning trajectory, that is, the speed and acceleration of the vehicle in each direction during driving are used to predict the driving trajectory of the vehicle; or it can also be based on a satellite positioning system, that is, the satellite positioning system is used to determine the latitude and longitude of the vehicle during driving, and then the trajectory of the vehicle is determined. The satellite positioning system mentioned here can include Beidou satellite navigation system BDS, global positioning system GPS, GLONASS satellite navigation system GLONASS, and GALILEO satellite navigation system GALILEO, etc.

[0057] In some embodiments of the present application, the predicted scanning trajectory of each scanning pass can be based on the laser inertial odometry LIO algorithm, which can use the original point cloud data to generate a high-precision predicted scanning trajectory of a single pass.

[0058] In some embodiments, the prediction result of the aforementioned LIO algorithm, the prediction result of the vehicle's inertial measurement unit, and the measurement data of the satellite positioning system can be combined to determine the scanning trajectory, which can improve the accuracy and robustness of position estimation.

[0059] The position information of each scanning pass can be the initial position of the vehicle at each scanning pass, which can provide a geographical reference.

[0060] In some embodiments, the initial position of each scanning pass can be determined by the satellite positioning system, that is, the latitude and longitude of the initial position are determined by positioning the vehicle through the aforementioned satellite positioning systems.

[0061] As a possible implementation manner, in addition to the satellite positioning system, other manners can be used to determine the initial position of the scanning pass. For example, a communication base station of wireless communication can be used to assist positioning of the vehicle, or a real-time kinematic (RTK) positioning method can be used, which has higher dynamic accuracy and can improve the positioning accuracy of the vehicle.

[0062] According to the above technical means, the vehicle trajectory and position information of each scanning pass are determined based on the LIO algorithm and the satellite positioning system, which can improve the accuracy of the vehicle scanning trajectory prediction and provide a geographical reference for the subsequent processing.

[0063] In step S230, at least part of the scanning data in the plurality of groups of scanning data is filtered based on a maximum clique algorithm according to the correlation of the plurality of predicted scanning trajectories in the geographical and / or time dimensions.

[0064] In the embodiment of the present application, the correlation in the geographical dimension can be determined according to the deviation between the predicted trajectories corresponding to different scanning data or the deviation between the initial positions of the corresponding scanning passes. The greater the deviation, the lower the correlation.

[0065] The maximum clique algorithm is a fast search algorithm used to find the clique with the most vertices from an undirected graph composed of multiple points. Each maximum clique in the plurality of maximum cliques based on the maximum clique algorithm includes a plurality of groups of target scanning data. The target scanning data in each maximum clique has a higher correlation in the geographical and / or time dimensions. For example, the plurality of groups of target scanning data can be a plurality of groups of point cloud data scanned along the same or almost the same scanning trajectory. It can be understood that the number of target scanning data in different maximum cliques can be the same or different, and the embodiment of the present application does not limit this.

[0066] The purpose of this step is to optimize and aggregate the data, improve the data quality and processing efficiency, and determine which trajectory set can be optimally aggregated through the maximum clique algorithm. In this process, each maximum clique identified contains trajectories that are closely related to each other, and these trajectories have obvious overlap in geography or time. Such aggregation can significantly reduce the need to process redundant data, thereby improving the overall data processing efficiency and system performance.

[0067] In step S240, a target feature in each group of target data is tracked to determine the pose information corresponding to the plurality of groups of target data in each scanning data set.

[0068] The target features are highly reflective landmarks in the intersection environment. These landmarks can be, for example, road signs. To enhance the nighttime visibility of road signs, their surfaces are typically covered with retroreflective materials such as reflective films. These films reflect vehicle headlights, making them easier for drivers to see. These materials are highly reflective not only of visible light but also of the invisible light emitted by lidar. Consequently, road signs and other landmarks appear as high-density areas in the lidar point cloud. Furthermore, these landmarks are usually fixed relative to the intersection and do not change with environmental factors. Therefore, in the technical solution provided in this application embodiment, the pose information of each set of target data can be determined by tracking landmarks in the point cloud data.

[0069] Based on the above-mentioned technical means, highly reflective road signs and other similar objects are used as target features. The strong reflection signals emitted by these objects are captured by the laser scanner and can be identified as stable features, thus avoiding the instability of traditional methods based on surface and line features.

[0070] Based on the extracted features, a detailed local map is constructed for each pass, providing sufficient detail to support accurate matching. Matching operations are performed between local maps from different passes to identify translation and rotation relationships at the same locations, thereby ensuring that data from multiple passes can be spatially aligned.

[0071] In step S250, based on the pose information corresponding to each set of target scan data, multiple sets of target scan data from multiple largest clusters are stitched together to obtain a three-dimensional map of the intersection environment.

[0072] Based on the aforementioned steps, by extracting features, the pose information of multiple sets of target data is determined. Based on the pose information of each set of target data, a continuous and seamless high-precision 3D map can be constructed from the data of each pass.

[0073] According to the method provided in the embodiments of this application, when synthesizing a 3D map using point cloud data from multiple trips, the correlation of vehicle trajectories is analyzed to filter multiple sets of scan data, reducing redundant data; during feature matching, high-reflectivity markers in the intersection environment are used to achieve stable cross-trip feature matching. This technical solution can improve the accuracy of 3D maps while reducing computational load.

[0074] In some embodiments, such as Figure 3 As shown, in step S230, before filtering at least a portion of the scan data from multiple sets of scan data based on the maximum clique algorithm according to the correlation of multiple predicted scan trajectories in the geographical and / or temporal dimensions, the data processing method provided in this application embodiment further includes steps S311-S313.

[0075] At step S311, a first predicted driving trajectory of the vehicle in each scan pass is determined based on an inertial measurement unit of the vehicle.

[0076] The inertial measurement unit generally includes an accelerometer and a gyroscope. The accelerometer is used to measure the acceleration of the vehicle in different directions in three-dimensional space, and the gyroscope is used to measure the rotation speed of the vehicle, to help determine the angular velocity, and then calculate the attitude and heading of the vehicle. In some implementations, the inertial measurement unit can also include a magnetometer, which is used to detect the earth's magnetic field to assist in correcting the drift of the gyroscope. With the inertial measurement unit, the driving trajectory of the vehicle can be estimated based on the motion parameters of the vehicle itself, i.e., the aforementioned first predicted driving trajectory.

[0077] At step S312, a second predicted driving trajectory of the vehicle in each scan pass is determined based on a satellite positioning system.

[0078] The satellite positioning system can be any one of the various satellite positioning systems described above. During the movement of the vehicle, the latitude and longitude information of the vehicle is collected in real time, i.e., the second predicted driving trajectory is obtained.

[0079] At step S313, according to the correlation of the first predicted driving trajectory, the second predicted driving trajectory, and the predicted scan trajectory, the multiple groups of scan data are screened, and the scan data whose correlation is less than a first correlation threshold are removed.

[0080] The correlation of the above multiple trajectories is used to represent the degree of deviation between different trajectories. For example, the correlation can be evaluated according to the size of the deviation of the initial position, the intermediate position, or the end position of each trajectory; or, the trajectories can be sampled according to the driving time of the vehicle, and multiple positions on different trajectories at the same time are selected, and the correlation is evaluated according to the deviation between the positions corresponding to the three trajectories at the same time.

[0081] When the correlation between the above multiple trajectories is less than the first correlation threshold, it can be considered that there is a large deviation in the group of data, for example, there is an error in the data collected by the inertial measurement unit, or the predicted scan trajectory deviates from the actual situation due to too many interference factors in the intersection environment, etc. In this case, the group of data should be removed to avoid interference in the subsequent splicing process and affect the quality and reliability of the overall map.

[0082] It should be noted that the evaluation index of the correlation of the above trajectories and the selection criteria of the first correlation threshold are only examples, and in actual application, appropriate evaluation index and specific value of the first correlation threshold can be selected according to the requirements of data processing speed and accuracy, and the embodiments of the present application do not make specific limitations.

[0083] According to the above technical means, by performing quality analysis on each single-pass data, and according to factors such as data integrity, accuracy and environmental variables, data with low quality is removed, so that the data for subsequent processing is reliable and useful, interference in the subsequent map stitching process can be avoided, and the quality and reliability of the map are improved.

[0084] In some embodiments, as shown in FIG. 2B, the step S230 in the foregoing text further includes steps S231-S232. Figure 4

[0085] In step S231, a directed graph is constructed based on the correlation of the plurality of predicted scan trajectories corresponding to the at least part of the scan data in the geographical and / or time dimensions.

[0086] Each node in the directed graph represents a predicted scan trajectory, and an edge in the directed graph represents that the correlation of the predicted scan trajectories corresponding to the two nodes in the geographical and / or time dimensions is greater than the second correlation threshold.

[0087] The correlation of the two predicted scan trajectories in the geographical dimension can be understood as the approximation degree of the two predicted scan trajectories. For example, the closer the starting positions of the two predicted scan trajectories, the higher the correlation, and the closer the intermediate points of the two predicted scan trajectories, the higher the correlation.

[0088] In step S232, a plurality of maximum cliques are determined in the directed graph based on the maximum clique algorithm. Each maximum clique includes a plurality of target scan data groups.

[0089] The maximum clique algorithm is a fast search algorithm for finding the maximum clique with the largest number of vertices from a directed graph composed of a plurality of points. Based on the correlation of the predicted trajectories of the scan data, the maximum clique algorithm in graph theory is used to identify and aggregate single-pass data with high correlation. In this process, each maximum clique identified contains trajectories that are closely related to each other, and the maximum clique found represents the trajectory set that should be aggregated the most. Such aggregation can significantly reduce the need to process redundant data, thereby improving the overall data processing efficiency and system performance.

[0090] In some embodiments, the step S240 in the foregoing further includes:

[0091] The plurality of target points corresponding to the target features in each target data group are tracked, and the translation and rotation relationship of each target scan data relative to the target features is determined.

[0092] ​According to the above technical means, by identifying the target features of the intersection environment, the extracted features are used to construct a detailed local map for each scan to provide sufficient details to support accurate matching.

[0093] In some embodiments, the step S240 of tracking the target features in each group of target data to determine the pose information corresponding to each group of target data in the scan data set further comprises: tracking a plurality of feature points corresponding to the target features in each group of target scan data, and determining the positional deviation and angular deviation between different target data according to the differences in shape, size and angle of the target features in different target scan data.

[0094] According to the foregoing description, the target features are features with high reflectivity such as road signs in the intersection environment, which appear as a dense area in the point cloud. By tracking this part of the point cloud, the positional and angular deviations between different point cloud data are determined according to the differences in shape, size and angle of the same target features in different point cloud data, so as to align the data of different scans in space.

[0095] According to the above technical means, the target features with high reflectivity are used as a more stable feature source. By identifying the same feature points in different scans, it is ensured that the multi-scan data can be correctly aligned in space, overcoming the instability problem of traditional methods based on surface and line features in multi-scan.

[0096] In some embodiments, as shown in Figure 5 The step S250 of splicing a plurality of groups of target scan data in the maximum cluster according to the pose information corresponding to each group of target scan data further comprises steps S251-S253.

[0097] In step S251, a pose graph is constructed according to the pose information corresponding to each group of target scan data.

[0098] It should be noted that the pose graph mentioned here is a pose graph constructed according to the spatial relationship within and between scans, including all single-scan poses and relative poses between scans.

[0099] In step S252, the pose of each target scan data is adjusted based on a pose graph optimization algorithm.

[0100] In step S253, the three-dimensional map of the intersection environment is spliced according to the adjusted pose.

[0101] According to the above technical means, the poses of each single scan are accurately adjusted based on the graph optimization technology, and then a continuous and seamless point cloud map is spliced according to the optimized pose, which can ensure the overall consistency and accuracy of the map.

[0102] The above describes the method embodiments of the present application in combination with Figures 1-5 The device embodiments of the present application are described in detail below in combination with Figure 6 and Figure 7 It should be understood that the description of the device embodiments corresponds to the method embodiments, and therefore, the parts not described in detail can be referred to the method embodiments described above.

[0103] Figure 6 is a schematic structural diagram of the point cloud data processing device provided by the embodiments of the present application, Figure 6 The point cloud data processing device 600 in the above

[0104] The acquisition unit 610 is configured to acquire a plurality of groups of scanning data, the plurality of groups of scanning data being laser point cloud data of an intersection environment collected by a vehicle in a plurality of scanning laps.

[0105] The first determination unit 620 is configured to determine a predicted scanning track of each scanning lap corresponding to each group of scanning data and position information of each scanning lap.

[0106] The screening unit 630 is configured to screen at least part of the scanning data in the plurality of groups of scanning data based on a maximum clique algorithm according to the correlation of the plurality of predicted scanning tracks in the geographical and / or time dimensions, to obtain a plurality of maximum cliques; wherein each maximum clique includes a plurality of groups of target scanning data.

[0107] The second determination unit 640 is configured to track a target feature in each group of target scanning data and determine pose information corresponding to the plurality of groups of target scanning data in each scanning data set, wherein the target feature is a high-reflectivity marker in the intersection environment.

[0108] The splicing unit 650 is configured to splice the plurality of groups of target scanning data in the plurality of maximum cliques according to the pose information corresponding to each group of target scanning data, to obtain a three-dimensional map of the intersection environment.

[0109] In some embodiments, the first determination unit 620 is further configured to process each group of scanning data by using a laser inertial odometer (LIO) algorithm to generate the predicted scanning track of each scanning lap, and use the coordinate information of the starting position of each scanning lap determined by the satellite positioning system as the position information.

[0110] In some embodiments, the first determination unit 620 is further configured to determine a first predicted driving track of the vehicle in each scanning lap based on an inertial measurement unit of the vehicle, and determine a second predicted driving track of the vehicle in each scanning lap based on a satellite positioning system.

[0111] The screening unit 630 is further configured to screen the multiple groups of scan data according to the trajectory correlation of the first predicted driving trajectory, the second predicted driving trajectory and the predicted scan trajectory, and eliminate scan data with trajectory correlation less than a first correlation threshold.

[0112] In some embodiments, the screening unit 630 is further configured to construct an undirected graph based on the correlation of the multiple predicted scan trajectories corresponding to the at least partial scan data in the geographical and / or time dimensions; each node in the undirected graph represents a predicted scan trajectory, and an edge in the undirected graph represents that the correlation of the predicted scan trajectories corresponding to the two nodes in the geographical and / or time dimensions is greater than a second correlation threshold; and determine multiple maximum cliques in the undirected graph based on a maximum clique algorithm, wherein each maximum clique includes multiple groups of target scan data.

[0113] In some embodiments, the second determining unit 640 is further configured to track multiple feature points corresponding to the target feature in each group of target scan data, and determine the translation and rotation relationship of each group of target scan data relative to the target feature.

[0114] In some embodiments, the target feature is a road sign in the intersection environment.

[0115] In some embodiments, the splicing unit 650 is further configured to construct a pose graph according to the pose information corresponding to each group of target scan data; adjust the pose of each target scan data based on a pose graph optimization algorithm; and splice the three-dimensional map of the intersection environment according to the adjusted pose.

[0116] Figure 7 FIG. 1 is a schematic structural diagram of a point cloud data processing device provided by an embodiment of the present application. Figure 7 The dashed line in FIG. 7 indicates that the unit or module is optional. The point cloud data processing device 700 can be used to implement the methods described in the above method embodiments. The point cloud data processing device 700 may, for example, be a server; or the point cloud data processing device 700 may also be a chip or integrated circuit capable of implementing the methods of the above embodiments.

[0117] The point cloud data processing device 700 can include one or more processors 710. The processor 710 can support the point cloud data processing device 700 to implement the method described in the foregoing method embodiments. The processor 710 can be a general processor or a special-purpose processor. For example, the processor can be a central processing unit (CPU). Alternatively, the processor 710 can also be other general processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The general processor can be a microprocessor or the processor can also be any conventional processor.

[0118] The point cloud data processing device 700 can further include one or more memories 720. The memory 720 stores a computer program, which can be executed by the processor 710, so that the processor 710 performs the method described in the foregoing method embodiments. The memory 720 can be independent of the processor 710 or integrated in the processor 710.

[0119] The point cloud data processing device 700 can further include a transceiver 730. The processor 710 can communicate with other devices or chips through the transceiver 730. For example, the processor 710 can perform data transceiving with other devices or chips through the transceiver 730.

[0120] The embodiments of the present application further provide a chip including a processor, which can be used to call and run a computer program from a memory, so that a device having the chip installed performs the method described in the foregoing method embodiments. It can be understood that the processor can be any type of processor mentioned above. It can be understood that the memory can be independent of the chip or integrated in the chip.

[0121] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. When the computer program is executed, the method in the embodiments of the present application can be implemented.

[0122] The embodiments of the present application further provide a computer program product. The computer program product includes a program, which causes a computer to execute the method in the embodiments of the present application.

[0123] The embodiments of the present application further provide a computer program. The computer program causes a computer to execute the method in the embodiments of the present application.

[0124] It should be understood that, in the embodiments of the present application, "B corresponding to A" means that B is associated with A, and B can be determined according to A. It should also be understood that the determination of B according to A does not mean that B is determined only according to A, but B can also be determined according to A and / or other information.

[0125] It should be understood that the term "and / or" herein is only a description of the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the character " / " herein generally represents that the front and rear associated objects are in an "or" relationship.

[0126] It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of the processes should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0127] In several embodiments provided by the present application, it should be understood that the disclosed system and device can be implemented by other means. For example, the device embodiments described above are only schematic, for example, the division of units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed mutual objects can be indirect coupling or communication connection through some interface, device or unit, which can be electrical, mechanical or other forms.

[0128] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on a plurality of network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the present embodiment.

[0129] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit.

[0130] In the above embodiments, all or part of the embodiments can be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of the present application are generated. The computer can be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions can be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) mode. The computer readable storage medium can be any available medium readable by a computer or a data storage device such as a server, data center, etc. integrated with one or more available media. The available media can be magnetic media (such as floppy disk, hard disk, magnetic tape), optical media (such as digital video disc (DVD)) or semiconductor media (such as solid state disk (SSD)) and the like.

[0131] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for processing point cloud data, characterized in that, The method comprises: acquiring a plurality of groups of scanning data, the plurality of groups of scanning data being laser point cloud data of an intersection environment collected by a vehicle in a plurality of scanning passes; determining a predicted scanning trajectory of each scanning pass corresponding to each group of scanning data and position information of each scanning pass; screening at least part of the plurality of groups of scanning data based on a maximum group algorithm according to correlation of the plurality of predicted scanning trajectories in a geographical and / or time dimension, to obtain a plurality of maximum groups; wherein each maximum group comprises a plurality of groups of target scanning data; tracking a target feature in each group of target scanning data to determine pose information of the plurality of groups of target scanning data in each maximum group; wherein the target feature is a high-reflectivity marker in the intersection environment; splicing the plurality of groups of target scanning data in the plurality of maximum groups according to the pose information of each group of target scanning data to obtain a three-dimensional map of the intersection environment; wherein the screening of at least part of the plurality of groups of scanning data based on the maximum group algorithm comprises: constructing an undirected graph based on correlation of a plurality of predicted scanning trajectories corresponding to the at least part of the plurality of groups of scanning data in a geographical and / or time dimension; wherein each node in the undirected graph represents a predicted scanning trajectory, and an edge in the undirected graph represents that correlation of the predicted scanning trajectories corresponding to two nodes in a geographical and / or time dimension is greater than a second correlation threshold; determining a plurality of maximum groups in the undirected graph based on a maximum group algorithm.

2. The method of claim 1, wherein, The determination of a predicted scanning trajectory of each scanning pass corresponding to each group of scanning data and position information of each scanning pass comprises: processing each group of scanning data by a laser inertial mileage calculation method to generate a predicted scanning trajectory of each scanning pass; and using coordinate information of a starting position of each scanning pass determined by a satellite positioning system as the position information.

3. The method of claim 2, wherein, Before the screening of the plurality of groups of scanning data based on the maximum group algorithm according to correlation of the plurality of predicted trajectories in a geographical and / or time dimension, the method further comprises: determining a first predicted driving trajectory of the vehicle in each scanning pass based on an inertial measurement unit of the vehicle; determining a second predicted driving trajectory of the vehicle in each scanning pass based on a satellite positioning system; screening the plurality of groups of scanning data according to trajectory correlation of the first predicted driving trajectory, the second predicted driving trajectory and the predicted scanning trajectory, and eliminating scanning data with trajectory correlation less than a first correlation threshold.

4. The method of claim 1, wherein, The tracking of a target feature in each group of target scanning data to determine pose information of the plurality of groups of target scanning data in each maximum group comprises: tracking a plurality of feature points corresponding to the target feature in each group of target scanning data to determine a translation and rotation relationship of each group of target scanning data relative to the target feature.

5. The method of claim 4, wherein, The tracking of a target feature in each group of target scanning data to determine pose information of the plurality of groups of target scanning data in each maximum group further comprises: tracking a plurality of feature points corresponding to a target feature in each group of the target scan data, and determining position deviation and angle deviation between different groups of the target scan data according to different shapes, sizes and angles of the target feature in different groups of the target scan data; wherein the target feature is a road sign in the intersection environment.

6. The method according to any one of claims 1-5, characterized in that, The splicing, according to the pose information corresponding to each group of the target scan data, of a plurality of groups of the target scan data in the plurality of maximal cliques includes: constructing a pose graph according to the pose information corresponding to each group of the target scan data; adjusting the pose of each group of the target scan data based on a pose graph optimization algorithm; splicing a three-dimensional map of the intersection environment according to the adjusted pose.

7. A point cloud data processing apparatus, characterized by comprising: The device includes: an acquisition unit configured to acquire a plurality of groups of scan data, the plurality of groups of scan data being laser point cloud data of an intersection environment collected by a vehicle in a plurality of scan passes; a first determination unit configured to determine a predicted scan trajectory of a scan pass corresponding to each group of the scan data and position information of each scan pass; a screening unit configured to screen at least part of the plurality of groups of scan data based on a maximal clique algorithm according to correlation of a plurality of predicted scan trajectories in a geographical and / or time dimension, to obtain a plurality of maximal cliques, wherein each maximal clique includes a plurality of groups of target scan data; a second determination unit configured to track a target feature in each group of the target scan data and determine pose information corresponding to a plurality of groups of the target scan data in each maximal clique, wherein the target feature is a high-reflectivity marker in the intersection environment; a splicing unit configured to splice a plurality of groups of the target scan data in the plurality of maximal cliques according to the pose information corresponding to each group of the target scan data, to obtain a three-dimensional map of the intersection environment; wherein the screening of at least part of the plurality of groups of scan data based on the maximal clique algorithm includes: constructing an undirected graph based on correlation of a plurality of predicted scan trajectories corresponding to the at least part of the scan data in a geographical and / or time dimension; wherein each node in the undirected graph represents a predicted scan trajectory, and an edge in the undirected graph represents that correlation of predicted scan trajectories corresponding to two nodes in a geographical and / or time dimension is greater than a second correlation threshold; determining a plurality of maximal cliques in the undirected graph based on the maximal clique algorithm.

8. A point cloud data processing device, characterized by, The point cloud data processing device includes a memory and a processor, the memory is configured to store a computer program, and the processor is configured to execute the computer program to implement the method of any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The storage medium is configured to store a computer program, and the computer program is executed to implement the method of any one of claims 1-6.

Images