Multi-track mapping method and device, vehicle and storage medium

Abstract

The application discloses a multi-track mapping method and device, a vehicle and a storage medium. The method comprises: acquiring key frame data corresponding to a plurality of historical driving tracks of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving track and at least one map feature corresponding to the vehicle pose information; performing alignment processing on the key frame data corresponding to different historical driving tracks to obtain a plurality of key frame pairing results; performing feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine the feature correspondence relationship between the plurality of key frame pairing results; and performing fusion processing based on the feature association result to generate target map data. The application solves the technical problems of low mapping efficiency and poor accuracy in offline mapping in the related art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving, in particular to a multi-trajectory mapping method and device, a vehicle and a storage medium. BACKGROUND

[0002] Offline mapping can improve positioning accuracy, speed up path planning, improve safety and reduce cost, and provide strong support for the development and application of intelligent driving technology. The traditional single-memory route offline mapping method has problems such as missing detection and false detection of mapping elements. Multiple repeated collection of information of the same route for mapping can alleviate the problems of missing detection and false detection. However, when using multiple collected information for offline mapping, there are differences in the time, road and observed area of each collection of information, and there are also problems such as unstable perception effect and positioning signal, which makes it difficult to accurately calculate the correlation between multiple collected information, further affecting the mapping efficiency and accuracy in the offline mapping process.

[0003] At present, there is no effective solution to the above problems. SUMMARY

[0004] Embodiments of the present application provide a multi-trajectory mapping method, device, vehicle and storage medium to at least solve the technical problems of low mapping efficiency and poor accuracy in offline mapping in related technologies.

[0005] According to an embodiment of the present application, a multi-trajectory mapping method is provided, comprising: obtaining key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data includes vehicle pose information in the historical driving trajectory and at least one map feature corresponding to the vehicle pose information; performing alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results; performing feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine the feature correspondence relationship between the plurality of key frame pairing results; and performing fusion processing based on the feature association result to generate target map data.

[0006] Optionally, the plurality of historical driving trajectories at least include a first driving trajectory and a second driving trajectory, and the key frame data at least include a plurality of first key frames corresponding to the first driving trajectory and a plurality of second key frames corresponding to the second driving trajectory. The aligning the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results includes: performing position association processing on the plurality of first key frames and the plurality of second key frames to obtain a position association result, wherein the position association result is used to represent an initial pairing result of the first key frames and the second key frames; performing spatio-temporal association processing based on the position association result to obtain a spatio-temporal association result, wherein the spatio-temporal association result is used to represent a key frame chain satisfying a preset spatio-temporal continuity condition; and determining the plurality of key frame pairing results according to the spatio-temporal association result.

[0007] Optionally, the performing feature association processing on the plurality of key frame pairing results to obtain a feature association result includes: obtaining vehicle pose information and lane line information based on the plurality of key frame pairing results; performing first association processing on the vehicle pose information and the lane line information to obtain a lane line association result, wherein the lane line association result is used to represent a lane line correspondence relationship in the key frame pairing result; and performing second association processing on the lane line association result to obtain the feature association result.

[0008] Optionally, the performing first association processing on the vehicle pose information and the lane line information to obtain a lane line association result includes: performing feature matching on the lane line information to obtain a first matching result; performing detection processing on the first matching result based on the vehicle pose information to obtain a target detection result, wherein the target detection result is used to determine whether a lane line association distance between adjacent key frames satisfies a preset distance condition; and determining the lane line association result based on the target detection result.

[0009] Optionally, the performing second association processing on the lane line association result to obtain a feature association result includes: performing adjustment processing on the vehicle pose information based on the lane line association result to obtain a pose adjustment result; and performing second association processing on at least one map feature in the key frame pairing result based on the pose adjustment result and the at least one map feature to obtain the feature association result.

[0010] Optionally, the performing second association processing on at least one map feature in the key frame pairing result based on the pose adjustment result and the at least one map feature to obtain a feature association result includes: performing feature matching on the at least one map feature in the key frame pairing result based on the pose adjustment result to obtain a second matching result; performing pose updating processing based on the second matching result to obtain a pose updating result; and in response to the pose updating result satisfying a preset convergence condition, determining the feature association result based on the pose updating result.

[0011] Optionally, the generating the target map data based on the feature association result comprises: performing screening processing on the feature association result by using a preset screening condition to obtain a screening result, wherein the preset screening condition is used to filter an error pairing relationship in the feature association result; and performing fusion processing based on the screening result to generate the target map data.

[0012] According to an embodiment of the present application, a multi-track mapping device is also provided, comprising: an acquisition module configured to acquire key frame data corresponding to a plurality of historical driving tracks of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving track and at least one map feature corresponding to the vehicle pose information; a processing module configured to perform alignment processing on the key frame data corresponding to different historical driving tracks to obtain a plurality of key frame pairing results; an association module configured to perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results; and a generation module configured to perform fusion processing based on the feature association result to generate target map data.

[0013] Optionally, the processing module is further configured to: perform position association processing on the plurality of first key frames and the plurality of second key frames to obtain a position association result, wherein the position association result is used to represent an initial pairing result of the first key frames and the second key frames; perform spatio-temporal association processing based on the position association result to obtain a spatio-temporal association result, wherein the spatio-temporal association result is used to represent a key frame chain satisfying a preset spatio-temporal continuity condition; and determine the plurality of key frame pairing results according to the spatio-temporal association result.

[0014] Optionally, the association module is further configured to: acquire vehicle pose information and lane line information based on the plurality of key frame pairing results; perform first association processing by using the vehicle pose information and the lane line information to obtain a lane line association result, wherein the lane line association result is used to represent a lane line correspondence relationship in the key frame pairing result; and perform second association processing by using the lane line association result to obtain the feature association result.

[0015] Optionally, the association module is further configured to: perform feature matching by using the lane line information to obtain a first matching result; perform detection processing on the first matching result based on the vehicle pose information to obtain a target detection result, wherein the target detection result is used to determine whether a lane line association distance between adjacent key frames satisfies a preset distance condition; and determine the lane line association result based on the target detection result.

[0016] Optionally, the association module is further configured to: perform adjustment processing on the vehicle pose information based on the lane line association result to obtain a pose adjustment result; perform the second association processing based on the pose adjustment result and at least one map feature in the key frame pairing result to obtain the feature association result.

[0017] Optionally, the association module is further configured to perform feature matching based on at least one map feature in the pose adjustment result and the key frame pairing result to obtain a second matching result; perform pose update processing based on the second matching result to obtain a pose update result; and determine the feature association result based on the pose update result in response to the pose update result satisfying a preset convergence condition.

[0018] Optionally, the generation module is further configured to perform screening processing on the feature association result by using a preset screening condition to obtain a screening result, wherein the preset screening condition is used to filter an incorrect pairing relationship in the feature association result; and perform fusion processing based on the screening result to generate the target map data.

[0019] According to an embodiment of the present application, a vehicle is provided, comprising: a memory storing an executable program; and a processor configured to execute the program, wherein the program performs the multi-track mapping method of any one of the embodiments of the present application when executed.

[0020] According to an embodiment of the present application, a computer readable storage medium is provided, comprising a stored executable program, wherein the executable program controls a device where the storage medium is located to perform the multi-track mapping method of any one of the embodiments of the present application when executed.

[0021] According to an embodiment of the present application, a computer program product is also provided, comprising a computer program, which, when executed by a processor, implements the multi-track mapping method of any one of the embodiments of the present application.

[0022] According to an embodiment of the present application, a computer program product is also provided, comprising a non-volatile computer readable storage medium, which stores a computer program, and the computer program, when executed by a processor, implements the multi-track mapping method of any one of the embodiments of the present application.

[0023] According to an embodiment of the present application, a computer program is also provided, which, when executed by a processor, implements the multi-track mapping method of any one of the embodiments of the present application.

[0024] In an embodiment of the present application, by aligning keyframe data corresponding to different historical driving trajectories and performing feature association and fusion processing, the accuracy and completeness of map data can be improved, reducing the inaccurate or missing map information caused by the inadequacy of a single data source. By processing keyframe data and performing feature association on multiple historical driving trajectories of a target vehicle along a preset driving route, the positioning accuracy of the target vehicle during driving can be improved, positioning errors can be reduced, and navigation accuracy can be improved. Fusion processing based on the feature association results to generate target map data can effectively reduce the cost and time of map updates, reduce the need for manual intervention, increase the speed and efficiency of map data updates, and provide more accurate and complete map information for the autonomous driving system, thereby improving the performance and stability of the autonomous driving system and enhancing the driving safety and efficiency of the vehicle. As a result, the multi-trajectory mapping method provided in the embodiment of the present application achieves the goal of efficiently and accurately generating target map data, thereby achieving the technical effect of improving the mapping efficiency and accuracy during the offline mapping process, thereby solving the technical problems of low mapping efficiency and poor accuracy in offline mapping in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 is a flowchart of a multi-trajectory mapping method according to one embodiment of the present application;

[0027] Figure 2 is a flowchart of another multi-trajectory mapping method according to one embodiment of the present application;

[0028] Figure 3 is a schematic diagram of a multi-trajectory mapping method according to one embodiment of the present application;

[0029] Figure 4 4 is a structural block diagram of a multi-trajectory mapping device according to one embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0031] It should be noted that the terms "first", "second", and the like in the description and in the claims of the present application and above-described accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular chronological or sequential order. It should be understood that the data thus used can be interchanged, where appropriate, so that the embodiments of the present application described herein can be carried out in sequences other than those illustrated or described herein. Moreover, the terms "comprise" and "have", and any variations thereof, are intended to cover non-exclusive inclusion, for example, processes, methods, systems, products, or devices that comprise a list of steps or units are not necessarily limited to those steps or units that are clearly listed, but can include other steps or units that are not clearly listed or inherent to such processes, methods, products, or devices.

[0032] Embodiments of the present application provide a multi-track mapping method. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0033] Offline mapping is a method of modeling and creating a map of an environment using pre-collected sensor data and map information without real-time positioning and sensor data. Offline mapping can improve the accuracy and stability of positioning using historical data and map information without real-time data, while reducing the use of real-time sensor data, saving energy and computing resources. Through offline mapping, the environment can be modeled and analyzed in advance to identify potential hazards and obstacles, improving safety, and in large-scale environments, the need for real-time data transmission and processing can be reduced, reducing cost and complexity.

[0034] Traditional single-memory route offline mapping methods have problems such as missing detection and false detection of mapping elements. Multiple repeated collection of information for the same route for mapping can alleviate the problem of missing detection and false detection. However, when using multiple collected information for offline mapping, there are differences in the time, road and observed area of each collection of information, as well as problems such as unstable perception effect and positioning signal, making it difficult to accurately calculate the correlation between multiple collected information, further affecting the mapping efficiency and accuracy in the offline mapping process.

[0035] The embodiment of the application provides a multi-track mapping method, which is mainly applied to the offline mapping process in the automatic driving technical field. Through efficient and accurate generation of map data, important basic data support is provided for positioning, navigation and other applications of automatic driving vehicles. Specifically, it can be applied to the creation and update of maps of commuting routes, tourist routes or other regular driving routes to adapt to the changing road environment and traffic conditions.

[0036] The multi-track mapping method in the embodiment of the application can effectively solve the problems of missing detection and false detection of mapping elements in the traditional offline mapping method, improve the accuracy of map data and the reliability of application. By analyzing the historical driving track of the target vehicle on the preset driving route, extracting key frame data, and performing alignment processing and feature association processing, high-precision map data fused with multi-source information is finally generated. The high-precision map data not only improves the safety and efficiency of the automatic driving vehicle, but also provides important road information resources for the traffic management department.

[0037] The multi-track mapping method provided by the embodiment of the application can be mainly applied to automatic driving, intelligent transportation system, vehicle navigation and traffic monitoring and other related scenes, and the core advantage is to enhance the technical application and service quality in the related technical scene by improving the generation efficiency and accuracy of map data.

[0038] Specifically, in the offline map generation scene of the automatic driving vehicle, the automatic driving vehicle needs accurate map data to assist navigation and decision making, and more accurate offline maps can be generated through historical driving tracks. In the infrastructure planning scene of the intelligent transportation system, through the accurate map data constructed offline, users can better understand and plan the urban transportation infrastructure. In the update and maintenance scene of the vehicle navigation system, the navigation system of the vehicle can provide more accurate navigation services through efficient update of map data, and more accurate map data can help the assisted driving system better understand the road conditions, help the vehicle to accurately position and efficiently plan the path, reduce the driving time and cost, and improve the driving safety.

[0039] Figure 1 is a flowchart of a multi-track mapping method according to an embodiment of the application, as shown in Figure 1 The method comprises the following steps:

[0040] Step S11, acquiring key frame data corresponding to a plurality of historical driving tracks of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving track and at least one map feature corresponding to the vehicle pose information;

[0041] Step S12, performing alignment processing on the key frame data corresponding to different historical driving tracks to obtain a plurality of key frame pairing results;

[0042] In step S13, feature association processing is performed on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine the feature correspondence relationship between the plurality of key frame pairing results.

[0043] In step S14, fusion processing is performed based on the feature association result to generate target map data.

[0044] The preset driving route can be a commuting route, a travel route, or other regular driving route, and the plurality of historical driving trajectories are trajectory data of the target vehicle driving on the preset driving route multiple times, which can be obtained by a vehicle-mounted sensor, a Global Positioning System (GPS) trajectory recorder, a traffic monitoring camera, or the like. For each historical driving trajectory, a trajectory analysis algorithm can be used to extract key frame data, such as filtering, interpolation, or the like, to process the trajectory data and extract the current vehicle pose at the key time and all observable map features at the current pose.

[0045] The key frame data can include, but is not limited to, a timestamp, vehicle pose information, at least one map feature, route information, environmental information, and the like. The timestamp is used to record the collection time of the vehicle pose information and the map feature, the vehicle pose information includes the position and attitude of the target vehicle, and is usually GPS data and Inertial Measurement Unit (IMU) data, which can be obtained by sensors on the target vehicle. The at least one map feature can be a landmark such as a road sign, an intersection, a building, or a traffic facility such as a lane line, a traffic signal, or the like. The route information is used to record the specific route of the target vehicle, including the starting point, the ending point, the passing place, and the like. The environmental information specifically includes the surrounding traffic situation, the road condition, and the like, which can help analyze the reasons and decision-making process of the historical driving trajectory.

[0046] After obtaining the key frame data, the key frame data corresponding to different historical driving trajectories is aligned to obtain a plurality of key frame pairing results. For example, the similarity between the key frames on different historical driving trajectories can be calculated. First, the similarity between each pair of key frames needs to be calculated, which can be calculated using a feature matching algorithm or a deep learning model. Then, the key frames are sorted according to the similarity results, and the key frames with the highest similarity are selected for pairing. Finally, a plurality of key frame pairing results can be obtained, thereby realizing the alignment of different historical driving trajectories. For example, the key frame data corresponding to historical driving trajectory A and the key frame data corresponding to historical driving trajectory B are aligned, a key frame B2 with the strongest correlation can be found for key frame A1 on historical driving trajectory A on historical driving trajectory B, thereby forming a key frame pair.

[0047] Further, the plurality of key frame pairing results are subjected to feature association processing to obtain a feature association result. Specifically, the feature association is performed between each key frame pairing result, thereby constructing the constraints between the key frames. The feature association processing can associate and align the map features between the key frame pairs, thereby determining the feature correspondence relationship between the plurality of key frame pairing results. For example, the feature association processing includes two steps of coarse association and fine association, wherein the coarse association processing only uses lane line information in the map features for association, and the fine association processing uses all map features for association.

[0048] After obtaining the feature association result, the feature association result is subjected to fusion processing to generate target map data. The target map data is an offline mapping result, which can be a two-dimensional map, a three-dimensional map or other forms of map data, and can provide accurate basic data support for subsequent positioning, navigation and other applications. The feature association result is fused to generate target map data, which can provide more accurate and comprehensive map information, thereby further meeting the offline mapping needs of users.

[0049] Based on the steps S11 to S14, by aligning the key frame data corresponding to different historical driving trajectories, and performing feature association and fusion processing, the accuracy and completeness of the map data can be improved, and the situation of inaccurate or missing map information caused by the deficiency of a single data source can be reduced. By processing the key frame data and feature association of the target vehicle on the preset driving route, the positioning accuracy of the target vehicle during driving can be improved, the positioning error can be reduced, and the navigation accuracy can be improved. Finally, based on the fusion processing of the feature association result, the target map data is generated, which can effectively reduce the cost and time of map updating, reduce the need for manual intervention, improve the updating speed and efficiency of the map data, and provide more accurate and complete map information for the automatic driving system, thereby improving the performance and stability of the automatic driving system, and improving the driving safety and efficiency of the vehicle. Therefore, the multi-trajectory mapping method provided by the embodiments of the present application achieves the purpose of efficiently and accurately generating target map data using key frame data collected in multiple historical driving trajectories, thereby achieving the technical effect of improving the mapping efficiency and accuracy in the offline mapping process, and further solving the technical problems of low mapping efficiency and poor accuracy in the related art during offline mapping.

[0050] The multi-trajectory mapping method in the embodiments of the present application will be further introduced below.

[0051] In an optional embodiment, the plurality of historical driving trajectories at least includes a first driving trajectory and a second driving trajectory, and the key frame data at least includes a plurality of first key frames corresponding to the first driving trajectory and a plurality of second key frames corresponding to the second driving trajectory. In step S12, the key frame data corresponding to different historical driving trajectories is aligned to obtain a plurality of key frame pairing results, including:

[0052] In step S121, the plurality of first key frames and the plurality of second key frames are subjected to position association processing to obtain a position association result, wherein the position association result is used to represent an initial pairing result of the first key frame and the second key frame.

[0053] In step S122, the position association result is subjected to spatio-temporal association processing to obtain a spatio-temporal association result, wherein the spatio-temporal association result is used to represent a key frame chain satisfying a preset spatio-temporal continuity condition.

[0054] In step S123, the spatio-temporal association result is used to determine the plurality of key frame pairing results.

[0055] Specifically, taking the first driving track as a historical driving track A and the second driving track as a historical driving track B as an example, the plurality of first key frames corresponding to the historical driving track A are A1, A2, A3, and A4, and the plurality of second key frames corresponding to the historical driving track B are B1, B2, B3, B4, B5, and B6. The position association processing is performed on the plurality of first key frames and the plurality of second key frames to obtain a position association result, so that the second key frame associated with each first key frame can be obtained.

[0056] Further, the spatio-temporal association processing is performed based on the position association result to obtain a spatio-temporal association result, and the spatio-temporal association result is a key frame chain satisfying a preset spatio-temporal continuity condition. For example, in the position association result, the second key frame B2 has the strongest association with the first key frame A1, and the second key frame B3 has the strongest association with the first key frame A2. In the process of aligning the key frame data corresponding to different historical driving tracks, the association of the key frames is realized based on the consistency and continuity of space and time. That is, the first key frames A1 and A2 adjacent in the historical driving track A are continuous in time and space, and the second key frames B2 and B3 associated with the first key frames A1 and A2 on the historical driving track B also need to be continuous in time and space. According to the spatio-temporal continuity, the longest continuous key frame chain satisfying the preset spatio-temporal continuity condition is the final spatio-temporal association result. That is, the time and space distance between A1 and A2 and the time and space distance between B2 and B3 are within a specified range. After obtaining the key frame chain satisfying the preset spatio-temporal continuity condition in the above first driving track and second driving track, the final plurality of key frame pairing results can be determined according to the key frame chain. For example, the plurality of key frame pairing results can be (A1-B2), (A2-B3), (A3-B4), and (A4-B6).

[0057] Based on the above optional embodiment, the position association processing is performed on the plurality of first key frames and the plurality of second key frames to obtain a position association result, and then the spatio-temporal association processing is performed based on the position association result to obtain a spatio-temporal association result, and finally the plurality of key frame pairing results are determined according to the spatio-temporal association result. Therefore, the time and space continuity before and after can be considered when the plurality of track information is associated, so that the association relationship between the plurality of collected information can be accurately calculated, and the mapping efficiency and accuracy in the offline mapping process are further improved.

[0058] In an optional embodiment, in step S13, the feature association processing is performed on the plurality of key frame pairing results to obtain a feature association result, including:

[0059] In step S131, vehicle pose information and lane line information are obtained based on the plurality of key frame pairing results.

[0060] In step S132, first association processing is performed on the vehicle pose information and the lane line information to obtain lane line association results, wherein the lane line association results are used to represent the lane line correspondence relationship in the key frame pairing results.

[0061] In step S133, second association processing is performed on the lane line association results to obtain feature association results.

[0062] Specifically, taking the multiple key frame pairing results (A1-B2), (A2-B3), (A3-B4), and (A4-B6) as examples, vehicle pose information and lane line information of each key frame are obtained based on the multiple key frame pairing results. In the vehicle pose information corresponding to A1, four lane lines a1, a2, a3, and a4 can be observed, and in the vehicle pose information corresponding to B2, three lane lines b1, b2, and b3 can be observed.

[0063] Further, first association processing is performed on the vehicle pose information and the lane line information to obtain lane line association results. The first association processing is a coarse association process, in which lane lines and lane boundaries need to be associated respectively. In the feature association, the spatial continuity of adjacent features needs to be ensured. If the positional deviation between the associated features of adjacent key frames is greater than a preset value, it is determined that the adjacent key frames are discontinuous in space. If the positional deviation is less than or equal to the preset value, it is determined that the adjacent key frames are continuous in space. The optimal feature chain that is continuous in time and space is the final lane line association result, that is, after coarse processing, the corresponding relationship of (a1-b1), (a2-b2), and (a3-b4) can be output.

[0064] After obtaining the lane line association results, second association processing is performed on the lane line association results. The second association processing is a fine association process, and the feature association results are obtained through the fine association processing.

[0065] Based on the above optional embodiments, by obtaining vehicle pose information and lane line information based on multiple key frame pairing results, then performing first association processing on the vehicle pose information and the lane line information to obtain lane line association results, and finally performing second association processing on the lane line association results, an accurate feature association result can be quickly obtained, which is used for accurate offline mapping and improves the mapping efficiency and accuracy.

[0066] In an optional embodiment, in step S132, the first association processing on the vehicle pose information and the lane line information to obtain the lane line association results includes:

[0067] In step S1321, feature matching is performed on the lane line information to obtain first matching results.

[0068] In step S1322, the first matching result is detected based on the vehicle pose information to obtain a target detection result, wherein the target detection result is used to determine whether the lane line association distance between adjacent key frames meets a preset distance condition.

[0069] In step S1323, the lane line association result is determined based on the target detection result.

[0070] Specifically, in the coarse association process, the lane line information is used for feature matching, and the lane lines and lane boundaries in the key frames can be associated respectively to obtain the first matching result. Further, in the feature association, the spatial continuity of adjacent features needs to be ensured, and the first matching result is detected based on the vehicle pose information to obtain the target detection result. For example, according to the lane line association of the first key frame A1 and the second key frame B2 determined by the first matching result, and the lane line association of the adjacent first key frame A2 and the second key frame B3, in order to ensure the time and spatial continuity of the association, the first matching result is detected based on the vehicle pose information. If the lane line association distance corresponding to A1 and B2 is consistent with the lane line association distance corresponding to A2 and B3, that is, the lane line association distance between adjacent key frames meets the preset distance condition, then the lane line association result can be determined according to the target detection result. If only the lane line association in A1 and B2 is considered in the coarse association, pairing errors may occur. By determining the lane line association result based on the target detection results of a plurality of key frames in a set, the time and spatial association can be effectively considered, thereby effectively reducing the association errors and improving the accuracy of offline mapping.

[0071] Based on the above optional embodiment, by using the lane line information for feature matching to obtain the first matching result, and then detecting the first matching result based on the vehicle pose information to obtain the target detection result, and finally determining the lane line association result based on the target detection result, the accuracy of the association relationship between the multiple collected information can be improved, and the mapping efficiency and accuracy of offline mapping can be further improved.

[0072] In an optional embodiment, in step S133, the second association process is performed based on the lane line association result to obtain a feature association result, including:

[0073] In step S1331, the vehicle pose information is adjusted based on the lane line association result to obtain a pose adjustment result.

[0074] In step S1332, the second association process is performed based on at least one of the pose adjustment result and the map feature in the key frame pairing result to obtain the feature association result.

[0075] Specifically, the lane line association result contains the lane corresponding relationship output after rough processing, i.e., (a1-b1), (a2-b2), (a3-b4), the vehicle pose information of the first key frame A1 and the second key frame B2 is adjusted based on the lane line association result, so that the lane lines are aligned to overlap in space, and the pose adjustment result is obtained. Based on at least one map feature in the pose adjustment result and the key frame pairing result, fine association processing is performed to obtain a feature association result. By adjusting the pose of the key frame and the position of the map feature, all map features between multiple trajectories can be aligned together, thereby performing fine association processing to improve the accuracy of the feature association result.

[0076] In an optional embodiment, at step S1332, performing second association processing based on at least one map feature in the pose adjustment result and the key frame pairing result to obtain a feature association result includes: performing feature matching based on at least one map feature in the pose adjustment result and the key frame pairing result to obtain a second matching result; performing pose update processing based on the second matching result to obtain a pose update result; and in response to the pose update result satisfying a preset convergence condition, determining the feature association result based on the pose update result.

[0077] Specifically, when performing feature matching, the discrete elements that need to be matched include but are not limited to lane lines, sidewalks, stop lines, arrows, road edges, etc. After performing feature matching based on at least one map feature in the pose adjustment result and the key frame pairing result, a second matching result is obtained, which contains the matching relationship between each discrete element. Based on the matching relationship, nonlinear optimization processing can be performed. The above feature matching process and pose update process are repeated until the preset convergence condition is met, i.e., the matching converges to the optimal result, thereby obtaining the final feature association result.

[0078] The goal of the above nonlinear optimization is to minimize the distance between each lane line and the matching pair of ground elements, while also satisfying the combined inertial navigation constraint between two consecutive key frames. Finally, the pose update result of each optimized key frame is output, which can ensure that the lane lines of the historical driving trajectory A and the historical driving trajectory B are aligned in the new pose.

[0079] Based on the above optional embodiment, by performing feature matching based on at least one map feature in the pose adjustment result and the key frame pairing result to obtain a second matching result, and then performing pose update processing based on the second matching result to obtain a pose update result, and finally determining the feature association result based on the pose update result in response to the pose update result satisfying a preset convergence condition, the reliability of the feature association result can be further improved through fine association processing, thereby improving the accuracy of offline mapping.

[0080] In an optional embodiment, in step S14, the target map data is generated by performing fusion processing based on the feature association result, which comprises:

[0081] In step S141, the feature association result is filtered by using a preset filtering condition to obtain a filtering result, wherein the preset filtering condition is used to filter the false pairing relationship in the feature association result.

[0082] In step S142, the target map data is generated by performing fusion processing based on the filtering result.

[0083] Specifically, after obtaining the feature association result, an accurate filtering error association mechanism is introduced, and the false pairing relationship in the feature association result is filtered by using a preset filtering condition, so that the false association can be effectively excluded, and the accuracy of the mapping is ensured. The association relationship of the road segment that is not aligned in the nonlinear optimization result is removed, and mapping is not performed on these road segments, so that the error map is not generated. Based on the association information in the filtering result, offline mapping is performed, and more accurate and complete target map data can be generated. The user can more accurately perform position positioning and navigation by using the target map data, and the practicality and user experience of the target map data are improved.

[0084] Based on the above optional embodiment, by using the preset filtering condition to filter the feature association result to obtain the filtering result, the association data that does not meet the condition can be effectively excluded, the errors in the data are reduced, and the accuracy and reliability of the target map data generated finally are further improved. In addition, the preset filtering condition and the automatic fusion processing can reduce manual intervention, improve the efficiency and speed of offline mapping, and save time and cost.

[0085] Figure 2 is a flowchart of another multi-track mapping method according to an embodiment of the present application, as shown in Figure 2 The method comprises the following steps:

[0086] In step S201, key frame data corresponding to a plurality of historical driving tracks of a target vehicle on a preset driving route is obtained, wherein the key frame data comprises vehicle pose information in the historical driving track and at least one map feature corresponding to the vehicle pose information.

[0087] In step S202, the plurality of first key frames and the plurality of second key frames are positionally associated to obtain a position association result, wherein the position association result is used to represent an initial pairing result of the first key frames and the second key frames.

[0088] In step S203, the spatio-temporal association result is obtained by performing spatio-temporal association processing based on the position association result, wherein the spatio-temporal association result is used to represent a key frame chain that meets a preset spatio-temporal continuity condition.

[0089] Step S204, determining a plurality of key frame pairing results according to the spatio-temporal association result;

[0090] Step S205, obtaining vehicle pose information and lane line information based on the plurality of key frame pairing results;

[0091] Step S206, performing first association processing using the vehicle pose information and the lane line information to obtain lane line association results, wherein the lane line association results are used to represent the lane line correspondence relationship in the key frame pairing results;

[0092] Step S207, performing second association processing using the lane line association results to obtain feature association results;

[0093] Step S208, performing screening processing on the feature association results using a preset screening condition to obtain screening results, wherein the preset screening condition is used to filter the incorrect pairing relationship in the feature association results;

[0094] Step S209, performing fusion processing based on the screening results to generate target map data.

[0095] Based on the above steps S201 to S209, by obtaining the key frame data corresponding to a plurality of historical driving trajectories of the target vehicle on the preset driving route, then performing alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results, subsequently performing feature association processing on the plurality of key frame pairing results to obtain feature association results, and finally performing fusion processing based on the feature association results to generate target map data, the purpose of efficiently and accurately generating target map data using the key frame data collected in a plurality of historical driving trajectories is achieved, thereby realizing the technical effect of improving the mapping efficiency and accuracy in the offline mapping process, and further solving the technical problems of low mapping efficiency and poor accuracy in the related art when performing offline mapping.

[0096] Figure 3 is a schematic diagram of a multi-trajectory mapping method according to one embodiment of the present application, as Figure 3As shown, the plurality of historical driving trajectories of the target vehicle on the preset driving route include: a historical driving trajectory A, a historical driving trajectory B, a historical driving trajectory C, and the like, key frame data corresponding to each historical driving trajectory is acquired, the key frame data includes vehicle pose information in the historical driving trajectory and at least one map feature corresponding to the vehicle pose information. The key frame data corresponding to different historical driving trajectories is aligned to obtain a plurality of key frame pairing results. Further, the vehicle pose information and the lane line information are acquired based on the plurality of key frame pairing results, the lane line association result is obtained by using the vehicle pose information and the lane line information for coarse association processing, and the feature association result is obtained by using the lane line association result for fine association processing. The feature association result is filtered by using a preset screening condition, and the preset screening condition is used to filter the error pairing relationship in the feature association result. The screening result that passes through the preset screening condition is fused for mapping to generate target map data, and the feature association result that does not pass through the preset screening condition is aligned again.

[0097] In the embodiment of the present application, the problem of insufficient single mapping capability can be solved by multiple data collection and offline mapping. In addition, the key frame association method based on time and spatial continuity in the embodiment of the present application can effectively solve the error association problem existing in the related art based on the spatial nearest method. The feature association method based on time and spatial continuity in the embodiment of the present application can effectively improve the accuracy of feature matching, and by using the iterative method for feature optimization and the accurate error association screening mechanism, the target map data can be efficiently and accurately generated by using the key frame data collected in the plurality of historical driving trajectories.

[0098] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by means of software and a general hardware platform as necessary, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, which is stored in a storage medium (such as a ROM / RAM, a magnetic disk, or an optical disc), and includes a plurality of instructions for causing a terminal device (which can be a mobile phone, a computer, a server, or a network device) to execute the methods described in the various embodiments of the present application.

[0099] In the embodiments of the present application, a multi-trajectory mapping device is also provided, which is used to implement the above embodiments and preferred embodiments, and will not be described again. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware, or a combination of software and hardware is also possible and contemplated.

[0100] Figure 4 is a structural block diagram of a multi-track mapping device according to an embodiment of the present application. As shown in the figure, the device comprises: Figure 4

[0101] The acquisition module 401 is configured to acquire key frame data corresponding to a plurality of historical driving tracks of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving track and at least one map feature corresponding to the vehicle pose information.

[0102] The processing module 402 is configured to perform alignment processing on the key frame data corresponding to different historical driving tracks to obtain a plurality of key frame pairing results.

[0103] The association module 403 is configured to perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results.

[0104] The generation module 404 is configured to perform fusion processing based on the feature association result to generate target map data.

[0105] Optionally, the processing module 402 is further configured to perform position association processing on the plurality of first key frames and the plurality of second key frames to obtain a position association result, wherein the position association result is used to represent an initial pairing result of the first key frames and the second key frames; perform spatio-temporal association processing based on the position association result to obtain a spatio-temporal association result, wherein the spatio-temporal association result is used to represent a key frame chain satisfying a preset spatio-temporal continuity condition; and determine the plurality of key frame pairing results according to the spatio-temporal association result.

[0106] Optionally, the association module 403 is further configured to acquire vehicle pose information and lane line information based on the plurality of key frame pairing results; perform first association processing on the vehicle pose information and the lane line information to obtain a lane line association result, wherein the lane line association result is used to represent a lane line correspondence relationship in the key frame pairing result; and perform second association processing on the lane line association result to obtain the feature association result.

[0107] Optionally, the association module 403 is further configured to perform feature matching on the lane line information to obtain a first matching result; perform detection processing on the first matching result based on the vehicle pose information to obtain a target detection result, wherein the target detection result is used to determine whether a lane line association distance between adjacent key frames satisfies a preset distance condition; and determine the lane line association result based on the target detection result.

[0108] ​Optionally, the association module 403 is further configured to: perform adjustment processing on the vehicle pose information based on the lane line association result, to obtain a pose adjustment result; perform second association processing on at least one map feature in the pose adjustment result and the key frame pairing result, to obtain a feature association result.

[0109] Optionally, the association module 403 is further configured to: perform feature matching based on at least one map feature in the pose adjustment result and the key frame pairing result, to obtain a second matching result; perform pose updating processing based on the second matching result, to obtain a pose updating result; and in response to the pose updating result satisfying a preset convergence condition, determine the feature association result based on the pose updating result.

[0110] Optionally, the generation module 404 is further configured to: perform screening processing on the feature association result by using a preset screening condition, to obtain a screening result, wherein the preset screening condition is used to filter an error pairing relationship in the feature association result; and perform fusion processing based on the screening result, to generate the target map data.

[0111] Embodiments of the present application also provide a vehicle, comprising: a memory storing an executable program; and a processor configured to execute the program, wherein the program performs the method in the embodiments of the present application when executed.

[0112] Optionally, in the embodiment, the processor can be configured to perform the following steps by using the computer program:

[0113] S1, obtaining key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information;

[0114] S2, performing alignment processing on the key frame data corresponding to different historical driving trajectories, to obtain a plurality of key frame pairing results;

[0115] S3, performing feature association processing on the plurality of key frame pairing results, to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results;

[0116] S4, performing fusion processing based on the feature association result, to generate target map data.

[0117] Embodiments of the present application also provide a computer readable storage medium, comprising a stored executable program, wherein the executable program controls a device where the computer readable storage medium is located to perform the method in the embodiments of the present application when executed.

[0118] Optionally, in the embodiment, the computer readable storage medium can be configured to store a computer program for performing the following steps:

[0119] S1, acquire key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information;

[0120] S2, perform alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results;

[0121] S3, perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results;

[0122] S4, perform fusion processing based on the feature association result to generate target map data.

[0123] Embodiments of the present application also provide a computer program product comprising a computer program which, when executed by a processor, implements the method in any of the embodiments of the present application.

[0124] Optionally, the computer program product comprises a computer program which, when executed by a processor, performs the following steps:

[0125] S1, acquire key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information;

[0126] S2, perform alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results;

[0127] S3, perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results;

[0128] S4, perform fusion processing based on the feature association result to generate target map data.

[0129] Embodiments of the present application also provide a computer program product comprising a non-volatile computer readable storage medium for storing a computer program, which, when executed by a processor, implements the method in any of the embodiments of the present application.

[0130] Optionally, the computer program stored in the non-volatile computer readable storage medium, when executed by a processor, performs the following steps:

[0131] S1, acquire key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information;

[0132] S2, perform alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results;

[0133] S3, perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results;

[0134] S4, perform fusion processing based on the feature association result to generate target map data.

[0135] Embodiments of the present application also provide a computer program, which, when executed by a processor, implements the method in any of the above embodiments of the present application.

[0136] Optionally, the computer program is executed by the processor to perform the following steps:

[0137] S1, acquire key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information;

[0138] S2, perform alignment processing on the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results;

[0139] S3, perform feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results;

[0140] S4, perform fusion processing based on the feature association result to generate target map data.

[0141] In the above embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0142] The serial numbers of the above embodiments of the present application only serve for description, and do not represent the advantages and disadvantages of the embodiments.

[0143] In the above embodiments of the present application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0144] In several embodiments provided in the present application, it should be understood that the disclosed technology can be implemented by other ways. Among them, the above-described device embodiments are only schematic, for example, the division of the units can be 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 units can be indirect coupling or communication connection through some interfaces, units or modules, and can be electrical or other forms.

[0145] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed to multiple units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0146] In addition, the functional units in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0147] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application essentially or the part of the prior art that contributes to the technical solutions or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The foregoing storage medium includes: a U disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0148] The above is only the preferred embodiment of the present application, and it should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, and these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A multi-track mapping method, characterized by, The method comprises: obtaining key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information; aligning the key frame data corresponding to different historical driving trajectories to obtain a plurality of key frame pairing results; performing feature association processing on the plurality of key frame pairing results to obtain a feature association result, wherein the feature association result is used to determine the feature correspondence relationship between the plurality of key frame pairing results; performing fusion processing based on the feature association result to generate target map data; wherein the feature association processing on the plurality of key frame pairing results to obtain the feature association result comprises: obtaining the vehicle pose information and lane line information based on the plurality of key frame pairing results; performing first association processing on the vehicle pose information and the lane line information to obtain a lane line association result, wherein the lane line association result is used to represent the lane line correspondence relationship in the key frame pairing result; performing second association processing on the lane line association result to obtain the feature association result.

2. The multi-track mapping method of claim 1, wherein, The plurality of historical driving trajectories at least include: a first driving trajectory and a second driving trajectory, and the key frame data at least includes: a plurality of first key frames corresponding to the first driving trajectory, and a plurality of second key frames corresponding to the second driving trajectory, and the aligning the key frame data corresponding to different historical driving trajectories to obtain the plurality of key frame pairing results comprises: performing position association processing on the plurality of first key frames and the plurality of second key frames to obtain a position association result, wherein the position association result is used to represent the initial pairing result of the first key frame and the second key frame; performing spatio-temporal association processing based on the position association result to obtain a spatio-temporal association result, wherein the spatio-temporal association result is used to represent a key frame chain that satisfies a preset spatio-temporal continuity condition; determining the plurality of key frame pairing results according to the spatio-temporal association result.

3. The multi-track mapping method of claim 1, wherein, The first association processing on the vehicle pose information and the lane line information to obtain the lane line association result comprises: performing feature matching on the lane line information to obtain a first matching result; performing detection processing on the first matching result based on the vehicle pose information to obtain a target detection result, wherein the target detection result is used to determine whether the lane line association distance between adjacent key frames satisfies a preset distance condition; determining the lane line association result based on the target detection result.

4. The multi-track mapping method of claim 1, wherein, The second association processing on the lane line association result to obtain the feature association result comprises: performing adjustment processing on the vehicle pose information based on the lane line association result to obtain a pose adjustment result; performing second association processing based on the pose adjustment result and the at least one map feature in the key frame pairing result to obtain the feature association result.

5. The multi-track mapping method of claim 4, wherein, The second association processing is performed based on the at least one map feature in the pose adjustment result and the key frame pairing result, and the feature association result is obtained. The feature matching is performed based on the at least one map feature in the pose adjustment result and the key frame pairing result, and a second matching result is obtained. The pose updating processing is performed based on the second matching result, and a pose updating result is obtained. The feature association result is determined based on the pose updating result in response to the pose updating result satisfying a preset convergence condition.

6. The multi-track mapping method of claim 1, wherein, The fusion processing is performed based on the feature association result, and the target map data is generated. The feature association result is filtered by using a preset screening condition, and a screening result is obtained, wherein the preset screening condition is used to filter an error pairing relationship in the feature association result. The fusion processing is performed based on the screening result, and the target map data is generated.

7. A multi-track mapping device, comprising: The method comprises: The acquisition module is configured to acquire key frame data corresponding to a plurality of historical driving trajectories of a target vehicle on a preset driving route, wherein the key frame data comprises vehicle pose information in the historical driving trajectories and at least one map feature corresponding to the vehicle pose information. The processing module is configured to perform alignment processing on the key frame data corresponding to different historical driving trajectories, and obtain a plurality of key frame pairing results. The association module is configured to perform feature association processing on the plurality of key frame pairing results, and obtain a feature association result, wherein the feature association result is used to determine a feature correspondence relationship between the plurality of key frame pairing results. The generation module is configured to perform fusion processing based on the feature association result, and generate target map data. The association module is further configured to acquire the vehicle pose information and lane line information based on the plurality of key frame pairing results, perform first association processing by using the vehicle pose information and the lane line information, and obtain a lane line association result, wherein the lane line association result is used to represent a lane line correspondence relationship in the key frame pairing result; and perform second association processing by using the lane line association result, and obtain the feature association result.

8. A vehicle characterized by comprising: The method comprises: A processor; A memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the multi-trajectory mapping method of any one of claims 1 to 6.

9. A non-volatile storage medium, comprising: The storage medium stores a computer program, wherein the computer program is configured to execute the multi-trajectory mapping method of any one of claims 1 to 6 when running.

10. A computer program product, characterised in that, The computer program product comprises computer instructions which, when executed by a processor, implement the multi-trajectory mapping method of any one of claims 1 to 6.

Images