Mapping method, device, computer equipment, storage medium and program product

Abstract

The present application relates to a mapping method, device, computer equipment, storage medium and program product. It relates to the field of intelligent driving technology. The method includes: based on the dead reckoning trajectory, optimizing the first GPS trajectory in the current trajectory optimization area to determine the optimized trajectory, wherein the trajectory optimization area includes: the partial area optimized last time and the unoptimized area; constructing vector map elements based on the perception information and the optimized trajectory; constructing the road structure based on the constructed vector map elements to obtain the target map. The method is used to achieve fast and accurate mapping, so that the map can be used after one construction.

Description

Technical Field

[0001] The present application relates to the field of intelligent driving technology, and in particular to a mapping method, apparatus, computer equipment, storage medium and program product. Background Art

[0002] At present, the mainstream method for building urban scene maps is crowdsourcing mapping. Crowdsourcing mapping solutions are generally divided into two main modules: vehicle-side and cloud-end. The vehicle-side is equipped with surveying and mapping equipment, and performs some lightweight data processing in real time. A large amount of processed data will be transmitted to the cloud, which is responsible for fusing the large amount of fragmented data sent back to achieve mapping. This method involves uploading and sending data, and the large amount of fragmented data uploaded by the vehicle-side needs to be processed by the cloud-end through complex aggregation algorithms to achieve mapping. Therefore, the efficiency is often low when building maps, and it is impossible to generate an instantly available reconstructed map based on the data on the vehicle-side in real time. Summary of the invention

[0003] Based on this, it is necessary to provide a lane centerline construction method, device, computer equipment, storage medium and program product to address the above technical problems.

[0004] In a first aspect, the present application provides a mapping method, comprising:

[0005] Based on the dead reckoning trajectory, optimizing the first GPS trajectory in the current trajectory optimization area to determine an optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area;

[0006] constructing vector map elements based on the perception information and the optimized trajectory;

[0007] The road structure is constructed based on the constructed vector map elements to obtain the target map.

[0008] In a second aspect, the present application also provides a mapping device, including:

[0009] A trajectory optimization module, for performing trajectory optimization on a first GPS trajectory in a current trajectory optimization area based on a dead reckoning trajectory to determine an optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area;

[0010] An element construction module, used for constructing vector map elements based on the perception information and the optimized trajectory;

[0011] The road structure construction module is used to construct the road structure based on the constructed vector map elements to obtain the target map.

[0012] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0013] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method described in the first aspect.

[0014] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0015] Projecting the vehicle trajectory points onto the constructed lane boundaries to determine a plurality of projection points;

[0016] Perform lane tracking and matching based on the multiple projection points to determine a target lane;

[0017] The lane centerline in the target lane is constructed.

[0018] The above-mentioned mapping method, device, computer equipment, storage medium and computer program product optimize the first GPS trajectory in the current trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory, wherein the trajectory optimization area includes: the partial area optimized last time and the unoptimized area; construct vector map elements based on the perception information and the optimized trajectory; and construct the road structure based on the constructed vector map elements to obtain the target map. Through this scheme, a segmented trajectory optimization strategy is proposed through the trajectory optimization area, so the vehicle can achieve real-time trajectory optimization and improve the mapping efficiency. Moreover, since the trajectory optimization area includes the area optimized last time, there is a certain overlap area, which can improve the mapping accuracy. Therefore, with the improvement of efficiency and accuracy, fast and accurate mapping can be achieved, and the effect of using the map after one construction can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0020] Figure 1A An application environment diagram of a mapping method in an embodiment;

[0021] Figure 1B A schematic diagram of a flowchart of a mapping method in an embodiment;

[0022] Figure 2A Schematic diagram 1 of a lane centerline construction method flow in one embodiment;

[0023] Figure 2B A schematic diagram of projecting a vehicle trajectory point onto a lane boundary in one embodiment;

[0024] Figure 3 is a schematic diagram of a trajectory-based linear natural coordinate system;

[0025] Figure 4 Schematic diagram 2 of a lane centerline construction method flow in one embodiment;

[0026] Figure 5 A schematic diagram of multiple layers in a memory route map in one embodiment;

[0027] Figure 6 A schematic diagram of a process of a method for quickly building a map on the vehicle side in an embodiment;

[0028] Figure 7 A schematic diagram of segmented trajectory optimization in one embodiment;

[0029] Figure 8 A schematic diagram of a process flow of trajectory optimization in one embodiment;

[0030] Fig. 9 A linearly connected road structure model in one embodiment;

[0031] Fig.10 A structural block diagram of a lane centerline construction device in an embodiment;

[0032] Fig.11 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0033] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0034] The mapping method provided in the embodiment of the present application can be applied to a vehicle-side device or a cloud-side device. The vehicle-side device can be a vehicle-mounted terminal, and the cloud-side device can be a cloud-side server.

[0035] The mapping method provided in the embodiment of the present application can be applied to Figure 1AIn the application environment shown. Among them, the vehicle terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The vehicle terminal 102 can execute the mapping method provided in the embodiment of the present application, and send the constructed map to the server 104 after the mapping is completed.

[0036] The terminal 102 may be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, IoT devices, and portable wearable devices. The IoT devices may be smart speakers, smart TVs, smart air conditioners, smart car-mounted devices, etc. The portable wearable devices may be smart watches, smart bracelets, head-mounted devices, etc. The server 104 may be implemented as an independent server or a server cluster consisting of multiple servers.

[0037] In an exemplary embodiment, Figure 1B As shown, a schematic flow chart of a mapping method is provided, and the method includes the following steps S11 to S13.

[0038] S11. Based on the dead reckoning trajectory, optimize the first GPS trajectory in the current trajectory optimization area to determine an optimized trajectory.

[0039] Before executing the above step S11, the first global positioning system GPS trajectory of the vehicle in the current trajectory optimization area can be obtained, and the dead reckoning trajectory of the vehicle in the current trajectory optimization area can be obtained. The above first GPS trajectory can reflect the position sequence and other conditions passed by the vehicle. The dead reckoning trajectory is a vehicle driving trajectory calculated according to a certain algorithm using some sensor information of the vehicle itself (such as vehicle speed, heading angle, etc.).

[0040] The trajectory optimization area includes: the partial area optimized last time and the unoptimized area.

[0041] In the embodiment of the present application, after obtaining the above two types of trajectory data, the first GPS trajectory in the current trajectory optimization area is optimized based on the dead reckoning trajectory. The specific optimization algorithm and method may vary depending on the specific application scenario and technical implementation. Its goal is to correct and improve the first GPS trajectory by combining the characteristics of the dead reckoning trajectory, and finally determine an optimized trajectory that has better performance in terms of accuracy, coherence, etc. compared to the original first GPS trajectory.

[0042] S12. Construct vector map elements based on the perception information and the optimized trajectory.

[0043] The above-mentioned perception information may generally include information about the surrounding environment acquired by the vehicle through various sensors (such as cameras, laser radars, etc.). This information can reflect the characteristics, positions, etc. of various objects such as roads, buildings, and traffic signs around the vehicle during driving.

[0044] In the embodiment of the present application, after obtaining the optimized trajectory and perception information, these data can be used to construct vector map elements. Among them, vector map elements are the basic units that constitute the vector map, such as line segments representing roads, polygons representing buildings, etc. By comprehensively analyzing and processing the perception information and the optimized trajectory, the relevant information is converted into vector map elements according to certain rules and algorithms, preparing for the subsequent construction of a complete map.

[0045] S13. Construct a road structure based on the constructed vector map elements to obtain a target map.

[0046] After a certain number of vector map elements have been constructed, the road structure can be constructed based on these vector map elements. The road structure may include, but is not limited to, road connection relationships, directions, number of lanes, and other aspects. For example, by analyzing the connection between line segments representing different road sections in the vector map elements and their directions and other attributes, the layout and structure of the entire road network can be determined.

[0047] Based on the constructed road structure, other relevant information is further improved and supplemented (such as traffic signs, surrounding buildings, etc., which can also be reflected through the previously constructed vector map elements), and finally a complete and usable target map is formed. This target map is the expected result of the entire mapping process, which can accurately reflect the road and surrounding environment of the vehicle driving area, achieving the effect of one-time mapping and ready for use.

[0048] In the above embodiment, a segmented trajectory optimization strategy is proposed through the trajectory optimization area, so the vehicle can achieve real-time trajectory optimization and improve the mapping efficiency. Moreover, since the trajectory optimization area includes the area optimized last time, there is a certain overlap area, which can improve the accuracy of mapping. Therefore, with the improvement of efficiency and accuracy, fast and accurate mapping can be achieved, so that the mapping can be used once it is built.

[0049] In some embodiments, the above-mentioned constructed vector map elements include constructed lane boundaries. In the embodiments of the present disclosure, the vehicle trajectory points can be projected onto the constructed lane boundaries to determine multiple projection points; lane tracking and matching can be performed based on the multiple projection points to determine the target lane; the lane centerline of the target lane is constructed, and the road structure is constructed based on the lane centerline of the target lane to obtain the target map.

[0050] In an exemplary embodiment, Figure 2A As shown, a flow chart of a lane centerline construction method is provided, and the method includes the following steps S201 to S203.

[0051] S201: Project the vehicle trajectory points onto the constructed lane boundary to determine a plurality of projection points.

[0052] The lane boundary may include: lane lines, road boundaries (eg, curbs), etc. The constructed lane boundary may be constructed based on visual perception information.

[0053] In the embodiment of the present application, the above-mentioned vehicle trajectory point can be any trajectory point in the vehicle trajectory.

[0054] In some embodiments, the vehicle trajectory may be a GPS trajectory or an Odometry (ODOM) trajectory.

[0055] Among them, in the process of projecting the vehicle trajectory point onto the lane boundary to determine multiple projection points, multiple projection vertical lines passing through the vehicle trajectory point and perpendicular to the lane boundary can be first determined; and the intersection points of the multiple projection vertical lines and the lane boundary are determined to determine the multiple projection points.

[0056] For example, Figure 2B As shown in , it is a schematic diagram of projecting the vehicle trajectory points to the lane boundary. Figure 2B As shown, the vehicle trajectory point A is projected onto the lane boundary 21 and the lane boundary 22 to determine the projection points A1 and A2.

[0057] In some embodiments, the lane centerline construction method provided in the embodiments of the present application can be applied to a Cartesian coordinate system or a linear natural coordinate system.

[0058] The origin of the linear natural coordinate system is the natural number 0, and the coordinate axis is the mileage of the vehicle.

[0059] The above-mentioned vehicle trajectory points may be trajectory points in a linear natural trajectory, and the linear natural trajectory is a trajectory of the vehicle driving trajectory in a linear natural coordinate system.

[0060] In the stage of vectorizing the map and the stage of constructing the map topology, the map elements can be processed in the above-mentioned linear natural coordinate system, and the mileage of the corresponding coordinate axis is the coordinate of the map element in the linear natural coordinate system.

[0061] For example, Figure 3 is a schematic diagram of a trajectory-based linear natural coordinate system, such as Figure 3As shown in , the origin of the linear natural coordinate system is 0, and the coordinates in the coordinate system are the mileage corresponding to the vehicle's driving trajectory. After the map element is projected onto the linear natural coordinate system, the mileage on the coordinate axis is the coordinate of the map element in the linear natural coordinate system (expressed by index), as shown in Figure 3 The coordinate index 0 (index0), coordinate index 1 (index1), coordinate index 2 (index2), and coordinate index (index3) shown in are respectively Figure 3 The coordinate points corresponding to the map elements indicated by the star marks.

[0062] The above-mentioned stage of vectorizing the map may be a stage of reconstructing the elements in the map. For example, the elements may include but are not limited to: lane lines, ground markings, traffic lights, and pole signs. Among them, the ground markings may include but are not limited to: stop lines, zebra crossings, and road arrows.

[0063] The above-mentioned map topology construction phase refers to the phase of constructing the map road structure.

[0064] In the stage of vectorizing the map and the stage of constructing the map topology, the above-mentioned linear natural coordinate system can be used for processing. On the one hand, the one-dimensional coordinate system can fully reflect the spatial relative position of the map elements, and the efficiency of the map element query operation is significantly improved; on the other hand, when querying the coordinate transformation matrix (for example, the elements need to be transformed in coordinates), it is more accurate than querying the corresponding information through timestamps.

[0065] S202: Perform lane tracking and matching based on multiple projection points to determine a target lane.

[0066] In some embodiments, lane tracking matching is performed based on multiple projection points to determine a target lane, which may include but is not limited to: pairing and combining based on multiple projection points to construct at least one set of lane point pairs; lane tracking matching is performed based on at least one set of lane point pairs to determine at least one lane.

[0067] Exemplarily, multiple projection points can be combined in pairs to determine multiple groups of lane point pairs. Then, lane point pairs in which two points are on both sides of the above-mentioned trajectory points can be screened out from the multiple groups of lane point pairs. At this time, the lane point pair can be considered to be the lane point pair of the lane where the vehicle is located. Lane tracking matching is performed based on the screened lane point pairs to determine the target lane.

[0068] For example, multiple projection points may be combined in pairs to determine multiple lane point pairs, and then lane point pairs with a difference less than a preset value may be determined from the multiple lane point pairs according to the difference between the preset lane width and the distance between the lane point pairs for lane tracking matching to determine the target lane. The preset lane width may prevent two or more lanes from being identified as one lane due to the missing of some lane boundaries. At this time, the selected lane point pairs may be lane point pairs of the lane where the vehicle is located, or lane point pairs of the lane on the side of the vehicle (such as the adjacent lane to the lane where the vehicle is located).

[0069] For example, multiple projection points may be combined in pairs to determine multiple lane point pairs, and then lane point pairs in which two points are located on both sides of the track point may be selected from the multiple lane point pairs. If there are multiple lane point pairs on both sides of the track point, then lane point pairs with a difference less than a preset value may be determined from the multiple lane point pairs according to the difference between the preset lane width and the distance between the lane point pairs to perform lane tracking matching to determine the target lane. The selected lane point pairs may be lane point pairs in the lane where the vehicle is located, or lane point pairs in the lane on one side of the vehicle (such as the adjacent lane to the lane where the vehicle is located).

[0070] Among them, the above-mentioned preset lane width is the standard value of the lane width on the current road.

[0071] In some embodiments, assuming that the vehicle determines multiple projection points at the current vehicle trajectory point, the vehicle can be sorted in the Y-axis direction according to the distance between the vehicle trajectory point, and the adjacent projection points among the multiple projection points are determined as lane point pairs, and the adjacent projection points are considered to form a lane. The Y-axis direction can be a direction perpendicular to the vehicle's driving direction.

[0072] In some embodiments, when the vehicle determines two projection points at the current vehicle trajectory point, and the distance between the two projection points is greater than the preset lane width, it can be determined whether the distance between the two projection points is greater than or equal to a multiple of the preset lane width. If such a situation is satisfied, multiple lanes can be determined based on the two projection points and the preset lane width. For example, assuming that the preset lane width is 3 meters and the distance between the two projection points is 6 meters, it can be determined that there are 2 lanes based on the two projection points.

[0073] The above situation where multiple projection points are determined at the current vehicle trajectory point corresponds to the following situation where the constructed lane point pair is bilateral.

[0074] In some embodiments, there is also a situation where no projection point is found when projecting the vehicle trajectory point onto the constructed lane boundary. In this case, two projection points can be determined based on the vehicle trajectory point and the preset lane width. For example, the vehicle trajectory point can be used as the midpoint to construct a line segment of the vehicle in the Y-axis direction with a length of the preset lane width, and the two endpoints of the line segment can be determined as the two projection points.

[0075] The above situation where no projection point is found at the current vehicle trajectory point corresponds to the following situation where the constructed lane point pair is boundless.

[0076] In some embodiments, when projecting the vehicle trajectory point onto the constructed lane boundary, there is also a situation where only one projection point is found. In this case, another projection point can be determined based on the position of the one projection point and the preset lane width. For example, the one projection point can be used as an endpoint of a line segment to construct a line segment of the vehicle in the Y-axis direction with a length of the preset lane width, and the other endpoint of the line segment can be determined as another projection point.

[0077] The above situation where only one projection point is found at the current vehicle trajectory point corresponds to the following situation where the constructed lane point pair is unilateral.

[0078] In some embodiments, lane tracking matching may be performed based on at least one set of lane point pairs based on a Kuhn-Munkres Algorithm (KM) algorithm to determine at least one lane.

[0079] In some embodiments, lane tracking matching is performed based on at least one set of lane point pairs to determine a target lane, which may include but is not limited to: if the target group lane point pair matches the tracking lane, the target group lane point pair is updated to the tracking lane to determine the target lane; if the target lane point pair does not match the tracking lane, a new lane is constructed based on the target group lane point pair to determine the target lane.

[0080] The target group of lane point pairs is any one of the at least one group of lane point pairs.

[0081] The above tracking lanes are composed of lane point pairs obtained by final lane matching based on the above KM algorithm.

[0082] When the target group lane point matches the tracking lane, it means that the target group lane point pair belongs to the tracking lane. At this time, the target lane point pair is added to the tracking lane. When the target group lane point does not match the tracking lane, it means that the target group lane point pair does not belong to the tracking lane and may belong to other lanes. At this time, a new lane can be constructed based on the target group lane point pair to determine the target lane. Subsequently, the constructed new lane can be used as a newly added tracking lane to continue matching with other lane point pairs.

[0083] S203: Construct a lane centerline of the target lane.

[0084] The target lane is composed of multiple lane point pairs. When constructing the lane centerline of the target lane, the centerline of the connection line of each lane point pair in the target lane may be determined, and the multiple connection centerlines obtained may be connected to obtain the lane centerline.

[0085] In some embodiments, after the lane centerline is determined by connecting multiple link centerlines, the lane centerline may be smoothed.

[0086] The lane centerline construction method projects the vehicle trajectory points onto the constructed lane boundary to determine multiple projection points, and performs lane tracking and matching based on the multiple projection points to determine the target lane and construct the lane centerline in the target lane. Through this solution, after the lane boundary is constructed, the target lane is determined by performing lane tracking and matching based on the multiple projection points of the vehicle trajectory points on the lane boundary, and the lane centerline is constructed based on the determined target lane, so that the lane centerline can be automatically constructed, thereby improving the efficiency and accuracy of lane centerline construction.

[0087] Furthermore, compared to directly constructing a lane centerline based on an already constructed lane boundary, in the present application, lane tracking and matching is performed through projection points to determine the target lane, and the lane centerline of the target lane is constructed. This can avoid the problem of inaccurate lane centerline directly determined due to the absence or offset of the already constructed lane boundary due to occlusion or unclear recognition.

[0088] In an exemplary embodiment, Figure 4 As shown, a second flowchart of a lane centerline construction method is provided, and the method includes the following steps S401 to S410.

[0089] S401, determining the current vehicle trajectory point.

[0090] The above vehicle trajectory point may be a trajectory point in a linear natural trajectory. Exemplarily, the vehicle trajectory point may be a current index in a linear natural coordinate system.

[0091] S402: Determine the projection point of the current vehicle trajectory point on the lane boundary in the current road.

[0092] Among them, in this application, the projection point of the current index on all lane boundaries in the current road can be calculated according to the current index.

[0093] S403: Construct lane point pairs according to the projection points.

[0094] Among them, lane point pairs (also called lane pairs) can be constructed based on the above projection points. The constructed lane point pairs may have a single side, double sides, or no side.

[0095] For example, a lane point pair consisting of one lane point on each side of the vehicle may be selected from the constructed lane point pairs, and subsequent lane tracking and matching may be continued to determine the lane centerline.

[0096] S404: Use an existing lane tracker to match lane point pairs.

[0097] Among them, the KM algorithm can be used to match lane point pairs for the existing lane tracker.

[0098] The lane tracker can be a tool or system for tracking and associating lane point pairs on a lane. It is used to match with existing lane point pairs and determine whether the match is successful. If the match is successful, the lane point pair is placed in the lane tracker; if it is not matched, a new lane tracker can be created based on the lane point pair to better track and manage the lane.

[0099] Exemplarily, when judging whether the lane point pair matches the existing lane tracker, it can be determined whether the lane point pair matches the existing lane tracker by judging whether the lane point pair and the lane point pair in the existing lane tracker form a continuous trajectory (i.e., the trajectory offset is small), and whether the distance between the lane point pair and the lane point pair in the existing lane tracker is within a certain range.

[0100] S405: Determine whether the match is successful.

[0101] When the match is successful, the following step S406 may be continued to be executed; when the match is unsuccessful, the following step S407 may be continued to be executed.

[0102] S406: Put the lane point pair into the lane tracker.

[0103] S407: Create a new lane tracker according to the lane point pair.

[0104] S408: Calculate the lane centerline.

[0105] In some embodiments, when calculating the lane centerline, the centerline of each lane point pair in the lane tracker may be determined, and the lane centerline may be obtained by connecting these centerlines.

[0106] S409: Move backward a preset step length based on the current vehicle trajectory point.

[0107] In the present application, the vehicle trajectory points can be selected with a preset step size and the above steps can be performed.

[0108] S410: Determine whether the route is exceeded.

[0109] After moving backward a preset step length based on the current vehicle trajectory point, if it exceeds the route range, it means that the currently selected vehicle trajectory has been executed and the process ends; after moving backward a preset step length based on the current vehicle trajectory point, if it does not exceed the route range, it means that the next vehicle trajectory point needs to be selected, and return to the above step S401 to continue to loop and execute the above steps.

[0110] In some embodiments, after all lane centerlines are constructed, the lane centerline center can be smoothed. The centerline of each lane point pair in the lane tracker can be determined, and these centerlines can be fitted to perform smoothing to obtain a lane centerline with a smooth trajectory.

[0111] The above lane centerline construction method can not only realize automatic construction of lane centerlines and improve the efficiency and accuracy of lane centerline construction, but also avoid the problem of inaccurate lane centerlines due to missing or offset of constructed lane boundaries due to occlusion or unclear recognition, thereby realizing lane centerline construction efficiently and accurately.

[0112] The embodiment of the present application also provides a method for rapid self-map construction on the vehicle side, which does not require human intervention or data upload and download. After the vehicle completes the driving route, the self-built map corresponding to the route will be saved on the local disk. The saved map can be used for automatic driving on the route that the vehicle has traveled. The solution can generate a map after one learning, and the map construction efficiency can reach the minute level. In addition, a risk layer is constructed at the same time to provide a basis for the upgrading and downgrading of automatic driving, so that the map can be used once it is built.

[0113] The lane centerline construction method provided in the embodiment of the present application can be applied to the vehicle-side rapid self-map construction method. The vehicle-side rapid self-map construction method will be described in detail below.

[0114] In the embodiment of the present application, the above-mentioned vehicle-side fast self-map construction method can construct a memory route map of the vehicle's traveled route on the vehicle side, and the map includes multiple layers of maps. For example, Figure 5 A schematic diagram of multiple layers in a memory route map. Figure 5 As shown, the memory route map may include: vector map, laser point cloud map, risk layer map. The above vector map (i.e., vector layer) includes: boundary (lane line and road boundary), lane (lane centerline), road (road), road_marker (ground marker), traffic_obj (traffic light, pole sign), etc. The vector layer is mainly used for vector high-precision positioning, prior perception, road structure topology reasoning, decision planning, etc. The laser point cloud map (i.e., laser point cloud layer) is mainly used for high-precision positioning and loop detection, etc. The risk layer map is used to mark the risk area in the memory route and is used for processing such as active degradation of the autonomous driving system.

[0115] In the embodiments of the present application, for example, Figure 6 The following is a flow chart of the above-mentioned vehicle-side rapid self-map building method, which includes the following steps in order of processing:

[0116] S601. Acquire perception information.

[0117] The perception information may include visual perception information and laser perception information. The visual perception information may include visual bird's eye view (BEV) perception information.

[0118] When acquiring perception information, environmental information can be identified and perceived, and a single-frame perception result can be output for subsequent mapping. The above-mentioned perception information may include visual perception information and laser perception information. When acquiring visual perception information, a BEV model may be used, and the obtained visual perception information may include single-frame lane lines, ground markings, traffic lights, pole signs, etc. Among them, ground markings may include stop lines, zebra crossings, road arrows, etc. The above-mentioned laser perception information includes laser point cloud information and dynamic obstacle information, etc.

[0119] S602: trajectory optimization.

[0120] The vehicle trajectory information can be used to convert the single-frame perception results from the vehicle body coordinate system into a global coordinate system for feature reconstruction.

[0121] The vehicle-side trajectory information mainly includes: the absolute position of the vehicle obtained by GPS signals, that is, the GPS trajectory, and the dead reckoning (ODOM) result obtained by inertial navigation and wheel speed sensors, that is, the dead reckoning trajectory.

[0122] Among them, although the GPS trajectory can give the absolute position of the vehicle more accurately, there will be problems such as position jitter and jumps in places with poor signals; the dead reckoning trajectory can obtain a continuous and smooth trajectory, and the relative position between frames has good accuracy, but due to the existence of sensor errors, cumulative errors will occur.

[0123] In some embodiments, the GPS trajectory and the dead reckoning trajectory may be fused to obtain an absolutely accurate and smooth trajectory to obtain an optimized trajectory.

[0124] In some embodiments, in order to ensure the efficiency of mapping, a real-time segmented trajectory optimization method is used to optimize the GPS trajectory and the dead reckoning trajectory. The trajectory optimization method may include but is not limited to the following process:

[0125] (1) Select the trajectory optimization area.

[0126] The trajectory optimization area includes the last optimized area and the unoptimized area. The trajectory position of the last optimized area remains fixed. The unoptimized area requires a continuous GPS signal stability at the end.

[0127] In some embodiments, when selecting a trajectory optimization area, when a continuous GPS signal stability is detected in the current GPS trajectory, the current optimization trajectory is selected, and the end of the selected current optimization trajectory includes the continuous GPS signal stable trajectory.

[0128] In some embodiments, whether there is a continuous period of GPS signal stability can be determined by whether the GPS signal confidence of multiple consecutive GPS track points is greater than a preset value. Exemplarily, the GPS signal confidence can be determined based on the GPS signal quality, and the GPS signal confidence is positively correlated with the GPS signal quality.

[0129] In some embodiments, the number of the above-mentioned multiple GPS track points is greater than a preset number.

[0130] In some embodiments, the length of the track composed of the above-mentioned multiple GPS track points is greater than or equal to a preset length.

[0131] For example, Figure 7 The figure shows a schematic diagram of segmented trajectory optimization. Figure 7 It can be seen that the current optimized trajectory includes the optimized overlapping area trajectory, that is, the trajectory of the overlapping area between the optimized trajectory and the current optimized trajectory.

[0132] (2) Optimize the trajectory within the trajectory optimization area.

[0133] In some embodiments, based on the dead reckoning trajectory, the first GPS trajectory in the current trajectory optimization area is trajectory optimized to determine an optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area.

[0134] In some embodiments, if it is detected that the GPS signal confidence of multiple consecutive GPS trajectory points is greater than a preset value, the first GPS trajectory in the trajectory optimization area and the dead reckoning trajectory are obtained; then the first GPS trajectory in the trajectory optimization area is trajectory optimized based on the dead reckoning trajectory to determine the optimized trajectory.

[0135] Among them, the multiple GPS track points are at the end of the first GPS track.

[0136] For example, Figure 8 The figure shows a flow chart of a trajectory optimization process, which may include but is not limited to the following steps:

[0137] S801. Determine an error parameter between a first GPS trajectory and a dead reckoning trajectory.

[0138] Among them, the first GPS track and the dead reckoning track are tracks within the selected range.

[0139] It should be noted that, in the process of performing trajectory optimization on the first GPS trajectory and the dead reckoning trajectory, the first GPS trajectory and the dead reckoning trajectory may be aligned according to timestamps, and the trajectory optimization process may be performed.

[0140] The error parameters include at least one of the following:

[0141] The relative distance error between frames, the error mean value obtained based on the statistics of the relative distance error between frames, and the covariance corresponding to the relative distance error between frames.

[0142] S802: Eliminate outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory.

[0143] The first GPS track is analyzed and screened according to the above error parameters, and abnormal points or points that do not conform to the normal pattern (for example, points that deviate far from the track) are identified as outliers and removed, thereby obtaining a second GPS track after outlier removal. This process helps to improve the accuracy and reliability of the GPS track.

[0144] S803: Match the dead reckoning trajectory to the global coordinate system to obtain a global trajectory.

[0145] By matching the dead reckoning trajectory to the global coordinate system, the trajectory information in the global scope can be obtained.

[0146] S804: Perform trajectory optimization based on the second GPS trajectory and the global trajectory to determine an optimized trajectory.

[0147] In some embodiments, an absolute constraint condition may be determined based on the second GPS trajectory, and an inter-frame relative constraint condition may be determined based on the global trajectory; then, based on the absolute constraint condition and the inter-frame relative constraint condition, the second GPS trajectory may be optimized to obtain an optimized trajectory.

[0148] Exemplarily, when performing trajectory optimization, Ceres is used to perform trajectory optimization on the second GPS trajectory based on the absolute constraint condition and the relative constraint condition between frames to obtain an optimized trajectory.

[0149] Among them, Ceres is a nonlinear optimization library. When performing trajectory optimization, you can use the Ceres optimization library to minimize errors or improve the accuracy of the trajectory. Specifically, Ceres can adjust the parameters in the trajectory, such as position information, to make the trajectory better meet the given constraints and goals. By using Ceres for trajectory optimization, you can improve the quality and accuracy of the trajectory to better meet application needs.

[0150] In some embodiments, after removing outliers from the first GPS trajectory based on the error parameter to obtain the second GPS trajectory, the covariance corresponding to the second GPS trajectory may be calculated, and the covariance is used to characterize the confidence of the absolute constraint condition.

[0151] The above-mentioned trajectory optimization of the second GPS trajectory based on the absolute constraint condition and the relative constraint condition between frames to obtain the optimized trajectory may include but is not limited to: based on the covariance corresponding to the second GPS trajectory, the absolute constraint condition, and the relative constraint condition between frames, the trajectory optimization of the second GPS trajectory to obtain the optimized trajectory.

[0152] In the above embodiment, by calculating the covariance corresponding to the second GPS trajectory, the uncertainty of the trajectory can be understood more accurately, so that the error factor can be better considered in trajectory optimization, and the accuracy of the optimized trajectory can be further improved. Considering the covariance helps to more comprehensively evaluate the confidence of the absolute constraints, making the optimization process more adaptable and robust, and improving the reliability of the results. Combining covariance, absolute constraints and relative constraints between frames for trajectory optimization can make fuller use of various information and obtain a better optimized trajectory.

[0153] (3) Maintain the cache and discard optimized trajectories that are far away from the ego vehicle.

[0154] Through the above trajectory optimization process, the GPS trajectory and the dead reckoning trajectory can be fused to obtain an absolutely accurate and smooth trajectory to obtain an optimized trajectory.

[0155] S603: Reconstructing map elements.

[0156] Vector map elements include lane lines, ground markings (stop lines, zebra crossings, road arrows), traffic lights, and pole signs. The basic principle of reconstruction is to project the single-frame perception results in the vehicle coordinate system into the global coordinate system through the optimized trajectory, and then match and fuse them. For lane lines, the fusion method is to sample the single-frame lane lines, then fit them using the Bezier curve fitting algorithm, and then sample them again after fitting to obtain discrete points; ground markings, traffic lights, and pole signs are weighted by multi-frame voting and observation distance to obtain fusion results.

[0157] S604, vectorization and topology construction.

[0158] The process in the vectorization and topology construction phase mainly includes the following steps:

[0159] S604a: map element preprocessing.

[0160] In order to improve the efficiency of vehicle-side mapping, this application proposes the following method, taking into account the "linear connection" characteristics of the memory route: Figure 3 The trajectory-based linear natural coordinate system shown can project all map elements onto the coordinate axis during the vectorization and topology construction stages and then be processed.

[0161] All subsequent processing is based on this linear natural coordinate system. On the one hand, the one-dimensional coordinate system can fully reflect the spatial relative position of map elements, which significantly improves the efficiency of map element query operations; on the other hand, when querying the coordinate transformation matrix (for example, when the elements need to be transformed), it is more accurate than querying the corresponding information through timestamps.

[0162] S604b, construction of basic road structure.

[0163] Considering the characteristic of "linear connection" of memory route, for example, Fig. 9 A linearly connected road structure model in one embodiment of the present application, such as Fig. 9 As shown in the figure, the basic road structure elements included in the road structure model are boundary (lane line and road boundary), lane (lane center line), and road (road); the topological connection of road only has predecessor and successor connections, and the predecessor and successor connections are one-to-one connection relationships, such as road 0, road 1 to road shown in the figure; the road contains multiple parallel lanes, and the topological connection of lanes has predecessor and successor connections and left and right successor connections, and the predecessor and successor have a one-to-many connection relationship.

[0164] The above basic road structure construction may include: boundary construction, lane construction and road construction.

[0165] Boundary construction:

[0166] The boundary construction process includes geometric verification, geometric repair, attribute aggregation, lane change point recognition, etc. Geometric verification verifies the smoothness, continuity, and semantic attributes of the boundary points to select the boundary set that meets the road structure characteristics; geometric repair aims to repair the boundary that is disconnected due to occlusion through geometric and semantic features; attribute aggregation can correct the location of the boundary attribute change points (virtuality, color, etc.) as much as possible through voting; lane change point recognition determines whether a lane change occurs during the mapping process through the location of the points under the boundary vehicle system, and identifies the lane change points to provide a reference for subsequent vector operations.

[0167] Lane construction:

[0168] The lane construction of the present application can be implemented with reference to the lane centerline construction method provided in the above embodiment. This method can quickly and efficiently construct and generate all lane centerlines on the memory route. This construction method has wide adaptability and can adapt to lane centerline construction in various scenarios.

[0169] Road build:

[0170] The memory route is composed of multiple sections of roads connected to each other. In this application, the road is interrupted to generate a new road only when the number of lanes changes. The shape points of the road are segmented vehicle trajectories. After the interruption, the topological relationship of the lanes in the road will be generated, and the topological relationship of the lanes between the connected roads will be generated.

[0171] S604c, construction of complex road structure.

[0172] Complex road structure construction includes but is not limited to construction of intersections, merge and split lanes.

[0173] In some embodiments, before constructing a complex road, topological binding of map elements may be performed first, for example, information such as stop lines and zebra crossings may be bound to corresponding lanes to facilitate subsequent intersection construction.

[0174] The intersection construction process may include but is not limited to the following steps:

[0175] Step a: Construct the intersection through the stop line, zebra crossing, traffic light and vehicle trajectory (turn and U-turn) information.

[0176] Step b, bind the intersection road, which is divided into entrance intersection road, intersection inside road, and exit intersection road.

[0177] Step c: Identify the turning attribute of the lane in the road entering the intersection through the vehicle trajectory and the road arrow.

[0178] Step d: For a single intersection, according to the turning attribute of the lane entering the intersection, a virtual lane and a lane centerline are generated by extending the trajectory of the own vehicle and the trajectory of other vehicles.

[0179] The process of constructing a merge lane may include but is not limited to the following steps:

[0180] Step A: Merge / exit scene recognition: The merge / exit scene is determined by the gradual reduction / increase of the width of the drivable lane, the lane line type, intersection restrictions and other conditions.

[0181] Step B, calculation of separation / merge point: the separation / merge point is identified by the lateral offset change rate of the merging / exiting lane shape point and the merging / exiting lane width change point, so that the merging / exiting lane can be accurately and smoothly generated.

[0182] Step C, merging in / out lane construction: refer to the shape points of the left and right lanes and the offset of the lateral position of the end point, linearly interpolate the lateral offset, and generate the merging in / out lane.

[0183] S605: Laser mapping.

[0184] In the above laser mapping process, a laser point cloud map may be constructed through a laser point cloud, and the laser point cloud map is mainly used for high-precision positioning and loop detection.

[0185] S606, map quality inspection.

[0186] Map quality inspection can automatically identify risky areas in the map of memorized routes and generate risk layer maps for processing such as active degradation of the autonomous driving system, thereby ensuring the integrity and robustness of the system.

[0187] For example, areas where lane lines are not smooth, where lane lines are interrupted, or where other information is missing in the map can be considered as risky areas in the map.

[0188] The vehicle-side rapid self-map building method provided in the embodiment of the present application does not require human intervention or data upload or download. After the vehicle completes the driving route, the self-built map corresponding to the route will be saved on the local disk. The saved map can be used for automatic driving on the route that the vehicle has traveled. The solution can generate a map after one learning, and the map building efficiency can reach the minute level. In addition, a risk layer is constructed at the same time to provide a basis for the upgrading and downgrading of automatic driving, so that the map can be used once it is built.

[0189] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0190] Based on the same inventive concept, the embodiment of the present application also provides a lane centerline device for implementing the lane centerline construction method involved above. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above method, so the specific limitations in one or more lane centerline device embodiments provided below can refer to the limitations of the lane centerline construction method above, and will not be repeated here.

[0191] In an exemplary embodiment, Fig.10 As shown, a mapping device is provided, comprising:

[0192] The trajectory optimization module 1001 is used to optimize the first GPS trajectory in the current trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area;

[0193] An element construction module 1002 is used to construct vector map elements based on the perception information and the optimized trajectory;

[0194] The road structure construction module 1003 is used to construct a road structure based on the constructed vector map elements to obtain a target map.

[0195] In some embodiments, the constructed vector map elements include constructed lane boundaries, and the road structure construction module 1003 is specifically used to:

[0196] Projecting the vehicle trajectory points onto the constructed lane boundaries to determine a plurality of projection points;

[0197] Perform lane tracking and matching based on the multiple projection points to determine a target lane;

[0198] A lane centerline of the target lane is constructed, and a road structure is constructed based on the lane centerline of the target lane to obtain the target map.

[0199] In some embodiments, the road structure construction module 1003 is specifically used to:

[0200] Perform pairing and combining based on the multiple projection points to construct at least one set of lane point pairs;

[0201] Lane tracking matching is performed based on the at least one set of lane point pairs to determine the target lane.

[0202] In some embodiments, the road structure construction module 1003 is specifically used to:

[0203] If the target group lane point pair matches the tracking lane, updating the target group lane point pair to the tracking lane to determine the target lane, wherein the target group lane point pair is any one group of lane point pairs in the at least one group of lane point pairs;

[0204] If the target lane point pair does not match the tracking lane, a new lane is constructed based on the target group lane point pair to determine the target lane.

[0205] In some embodiments, the road structure construction module 1003 is specifically used to:

[0206] The center point of the line connecting the lane point pairs of the target group is determined as a point on the lane center line.

[0207] In some embodiments, the road structure construction module 1003 is specifically used to:

[0208] Determine a plurality of projection vertical lines passing through the vehicle trajectory point and perpendicular to the lane boundary;

[0209] The intersection points of the multiple projection vertical lines and the lane boundary are determined as the multiple projection points.

[0210] In some embodiments, the method is applied to a linear natural coordinate system, the origin of which is the natural number 0, and the coordinate axis is the mileage of the vehicle;

[0211] The vehicle trajectory point is a trajectory point in a linear natural trajectory, and the linear natural trajectory is a trajectory of the vehicle's driving trajectory in the linear natural coordinate system.

[0212] In some embodiments, the trajectory optimization module 1001 is further configured to perform trajectory optimization on the first GPS trajectory in the trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory. If it is detected that the GPS signal confidence of a plurality of consecutive GPS trajectory points is greater than a preset value, then the first GPS trajectory and the dead reckoning trajectory in the trajectory optimization area are acquired;

[0213] Wherein, the multiple GPS track points are at the end of the first GPS track.

[0214] In some embodiments, the trajectory optimization module 1001 is specifically used to:

[0215] determining an error parameter between the first GPS trajectory and the dead reckoning trajectory;

[0216] Eliminate outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory;

[0217] Matching the dead reckoning trajectory to a global coordinate system to obtain a global trajectory;

[0218] Trajectory optimization is performed based on the second GPS trajectory and the global trajectory to determine the optimized trajectory.

[0219] In some embodiments, the trajectory optimization module 1001 is specifically used to:

[0220] determining an absolute constraint condition based on the second GPS trajectory, and determining an inter-frame relative constraint condition based on the global trajectory;

[0221] Based on the absolute constraint condition and the inter-frame relative constraint condition, the second GPS trajectory is optimized to obtain the optimized trajectory.

[0222] In some embodiments, the trajectory optimization module 1001 is further configured to: after removing outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory,

[0223] Calculating a covariance corresponding to the second GPS trajectory, where the covariance is used to characterize the confidence of the absolute constraint condition;

[0224] The step of optimizing the second GPS trajectory based on the absolute constraint condition and the relative constraint condition between frames to obtain an optimized trajectory includes:

[0225] Based on the covariance corresponding to the second GPS trajectory, the absolute constraint condition, and the inter-frame relative constraint condition, trajectory optimization is performed on the second GPS trajectory to obtain the optimized trajectory.

[0226] Each module in the above-mentioned mapping device can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0227] In an exemplary embodiment, a computer device is provided. The computer device may be a vehicle-mounted terminal, and its internal structure diagram may be as shown in FIG. Fig.11 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner. When the computer program is executed by the processor, a mapping method is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a real imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen.

[0228] Those skilled in the art will understand that Fig.11 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0229] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, each process involved in the above method embodiment is implemented.

[0230] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0231] Projecting the vehicle trajectory points onto the constructed lane boundaries to determine a plurality of projection points;

[0232] Perform lane tracking and matching based on the multiple projection points to determine a target lane;

[0233] A lane centerline of the target lane is constructed.

[0234] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0235] The performing lane tracking and matching based on the multiple projection points to determine the target lane includes:

[0236] Perform pairing and combining based on the multiple projection points to construct at least one set of lane point pairs;

[0237] Lane tracking matching is performed based on the at least one set of lane point pairs to determine the target lane.

[0238] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0239] The performing lane tracking and matching based on the at least one set of lane point pairs to determine the target lane includes:

[0240] If the target group lane point pair matches the tracking lane, updating the target group lane point pair to the tracking lane to determine the target lane, wherein the target group lane point pair is any one group of lane point pairs in the at least one group of lane point pairs;

[0241] If the target lane point pair does not match the tracking lane, a new lane is constructed based on the target group lane point pair to determine the target lane.

[0242] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0243] The constructing the lane centerline of the target lane includes:

[0244] The center point of the line connecting the lane point pairs of the target group is determined as a point on the lane center line.

[0245] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0246] The vehicle trajectory point is projected onto the lane boundary to determine a plurality of projection points, including:

[0247] Determine a plurality of projection vertical lines passing through the vehicle trajectory point and perpendicular to the lane boundary;

[0248] An intersection point between the multiple projection vertical lines and the lane boundary is determined as the at least one projection point.

[0249] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0250] The method is applied in a linear natural coordinate system, the origin of which is the natural number 0, and the coordinate axis is the mileage of the vehicle;

[0251] The vehicle trajectory point is a trajectory point in a linear natural trajectory, and the linear natural trajectory is a trajectory of the vehicle's driving trajectory in the linear natural coordinate system.

[0252] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the lane centerline construction method described in the above method embodiment is implemented.

[0253] In one embodiment, a computer program product is provided, including a computer program, which, when executed by a processor, implements the lane centerline construction method described in the above method embodiment.

[0254] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0255] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0256] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A mapping method, characterized in that: include: Based on the dead reckoning trajectory, optimizing the first GPS trajectory in the current trajectory optimization area to determine an optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area; constructing vector map elements based on the perception information and the optimized trajectory; Constructing a road structure based on the constructed vector map elements to obtain a target map; The method of optimizing the first GPS trajectory in the trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory includes: determining an error parameter between the first GPS trajectory and the dead reckoning trajectory; removing outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory; matching the dead reckoning trajectory to a global coordinate system to obtain a global trajectory; and optimizing the trajectory based on the second GPS trajectory and the global trajectory to determine the optimized trajectory.

2. The method according to claim 1, characterized in that The constructed vector map elements include constructed lane boundaries, and constructing a road structure based on the constructed vector map elements to obtain a target map includes: Projecting the vehicle trajectory points onto the constructed lane boundaries to determine a plurality of projection points; Perform lane tracking and matching based on the multiple projection points to determine a target lane; A lane centerline of the target lane is constructed, and a road structure is constructed based on the lane centerline of the target lane to obtain the target map.

3. The method according to claim 2, characterized in that The performing lane tracking and matching based on the multiple projection points to determine the target lane includes: Perform pairing and combining based on the multiple projection points to construct at least one set of lane point pairs; Lane tracking matching is performed based on the at least one set of lane point pairs to determine the target lane.

4. The method according to claim 3, characterized in that The performing lane tracking and matching based on the at least one set of lane point pairs to determine the target lane includes: If the target group lane point pair matches the tracking lane, updating the target group lane point pair to the tracking lane to determine the target lane, wherein the target group lane point pair is any one group of lane point pairs in the at least one group of lane point pairs; If the target group lane point pair does not match the tracking lane, a new lane is constructed based on the target group lane point pair to determine the target lane.

5. The method according to claim 4, characterized in that The constructing the lane centerline of the target lane includes: The center point of the line connecting the lane point pairs of the target group is determined as a point on the lane center line.

6. The method according to claim 2, characterized in that The vehicle trajectory point is projected onto the lane boundary to determine a plurality of projection points, including: Determine a plurality of projection vertical lines passing through the vehicle trajectory point and perpendicular to the lane boundary; The intersection points of the multiple projection vertical lines and the lane boundary are determined as the multiple projection points.

7. The method according to any one of claims 2 to 6, characterized in that: The method is applied in a linear natural coordinate system, the origin of which is the natural number 0, and the coordinate axis is the mileage of the vehicle; The vehicle trajectory point is a trajectory point in a linear natural trajectory, and the linear natural trajectory is a trajectory of the vehicle's driving trajectory in the linear natural coordinate system.

8. The method according to claim 1, characterized in that Before performing trajectory optimization on the first GPS trajectory in the trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory, the method further includes: If it is detected that the GPS signal confidence of a plurality of consecutive GPS trajectory points is greater than a preset value, obtaining a first GPS trajectory and the dead reckoning trajectory within the trajectory optimization area; Wherein, the multiple GPS track points are at the end of the first GPS track.

9. The method according to claim 1, characterized in that: The performing trajectory optimization based on the second GPS trajectory and the global trajectory to determine the optimized trajectory includes: determining an absolute constraint condition based on the second GPS trajectory, and determining an inter-frame relative constraint condition based on the global trajectory; Based on the absolute constraint condition and the inter-frame relative constraint condition, the second GPS trajectory is optimized to obtain the optimized trajectory.

10. The method according to claim 9, characterized in that After removing outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory, the method further includes: Calculating a covariance corresponding to the second GPS trajectory, where the covariance is used to characterize the confidence of the absolute constraint condition; The step of optimizing the second GPS trajectory based on the absolute constraint condition and the relative constraint condition between frames to obtain an optimized trajectory includes: Based on the covariance corresponding to the second GPS trajectory, the absolute constraint condition, and the inter-frame relative constraint condition, trajectory optimization is performed on the second GPS trajectory to obtain the optimized trajectory.

11. A mapping device, characterized in that: The device comprises: A trajectory optimization module, for performing trajectory optimization on a first GPS trajectory in a current trajectory optimization area based on a dead reckoning trajectory to determine an optimized trajectory, wherein the trajectory optimization area includes: a partial area optimized last time and an unoptimized area; An element construction module, used for constructing vector map elements based on the perception information and the optimized trajectory; A road structure construction module is used to construct a road structure based on the constructed vector map elements to obtain a target map; The method of optimizing the first GPS trajectory in the trajectory optimization area based on the dead reckoning trajectory to determine the optimized trajectory includes: determining an error parameter between the first GPS trajectory and the dead reckoning trajectory; removing outliers from the first GPS trajectory based on the error parameter to obtain a second GPS trajectory; matching the dead reckoning trajectory to a global coordinate system to obtain a global trajectory; and optimizing the trajectory based on the second GPS trajectory and the global trajectory to determine the optimized trajectory.

12. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 10 are implemented.

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

14. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

Images