Method, device, equipment, and medium for obtaining lane topology based on crowdsourcing trajectory

Abstract

The present invention provides a method, apparatus, device, and medium for obtaining lane topology based on crowdsourced trajectories. The method comprises the following steps: obtaining trajectory point information collected by crowdsourced vehicles; identifying intersections and non-intersections based on the trajectory point information; grouping a number of trajectory points in the intersection and non-intersection areas according to their timestamps and distance intervals to obtain trajectory groups for the intersection and non-intersection areas; clustering the trajectory groups for the intersection to generate virtual lane lines for the intersection; clustering the trajectory groups for the non-intersection to generate lane topology for the non-intersection; and merging the virtual lane lines for the intersection and the lane topology for the non-intersection to generate a final lane topology map. The present invention utilizes crowdsourced trajectory data and employs segmentation, clustering, and splicing methods to generate lane lines with less redundancy and closer to reality.

Description

Technical Field

[0001] The present application relates to the technical field of crowdsourced map data processing, and in particular to a method, apparatus, device, and medium for obtaining lane line topology based on crowdsourced trajectories. Background Art

[0002] In the field of autonomous driving, lane-level high-precision maps are crucial for functions such as path calculation, navigation, and positioning. Although professionally collected high-precision maps are highly accurate, due to production cost and efficiency issues, the update frequency often cannot keep up with road changes.

[0003] With the rapid development of sensor technology, wireless communications, and network technologies, people's travel generates a massive amount of spatiotemporal trajectory data, containing a wealth of detailed road information and human behavior information. Trajectory data collection has gradually evolved from specialized measurement vehicles or professional personnel to voluntary recording of travel trajectories by non-professionals, and data collection has begun to shift to a crowdsourcing model. Crowdsourcing vehicle trajectory data (crowdsourced big data) is undoubtedly the best data source for lane-level road information extraction. However, the lane lines currently generated using crowdsourced trajectories suffer from high redundancy and differ from the actual road surface. Summary of the Invention

[0004] In view of the above-mentioned shortcomings of the prior art, the present invention provides a method, device, equipment, and medium for obtaining lane topology based on crowdsourcing trajectories to solve the above-mentioned technical problems.

[0005] The method for obtaining lane topology based on crowdsourcing trajectories provided by the present invention includes:

[0006] Obtaining track point information collected by crowdsourcing vehicles, wherein the track point information includes timestamps and location information of several track points;

[0007] identifying intersection areas and non-intersection areas based on the trajectory point information;

[0008] Grouping the plurality of trajectory points in the intersection area and the non-intersection area according to the timestamps of the trajectory points and the distance intervals of the trajectory points, respectively, to obtain a trajectory group in the intersection area and a trajectory group in the non-intersection area;

[0009] performing distance clustering on the trajectory group in the intersection area to generate a virtual lane line in the intersection area;

[0010] performing distance clustering on the trajectory group in the non-intersection area to generate a lane line topology in the non-intersection area;

[0011] The virtual lane lines in the intersection area and the lane line topology in the non-intersection area are merged to generate a final lane line topology map.

[0012] In one embodiment of the present invention, identifying intersection areas and non-intersection areas based on information of the trajectory points includes:

[0013] Calculating the angular velocity of the trajectory point according to the timestamp and direction angle of the trajectory point;

[0014] Taking the trajectory point with an angular velocity greater than a first threshold as a candidate turning point;

[0015] Merging the consecutive turning candidate points into a group, and if the number of non-turning candidate points between the two groups is less than a second threshold, further merging the two groups of non-turning candidate points, and marking the non-turning candidate points between the two groups as turning candidate points;

[0016] Calculate the direction angle difference and timestamp difference between the start point and the end point in each group of candidate turning points, filter out the groups with differences less than the third threshold, and obtain the turning point combination;

[0017] MeanShift clustering is performed on the turning points in the turning point combination, and the intersection area is identified according to the speed of the trajectory points and / or the proportion of turning points.

[0018] In one embodiment of the present invention, mean-shift clustering is performed on the turning points in the turning point combination, and the intersection area is identified based on the speed of the trajectory points and / or the proportion of turning points, including: mean-shift clustering is performed on the turning points in the turning point combination, and the cluster center point is the intersection center point; with the intersection center point as the center of the circle, multiple circular rings are constructed according to preset radius and interval; the average speed of the trajectory points within the circular range and the ratio of the number of the turning points to the trajectory points are calculated from small to large; if the average speed is greater than the fourth threshold or the proportion of the turning points is lower than the fifth threshold or the radius of the circular ring exceeds the sixth threshold, the calculation is stopped, and at this time, the circular area between the intersection center point and the circular ring is the intersection area.

[0019] In one embodiment of the present invention, a number of trajectory points in the intersection area and the non-intersection area are grouped according to the timestamps of the trajectory points and the distance intervals of the trajectory points, respectively, including: traversing the trajectory points in the order of the timestamps to determine the interruption points; grouping the trajectory points according to the areas where the trajectory points are located and the interruption points to obtain the trajectory group of the intersection area and the trajectory group of the non-intersection area.

[0020] In one embodiment of the present invention, the interruption point satisfies at least one of the following conditions: a trajectory point at the boundary between the intersection area and the non-intersection area; a timestamp interval between adjacent trajectory points exceeds a preset threshold; a distance interval between adjacent trajectory points exceeds a preset threshold.

[0021] In one embodiment of the present invention, distance clustering is performed on the trajectory group in the intersection area to generate a virtual lane line in the intersection area, including:

[0022] Calculate the Hausdorff distance of each two trajectory groups in the intersection area. If the distances between the first and last parts are both less than the lane width threshold, the smaller distance is used as the Hausdorff distance of the two trajectory groups. Otherwise, the distance between the middle parts of the two trajectory groups is used as the Hausdorff distance.

[0023] Clustering the trajectory groups in the intersection area according to the Hausdorff distance;

[0024] From each clustered trajectory cluster, a relatively smooth trajectory group is selected as a representative trajectory to generate a virtual lane line. The smoothness is calculated by averaging the azimuth angle differences between adjacent trajectory points, with smaller values representing smoother paths.

[0025] In one embodiment of the present invention, clustering the trajectory groups in the non-intersection area to generate a lane centerline topology in the non-intersection area includes:

[0026] Clustering the trajectory groups in the non-intersection area according to travel direction and distance to obtain road segments;

[0027] Taking the road segment as a unit, performing distance-based clustering on the trajectory groups in each road segment to generate a lane line topology in a non-intersection area.

[0028] In one embodiment of the present invention, clustering the trajectory groups in the non-intersection area according to the travel direction and distance includes:

[0029] Calculate the distance between the starting points and the end points of each two trajectory groups, take the smaller value of the two distances as the distance between the trajectory groups, and then perform preliminary clustering on each trajectory group based on the distance;

[0030] The Hausdorff distance between the road segments after the preliminary clustering is calculated. If the Hausdorff distance is less than the lane width, the two road segments are merged. Otherwise, the minimum distance between the trajectory groups in the road segment is calculated. If the minimum distance is less than the lane width, the two road segments are merged.

[0031] In one embodiment of the present invention, if the road segment is longer than 100 meters, the preliminary clustering further includes cutting the road segment into small segments, and cutting the road segment into small segments includes:

[0032] In the longest trajectory group in the road segment, find the trajectory points at the separation position in the longest trajectory group according to the preset interval and use them as cutting points; traverse the trajectory points in the trajectory group in the road segment and segment the trajectory group according to the distance between the trajectory point and the cutting point.

[0033] In one embodiment of the present invention, taking the road segments as units, performing distance-based clustering on the trajectory groups in each of the road segments to generate a lane line topology in the non-intersection area includes:

[0034] Performing clustering based on Hausdorff distance on the trajectory groups in each road segment, and selecting a trajectory group close to the center of each cluster from the multiple trajectory groups in each cluster as a candidate lane line;

[0035] A graph is generated using the candidate lane lines as points and the Hausdorff distances between the candidate lane lines as edges, and a maximum clique algorithm is used to find mutually non-intersecting lane lines as the main lane lines;

[0036] Based on the trajectory group data in each of the road segments and the main lane line, a diverging and merging lane line is generated.

[0037] In one embodiment of the present invention, generating a diverging and merging lane line based on the trajectory group data in each road segment and the main lane line includes:

[0038] In the road segment, at regular intervals, the number of lanes at the corresponding position is obtained by using a Gaussian mixture model for trajectory points at a cross section of the road segment;

[0039] Combining the number of lanes at each position and the information of the main lane line, identifying lanes other than the main lane line as candidate diverging and merging lanes;

[0040] Calculate the minimum distance between the starting point of each candidate diverging and merging lane and each main lane, the minimum distance between the end point and each main lane, and the number of main lanes intersecting with each candidate diverging and merging lane. Determine the lane change lane based on the above calculation results.

[0041] The intersection point of the candidate diverging and merging lane line and the main lane line after filtering out the lane change lane line is the diverging and merging point; if the distance between the starting point of the candidate diverging and merging lane line and the main lane line is not less than the lane width threshold, then the trajectory between the starting point of the candidate diverging and merging lane line and the diverging and merging point is the diverging and merging lane line; if the distance between the end point of the candidate diverging and merging lane line and the main lane is not less than the lane width threshold, then the trajectory between the end point of the candidate diverging and merging lane line and the diverging and merging point is the diverging and merging lane line.

[0042] In one embodiment of the present invention, the virtual lanes in the intersection area and the lane line topology of the non-intersection area are merged to generate a final lane line topology map, including:

[0043] Generate nodes at the start and end points of each lane line;

[0044] Interrupting the main lane line at the diverging and merging position to add a node point;

[0045] Merge the node points whose distance is close to the threshold;

[0046] If a node point only connects two lane lines, the node point is deleted and the lane lines are merged to generate the final lane line topology map.

[0047] The present invention provides an apparatus for obtaining lane topology based on crowdsourced trajectories, comprising: a crowdsourced data acquisition module, an information identification module, a grouping module, an intersection area processing module, a non-intersection area processing module, and a fusion module. The crowdsourced data acquisition module is configured to acquire trajectory point information collected by crowdsourced vehicles, the trajectory point information including timestamps and location information of multiple trajectory points; the information identification module is configured to identify intersection areas and non-intersection areas based on the trajectory point information; the grouping module is configured to group multiple trajectory points in the intersection area and non-intersection area according to the timestamps and distance intervals of the trajectory points, respectively, to obtain trajectory groups for the intersection area and trajectory groups for the non-intersection area; the intersection area processing module is configured to perform distance clustering on the trajectory groups in the intersection area to generate virtual lane lines for the intersection area; the non-intersection area processing module is configured to perform distance clustering on the trajectory groups in the non-intersection area to generate lane line topology for the non-intersection area; and the fusion module is configured to merge the virtual lanes in the intersection area with the lane line topology of the non-intersection area to generate a final lane line topology map.

[0048] The electronic device provided by the present invention includes one or more processors and a storage device, wherein the storage device is used to store one or more programs. When the one or more programs are executed by the one or more processors, the electronic device implements the method of obtaining lane line topology based on crowdsourcing trajectories described in the present invention.

[0049] The computer-readable storage medium provided by the present invention stores a computer program thereon. When the computer program is executed by a processor of a computer, the computer is caused to execute the method of obtaining lane topology based on crowdsourcing trajectories described in the present invention.

[0050] The present invention utilizes crowdsourced vehicle trajectory data as input, generating lane-level road network topology through segmentation, aggregation, and splicing. The resulting lane lines are less redundant and more closely resemble real-world lanes.

[0051] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0053] Figure 1 1 is a schematic diagram of an implementation environment for obtaining lane topology based on crowdsourced trajectories, as shown in an exemplary embodiment of the present application;

[0054] Figure 2 is a flowchart of a method for obtaining lane topology based on crowdsourced trajectories, shown in an exemplary embodiment of the present application;

[0055] Figure 3 is a flowchart illustrating an exemplary embodiment of the present application for identifying intersection areas and non-intersection areas based on the trajectory point information;

[0056] Figure 4 is a flowchart illustrating an exemplary embodiment of the present application for performing MeanShift clustering on the turning points in the turning point group and identifying the intersection area based on the speed of the trajectory point and / or the turning point;

[0057] Figure 5is a flow chart showing an exemplary embodiment of the present application, which shows how to group a plurality of trajectory points in the intersection area and the non-intersection area according to the timestamps of the trajectory points and the distance intervals of the trajectory points;

[0058] Figure 6 is a flowchart illustrating an exemplary embodiment of the present application for clustering trajectory groups in the intersection area to generate virtual lane lines in the intersection area;

[0059] Figure 7 is a schematic diagram of virtual lane lines in an intersection area shown in an exemplary embodiment of the present application;

[0060] Figure 8 is a flowchart illustrating an exemplary embodiment of the present application for clustering the trajectory group in the non-intersection area to generate a lane line topology in the non-intersection area;

[0061] Figure 9 is a flowchart illustrating clustering of trajectory groups in the non-intersection area according to travel direction and distance, as shown in an exemplary embodiment of the present application;

[0062] Figure 10 is a flow chart showing a method of cutting a long road segment according to an exemplary embodiment of the present application;

[0063] Figure 11 is a flowchart illustrating, in an exemplary embodiment of the present application, a method of clustering trajectory groups in each road segment based on distance, taking the road segment as a unit;

[0064] Figure 12 is a flowchart illustrating an exemplary embodiment of the present application for generating diverging and merging lane lines based on the trajectory group data in each of the road segments and the main lane line;

[0065] Figure 13 1 is a schematic structural diagram of a main lane line and a diverging lane line shown in an exemplary embodiment of the present application;

[0066] Figure 14 1 is a schematic structural diagram of a main lane line and a merging lane line shown in an exemplary embodiment of the present application;

[0067] Figure 15 This is a flowchart illustrating an exemplary embodiment of the present application for merging the virtual lane lines of the intersection area and the lane line topology of the non-intersection area to generate a final lane line topology map;

[0068] Figure 16 is a block diagram of an apparatus for obtaining lane topology based on crowdsourced trajectories, shown in an exemplary embodiment of the present application;

[0069] Figure 17 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application, shown as an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0070] The following describes the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art will readily appreciate the other advantages and benefits of the present invention from the disclosure herein. The present invention may also be implemented or applied through various other specific embodiments, and the various details in this specification may be modified or altered based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are intended only to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0071] It should be noted that the illustrations provided in the following embodiments are merely schematic illustrations of the basic concept of the present invention. Therefore, the illustrations only show components related to the present invention and are not drawn according to the number, shape, and size of components in actual implementation. In actual implementation, the type, quantity, and proportion of each component may be changed arbitrarily, and the component layout may also be more complex.

[0072] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the embodiments of the present invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0073] First of all, it should be explained that crowdsourcing trajectories are trajectory data collected by crowdsourcing vehicles. The trajectory data collected by crowdsourcing vehicles refers to the spatiotemporal location information fed back by crowdsourcing vehicles through smart terminal devices (in-vehicle navigation, smart phones, in-vehicle computers, tablets, laptops, etc.) in a crowdsourcing manner.

[0074] Lane line topology refers to the topological information that can reflect the lane lines on the road and the relative positions between lane lines.

[0075] A timestamp is data generated using digital signature technology. The signature includes information such as the original file, signature parameters, and the signing time. The main purpose of a timestamp is to authenticate the time when data was generated through certain technical means, thereby verifying whether the data has been tampered with since its generation.

[0076] MeanShift clustering, as the name suggests, consists of Mean and Shift. It is a sliding window-based algorithm that attempts to find areas with dense data points.

[0077] The drift vector of the center point is calculated by the change in data density in the region of interest, and the center point is moved for the next iteration until the density reaches the maximum (the center point remains unchanged). This operation can be performed for each data point, and the number of times the data appears in the region of interest is counted. This parameter will be used as the basis for classification in the end.

[0078] The Hausdorff distance is a measure of the similarity between two sets of points. It is a definition of the distance between two point sets: Assume there are two sets A = {a1…,ap), B = {b1…,bq}, then the Hausdorff distance between these two point sets is defined as h(A, B) = max a∈A{minb∈B{d(a,b)}}

[0079] Where a and b are points in sets A and B respectively, and d(a, b) represents the Euclidean distance between a and b. As shown in the above formula, the Hausdorff distance h(A, B) measures the maximum mismatch between two point sets. The smaller the distance, the higher the match.

[0080] The Gaussian mixture model uses the Gaussian probability density function (normal distribution curve) to accurately quantify things. It is a model that decomposes things into several Gaussian probability density functions (normal distribution curves).

[0081] Figure 1 FIG. 1 is a schematic diagram of an exemplary embodiment of the present application showing an implementation environment for obtaining lane topology based on crowdsourced trajectories. Figure 1 As shown, during the driving process, the vehicle uses the navigation map software installed on the smart terminal 110 to achieve navigation. The navigation map software will refresh the road conditions, that is, the navigation map software will make a network request to the navigation server 120 according to the domain name of the navigation server 120, and then the navigation server 120 will return the corresponding navigation path to the navigation map software. Figure 1The smart terminal 110 shown can be a terminal device that supports the installation of navigation map software, such as a smart phone, a car computer, or a tablet computer, but is not limited to this. The navigation server 120 is a navigation server, which can be, for example, an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms, and is not limited here. The smart terminal 110 can communicate with the navigation server 220 through wireless networks such as 3G (third generation mobile information technology), 4G (fourth generation mobile information technology), and 5G (fifth generation mobile information technology), and is not limited here.

[0082] In the field of autonomous driving, lane-level high-precision maps are often used. Although high-precision maps offer high accuracy, due to production cost and efficiency issues, their update frequency often cannot keep up with road changes. To address these issues, the present application provides a method for obtaining lane topology based on crowdsourced trajectories, an apparatus for obtaining lane topology based on crowdsourced trajectories, an electronic device, a computer-readable storage medium, and a computer program product. These embodiments are described in detail below.

[0083] See also Figure 2 , Figure 2 This is a flow chart of a method for obtaining lane topology based on crowdsourced trajectories, as shown in an exemplary embodiment of the present application. Figure 1 The implementation environment shown is specifically executed by the smart terminal 110 in the implementation environment. It should be understood that the method can also be applied to other exemplary implementation environments and specifically executed by devices in other implementation environments. This embodiment does not limit the implementation environment to which the method is applicable.

[0084] like Figure 2 As shown, in an exemplary embodiment, the method for obtaining lane topology based on crowdsourced trajectories includes at least steps S210 to S260, which are described in detail as follows:

[0085] Step S210: Obtain trajectory point information collected by crowdsourcing vehicles.

[0086] The navigation server 120 can obtain crowdsourced trajectory data for corresponding roads from crowdsourced vehicles. Crowdsourced vehicles collect trajectory data through on-board sensors while driving on the road. The crowdsourced trajectory data is obtained through a crowdsourcing model. On-board sensors can include Global Navigation Satellite Systems (GNSS), cameras, ultrasonic radars, lidars, or millimeter-wave radars. The corresponding crowdsourced trajectory data can include timestamps and location information of trajectory points, such as direction angles.

[0087] Step 220 : Identify intersection areas and non-intersection areas based on the trajectory point information.

[0088] Roads usually contain intersection sections and non-intersection sections, and the vehicle trajectories in intersection sections and non-intersection sections are different. Therefore, after obtaining the trajectory points collected by crowdsourcing vehicles, the intersection area and non-intersection area are first identified based on the trajectory point information for subsequent trajectory segmentation.

[0089] Figure 3 yes Figure 2 Step S220 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 3 As shown, identifying the intersection area and the non-intersection area based on the trajectory point information may include steps S310 to S350, which are described in detail as follows:

[0090] Step S310: Calculate the angular velocity of the trajectory point according to the timestamp and direction angle of the trajectory point.

[0091] The trajectory point information obtained in step S210 includes the timestamp and direction angle of each trajectory point. According to the timestamp and direction corresponding to each trajectory point, the angular velocity of each trajectory point is calculated according to the formula ω = Δθ / Δt, where ω is the angular velocity, Δt represents time, and Δθ represents the angle turned within Δt time.

[0092] Step S320: taking the trajectory point whose angular velocity is greater than the first threshold as a candidate turning point.

[0093] The angular velocity corresponding to each trajectory point calculated in step S310 is compared with a preset threshold, denoted as the first threshold. If the angular velocity corresponding to the trajectory point is greater than the first threshold, the trajectory point is marked as a turn candidate point; otherwise, it is marked as a non-turn candidate point. It should be noted that the selection of the first threshold is related to factors such as the timestamp interval of the trajectory points and is not limited here. It can be set according to actual needs. The same applies to the thresholds below.

[0094] Step S330: merge the consecutive turning candidate points into a group; if the number of non-turning candidate points between the two groups is less than a second threshold, further merge the two groups, and mark the non-turning candidate points between the two groups as turning candidate points.

[0095] By calculating the angular velocity of each trajectory point and comparing it with a preset threshold, several trajectory points are divided into turning candidate points and non-turning candidate points. If several consecutive trajectory points are all turning candidate points, they are merged into a group. This method is continued until all consecutive turning candidate points are merged. The turning candidate point groups are then compared pairwise. If the number of non-turning candidate points between two groups of turning candidate points is less than a preset threshold, here denoted as the second threshold, the two groups of turning candidate points are further merged, and the non-turning candidate points between the two groups are marked as turning candidate points.

[0096] Step S340: Calculate the direction angle difference and timestamp difference between the starting point and the end point in each group of candidate turning points, filter out the groups with differences less than a third threshold, and obtain a turning point group.

[0097] The azimuth angle difference and timestamp difference between the start and end points of each set of candidate turning points are calculated and compared with a preset threshold, referred to as a third threshold. The third threshold includes a azimuth angle difference threshold and a timestamp difference threshold. For example, the azimuth angle difference threshold is 50°, and the timestamp difference threshold is 3 seconds. If the azimuth angle difference between the start and end points of a candidate turning point group is less than the azimuth angle difference threshold and / or the timestamp difference between the start and end points is less than the timestamp difference threshold, the candidate turning point group is filtered out, and the remaining points are considered turning point groups.

[0098] Step S350: Perform mean shift clustering on the turning points in the turning point group, and identify the intersection area based on the speed of the trajectory point and / or the turning point.

[0099] Figure 4 yes Figure 3 Step S350 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 4 As shown, the process of performing MeanShift clustering on the turning points in the turning point group and identifying the intersection area based on the speed of the trajectory point and / or the turning point may include steps S410 to S430, which are described in detail as follows:

[0100] Step S410: MeanShift clustering is performed on the filtered turning points (with a bandwidth of 50 meters), and the cluster center is the intersection center.

[0101] Step S420: construct multiple circular rings with the center point of the intersection as the center and at preset radius and intervals;

[0102] The preset radius and interval can be set according to actual conditions. For example, starting with a radius of 30 meters and an interval of 5 meters between circles, multiple concentric circles can be constructed in order from small to large.

[0103] Step S430 , calculating the average speed of the trajectory points and the ratio of the number of the turning points to the number of the trajectory points within the circular range from small to large, and identifying the intersection area based on the calculation results.

[0104] Specifically, the annular range is calculated from small to large. For example, the average speed V of all trajectory points within the annular range between a circle with a radius of 30 meters and a circle with a radius of 35 meters is first calculated as (V1+V2+V3+…+Vn) / n, where Vn represents the speed of the nth trajectory point and n represents the number of trajectory points; and the ratio of the number of turning points to the total number of trajectory points within this annular range. Then, the average speed and the proportion of turning points of all trajectory points within the annular range between a circle with a radius of 35 meters and a circle with a radius of 40 meters are calculated in the same manner, and so on until the calculation stops. If the average speed V is greater than the preset threshold, it is recorded as the fourth threshold, or the ratio of the number of turning points to the number of trajectory points is lower than the preset threshold, it is recorded as the fifth threshold, or the radius of the ring exceeds the preset threshold, it is recorded as the sixth threshold and the calculation stops. At this time, the circular area between the center point of the intersection and the ring is the intersection area. The area outside the intersection area is the non-intersection area.

[0105] Please continue reading Figure 2 In step S230, the trajectory points of the intersection area and the non-intersection area are grouped according to the timestamps of the trajectory points and the distance intervals of the trajectory points to obtain a trajectory group of the intersection area and a trajectory group of the non-intersection area.

[0106] Figure 5 yes Figure 2 Step S230 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 5 As shown, the process of grouping the plurality of trajectory points in the intersection area and the non-intersection area according to the timestamps of the trajectory points and the distance intervals of the trajectory points may include steps S510 to S520, which are described in detail as follows:

[0107] Step S510: traverse the trajectory points in the order of timestamps and determine the interruption point.

[0108] The trajectory point information contains timestamps. The trajectory points are traversed in the order of timestamps and the interruption points are selected. The interruption points meet at least one of the following conditions: (1) the trajectory point is located at the boundary between the intersection area and the non-intersection area; (2) the time stamp interval between adjacent trajectory points exceeds the preset threshold; (3) the distance interval between adjacent trajectory points exceeds the preset threshold.

[0109] Step S520: grouping the track points according to the areas where the track points are located and the interruption points.

[0110] The area where the trajectory points are located includes the intersection area and the non-intersection area. According to the interruption point, several trajectory points in the intersection area are divided into trajectory groups of multiple intersection areas; and several trajectory points in the non-intersection area are divided into trajectory groups of multiple non-intersection areas.

[0111] Please continue reading Figure 2 Step S240: Cluster the trajectory groups in the intersection area to generate virtual lane lines in the intersection area. The goal of intersection area processing is to obtain virtual connected lanes, that is, the connection relationship between lanes on each road associated with the intersection.

[0112] See also Figure 6 , Figure 6 yes Figure 2 Step S240 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 6 As shown, the process of clustering the trajectory groups in the intersection area to generate virtual lane lines in the intersection area may include steps S610 to S630, which are described in detail as follows:

[0113] Step S610: Calculate the Hausdorff distance between every two trajectory groups in the intersection area.

[0114] The Hausdorff distance in this embodiment is an improved Hausdorff distance. Specifically, the Hausdorff distance between the beginning and end of the two trajectory groups is first calculated. If the distance between the beginning and end is less than the lane width threshold, the distance is used as the Hausdorff distance of the two trajectory groups; otherwise, the distance between the middle parts of the two trajectories is calculated according to the standard Hausdorff distance. This embodiment adopts the improved method because the standard Hausdorff distance is the distance between the middle parts of the two trajectory groups. If the beginning and end parts of the two trajectory groups are close, and the middle parts have a larger distance, the two trajectory groups will be divided into different clusters when clustered by distance. However, for the virtual lanes at the intersection, if the beginning and end parts of the two trajectories are close, it means that their starting lanes and ending lanes are the same and they should belong to the same virtual lane. Based on this, this embodiment improves the Hausdorff distance to prevent the above problems.

[0115] Step 620 : Clustering the trajectory groups in the intersection area according to the Hausdorff distance.

[0116] Distance clustering is performed according to the Hausdorff distance calculated in step S610.

[0117] Step S630 : Select a smoother trajectory group from each clustered trajectory cluster as a representative trajectory to generate a virtual lane line.

[0118] When selecting a representative track, the track's steering smoothness is used as the weight. The smoothness is calculated by averaging the angle differences between adjacent track points within the track group. The smaller the average value, the smoother the track. Figure 7 , Figure 7 yes Figure 2 The virtual lane line generated in step S240 in the illustrated embodiment.

[0119] Please continue reading Figure 2 , step S250, clustering the trajectory groups in the non-intersection area to generate a lane line topology in the non-intersection area.

[0120] The goal of non-intersection area processing is to obtain the topology of lane lines, including diverging lane lines and merging lane lines.

[0121] See also Figure 8 , Figure 8 yes Figure 2 Step S250 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 8 As shown, the process of clustering the trajectory groups in the non-intersection area to generate the lane line topology in the non-intersection area may include steps S810 to S820, which are described in detail as follows:

[0122] Step 810: Cluster the trajectory groups in the non-intersection area according to the travel direction and distance to obtain road segments.

[0123] That is, the trajectory groups located between two intersections and with the same travel direction are grouped into one road segment.

[0124] See also Figure 9 , Figure 9 yes Figure 8 Step S810 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 9 As shown, the process of clustering the trajectory groups in the non-intersection area according to the travel direction and distance may include steps S910 to S920, which are described in detail as follows:

[0125] Step S910 : Calculate the distance between the starting points and the distance between the end points of each two trajectory groups, take the smaller value of the two distances as the distance between the trajectory groups, and then perform preliminary clustering (hierarchical clustering) on each trajectory group based on the distance.

[0126] Considering that the trajectory group may be discontinuous or incomplete, only considering the distance between the starting point and the end point will lead to incorrect allocation of road segments for isolated trajectory groups, so further optimization is required. The optimization step is S920.

[0127] Step S920: Calculate the Hausdorff distance between the road segments after the preliminary clustering.

[0128] If the Hausdorff distance between the road segments is less than the lane width, the two road segments are merged; otherwise, the minimum distance between the trajectory groups in the road segment is calculated, and if the minimum distance is less than the lane width, the two road segments are merged.

[0129] Due to the existence of lane changes in reality, if the road segment generated in step S810 is too long, the distance between trajectory groups within the road segment will change frequently, resulting in poor clustering effect. To solve this problem, for road segments longer than 100 meters, the road segment is first segmented before clustering. The segmentation steps are as follows:

[0130] Step S1010: Find the track points at the separation positions in the track group according to the preset intervals and use them as cutting points.

[0131] The above-mentioned preset interval is less than 100 meters and can be set according to actual conditions. The purpose is to divide a long road section exceeding 100 meters into small sections.

[0132] Step S1020 : traverse the trajectory points in the trajectory group, and segment the trajectory group according to the distance between the trajectory points and the cutting points.

[0133] The trajectory points between adjacent cutting points are a small road segment.

[0134] Please continue reading Figure 8 In step S820, taking the road segment as a unit, performing distance-based clustering on the trajectory groups in each of the road segments to generate a lane line topology.

[0135] It should be noted that lane lines include main lane lines, diverging lane lines and merging lane lines.

[0136] See also Figure 11 , Figure 11 yes Figure 8 Step S820 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 11 As shown, the process of performing distance-based clustering on the trajectory groups in each road segment to generate lane line topology may include steps S1110 to S1130, which are described in detail as follows:

[0137] Step S1110 : Clustering is performed based on Hausdorff distance on the trajectory groups in each road segment. From the multiple trajectory groups in each cluster, a trajectory group that is closer to the center of the cluster is selected as a candidate lane line.

[0138] Step S1120: Clustering is performed on the candidate lane lines to generate main lane lines.

[0139] The specific process is as follows: a graph is generated using the candidate lane lines as points and the Hausdorff distances between the candidate lane lines as edges. The maximum clique algorithm (max_weight_clique) is used to find non-intersecting lane lines as the main lane lines. In the maximum clique algorithm, the smoothness weight of the lane lines is taken into consideration, and smoother lane lines are preferred.

[0140] Step S1130: Generate diverging and merging lane lines based on the trajectory group data in each road segment and the main lane line. Figure 12 , specifically including steps S1210 to S1240:

[0141] Step S1210: Within a road segment, at regular intervals, a Gaussian mixture model is used to obtain the number of lanes at corresponding locations for trajectory points at the cross section of the road segment.

[0142] There is no restriction on the interval distance here, and it can be set according to the specific situation. For example, the interval distance is 20 meters, 30 meters, etc. It should be noted that the interval distance is less than the length of the road section.

[0143] Step S1220: Combine the number of lanes at each position and the information of the main lane line to identify lanes other than the main lane line as candidate diverging and merging lane lines.

[0144] Step S1230: Determine a lane change lane based on the candidate diverging and merging lane markings and the main lane markings.

[0145] Specifically, the minimum distance min_dis_s between the starting point of each candidate diverging and merging lane and each main lane, the minimum distance min_dis_e between the end point of each candidate diverging and merging lane and each main lane, and the number of main lanes intersecting each candidate diverging and merging lane, inter_cnt, are calculated. A candidate diverging and merging lane that meets one of the following conditions is considered a lane change lane: (1) max(min_dis_s, min_dis_e) < lane width threshold; inter_cnt > 1.

[0146] Step S1240: Find a diverging and merging lane marking based on the candidate diverging and merging lane markings, the main lane marking, and the lane change lane marking.

[0147] The specific process is as follows: the intersection point of the candidate diverging and merging lane line and the main lane line after filtering out the lane change lane line is the diverging and merging point; the starting point of the candidate diverging and merging lane and the diverging and merging point generate a trajectory, recorded as s_lane, and the end point of the candidate diverging and merging lane and the diverging and merging point generate a trajectory, recorded as e_lane; if the distance between the starting point of the candidate diverging and merging lane and the main lane is not less than the lane width threshold, s_lane is retained as the diverging and merging lane line; if the distance between the end point of the candidate diverging and merging lane and the main lane is not less than the lane width threshold, e_lane is retained as the diverging and merging lane line. Examples of diverging and merging lane lines are as follows: Figure 13 and Figure 14 ,in Figure 13 Schematic diagram of the main lane and diverging lane (the dotted line represents the diverging lane). Figure 14 Schematic diagram of the main lane and merging lane (the dotted line is the merging lane).

[0148] Please continue reading Figure 2 In step S260, the virtual lane lines in the intersection area and the lane line topology in the non-intersection area are merged to generate a final lane line topology map.

[0149] See also Figure 15 , Figure 15 yes Figure 2 Step S260 in the embodiment shown is a flow chart in an exemplary embodiment. Figure 15 As shown, the process of merging the virtual lane lines of the intersection area and the lane line topology of the non-intersection area to generate a final lane line topology map may include steps S1510 to S1530, which are detailed as follows:

[0150] Step S1510: Generate node points at the starting point and end point of each lane line.

[0151] The lanes include virtual lane lines in the intersection area, main lane lines in the non-intersection area, and diverging and merging lane lines. The starting point and end point of each lane are marked as nodes.

[0152] Step S1520: interrupt the main lane line at the diverging and merging position to add a node point.

[0153] At the intersection of the diverging lane line and the main lane line, and at the intersection of the merging lane line and the main lane line, the main lane line is interrupted to add a node point.

[0154] Step S1530: Merge the nodes whose distances are close to the threshold.

[0155] The threshold here is also not limited and can be set according to the actual situation. If a node point only connects two lane lines, the node point is deleted and the lane lines are merged to generate the final lane centerline topology.

[0156] Figure 16 This is a block diagram of an exemplary embodiment of the present application showing a device for obtaining lane topology based on crowdsourced trajectories. The device can be applied to Figure 1 The implementation environment shown is specifically configured in the smart terminal 110. The apparatus may also be applicable to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the apparatus is applicable.

[0157] like Figure 16 As shown, the exemplary device for obtaining lane line topology based on crowdsourcing trajectory includes: a crowdsourcing data acquisition module 1601, an information recognition module 1602, a grouping module 1603, an intersection area processing module 1604, a non-intersection area processing module 1605 and a fusion module 1606, wherein the crowdsourcing data acquisition module 1601 is configured to obtain trajectory point information collected by crowdsourcing vehicles, and the trajectory point information includes timestamps and position information of several trajectory points; the information recognition module 1602 is configured to identify intersection areas and non-intersection areas based on the information of the trajectory points; the grouping module 1603 is configured to identify intersection areas and non-intersection areas based on the timestamps of the trajectory points and the distances of the trajectory points The plurality of trajectory points in the intersection area and the non-intersection area are grouped at intervals to obtain a trajectory group in the intersection area and a trajectory group in the non-intersection area; the intersection area processing module 1604 is configured to perform distance clustering on the trajectory group in the intersection area to generate a virtual lane line in the intersection area; the non-intersection area processing module 1605 is configured to perform distance clustering on the trajectory group in the non-intersection area to generate a lane line topology in the non-intersection area; the fusion module 1606 is configured to merge the virtual lane in the intersection area and the lane line topology in the non-intersection area to generate a final lane line topology map.

[0158] It should be noted that the device for obtaining lane line topology based on crowdsourced trajectories provided in the above embodiment and the method for obtaining lane line topology based on crowdsourced trajectories provided in the above embodiment are of the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the road condition refresh device provided in the above embodiment can, as needed, allocate the above functions to different functional modules, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0159] An embodiment of the present application also provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the method for obtaining lane line topology based on crowdsourcing trajectories provided in the above-mentioned embodiments.

[0160] Figure 17 The following is a schematic diagram showing the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application. Figure 17 The computer system 1700 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0161] like Figure 17 As shown, computer system 1700 includes a central processing unit (CPU) 1701, which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1702 or programs loaded from storage unit 1708 into random access memory (RAM) 1703, such as executing the methods described in the above embodiments. Various programs and data required for system operation are also stored in RAM 1703. CPU 1701, ROM 1702, and RAM 1703 are connected to each other via bus 1704. Input / output (I / O) interface 1705 is also connected to bus 1704.

[0162] The following components are connected to the I / O interface 1705: an input section 1706 including a keyboard, a mouse, and the like; an output section 1707 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 1708 including a hard disk; and a communication section 1709 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 1709 performs communication processing via a network such as the Internet. A drive 1710 is also connected to the I / O interface 1705 as needed. Removable media 1711, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 1710 as needed, so that computer programs read from the removable media can be installed in the storage section 1708 as needed.

[0163] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1709, and / or installed from a removable medium 1711. When the computer program is executed by the central processing unit (CPU) 1701, the various functions defined in the system of the present application are executed.

[0164] It should be noted that the computer-readable medium shown in the embodiments of the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. This propagated data signal can take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0165] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0166] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0167] Another aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a computer processor, causes the computer to perform the aforementioned method for obtaining lane topology based on crowdsourced trajectories. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently and not be incorporated into the electronic device.

[0168] Another aspect of the present application provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method for obtaining lane topology based on crowdsourced trajectories provided in each of the above embodiments.

[0169] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Anyone skilled in the art may modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, any equivalent modifications or alterations made by one of ordinary skill in the art without departing from the spirit and technical principles disclosed herein are intended to be covered by the claims of the present invention.

Claims

1. A method for obtaining lane topology based on crowdsourcing trajectories, characterized in that: The method comprises: Obtaining track point information collected by crowdsourcing vehicles, wherein the track point information includes timestamps and location information of several track points; identifying intersection areas and non-intersection areas based on the trajectory point information; Grouping the plurality of trajectory points in the intersection area and the non-intersection area according to the timestamps of the trajectory points and the distance intervals of the trajectory points, respectively, to obtain a trajectory group in the intersection area and a trajectory group in the non-intersection area; Clustering the trajectory groups in the intersection area to generate virtual lane lines in the intersection area; Clustering the trajectory group in the non-intersection area to generate a lane line topology in the non-intersection area; Merging the virtual lane lines in the intersection area and the lane line topology in the non-intersection area to generate a final lane line topology map; The step of clustering the trajectory groups in the intersection area to generate virtual lane lines in the intersection area includes: Calculate the Hausdorff distance between each two trajectory groups in the intersection area. If the distances between the first and last parts are both less than the lane width threshold, the smaller distance is used as the Hausdorff distance between the two trajectory groups. Otherwise, the distance between the middle parts of the two trajectory groups is used as the Hausdorff distance. Clustering the trajectory groups in the intersection area according to the Hausdorff distance; From each clustered trajectory cluster, a smoother trajectory group is selected as a representative trajectory to generate a virtual lane line; Clustering the trajectory groups in the non-intersection area to generate a lane line topology in the non-intersection area includes: Clustering the trajectory groups in the non-intersection area according to travel direction and distance to obtain road segments; Taking the road segment as a unit, performing distance-based clustering on the trajectory groups in each of the road segments to generate a lane line topology in a non-intersection area; Clustering the trajectory groups in the non-intersection area according to the travel direction and distance to obtain road segments, including: calculating the distance between the starting points and the distance between the end points of each two trajectory groups, taking the smaller value of the two distances as the distance between the trajectory groups, and then preliminarily clustering the trajectory groups according to the distance; The Hausdorff distance between the road segments after the preliminary clustering is calculated. If the Hausdorff distance is less than the lane width, the two road segments are merged. Otherwise, the minimum distance between the trajectory groups in the road segment is calculated. If the minimum distance is less than the lane width, the two road segments are merged.

2. The method according to claim 1, characterized in that Identifying intersection areas and non-intersection areas based on the trajectory point information includes: Calculating the angular velocity of the trajectory point according to the timestamp and direction angle of the trajectory point; Taking the trajectory point with an angular velocity greater than a first threshold as a candidate turning point; Merging the consecutive turning candidate points into a group, and if the number of non-turning candidate points between the two groups is less than a second threshold, further merging the two groups of non-turning candidate points, and marking the non-turning candidate points between the two groups as turning candidate points; Calculate the direction angle difference and timestamp difference between the start point and the end point in each group of candidate turning points, filter out the groups with differences less than the third threshold, and obtain the turning point combination; Mean shift clustering is performed on the turning points in the turning point combination, and the intersection area is identified according to the speed of the trajectory points and / or the proportion of turning points.

3. The method according to claim 2, characterized in that Performing mean shift clustering on the turning points in the turning point combination, and identifying the intersection area based on the speed of the trajectory points and / or the proportion of turning points, including: Performing mean shift clustering on the turning points in the turning point combination, where the cluster center is the intersection center; Taking the center point of the intersection as the center of the circle, construct multiple circles according to the preset radius and interval; Calculate the average speed of the track points within the annular range and the ratio of the number of the turning points to the number of the track points from small to large; If the average speed is greater than the fourth threshold, the proportion of the turning points is lower than the fifth threshold, or the radius of the ring exceeds the sixth threshold, the calculation is stopped. At this time, the circular area between the center point of the intersection and the ring is the intersection area.

4. The method according to claim 1, wherein Grouping the plurality of trajectory points in the intersection area and the non-intersection area respectively according to the timestamps of the trajectory points and the distance intervals of the trajectory points, including: Traversing the trajectory points in order of the timestamps to determine the interruption point; The trajectory points are grouped according to the areas where the trajectory points are located and the interruption points to obtain a trajectory group for the intersection area and a trajectory group for the non-intersection area.

5. The method according to claim 4, characterized in that The interruption point satisfies at least one of the following conditions: A trajectory point at a critical point between the intersection area and the non-intersection area; The time stamp interval between adjacent trajectory points exceeds a preset threshold; The distance between adjacent trajectory points exceeds a preset threshold.

6. The method according to claim 1, characterized in that If the road segment is longer than 100 meters, the preliminary clustering further includes cutting the road segment into small segments, and cutting the road segment into small segments includes: In the longest track group in the road segment, finding track points at separation positions in the longest track group at a preset interval as cutting points; Traverse the trajectory points in the trajectory group in the road segment, and segment the trajectory group according to the distance between the trajectory points and the cutting point.

7. The method according to claim 1, characterized in that Taking the road segment as a unit, performing distance-based clustering on the trajectory groups in each road segment to generate a lane line topology in a non-intersection area, including: Clustering is performed based on Hausdorff distance on the trajectory groups in each of the road segments, and a trajectory group close to the center of each cluster is selected from the multiple trajectory groups in each cluster as a candidate lane line; A graph is generated using the candidate lane lines as points and the Hausdorff distances between the candidate lane lines as edges, and a maximum clique algorithm is used to find mutually non-intersecting lane lines as the main lane lines; Based on the trajectory group data in each of the road segments and the main lane line, a diverging and merging lane line is generated.

8. The method according to claim 7, characterized in that Generating a diverging and merging lane line based on the trajectory group data in each of the road segments and the main lane line, including: In the road segment, at regular intervals, the number of lanes at the corresponding position is obtained by using a Gaussian mixture model for trajectory points at a cross section of the road segment; Combining the number of lanes at each position and the information of the main lane line, identifying lanes other than the main lane line as candidate diverging and merging lane lines; Calculate the minimum distance between the starting point of each candidate diverging and merging lane and each main lane, the minimum distance between the end point and each main lane, and the number of main lanes intersecting with each candidate diverging and merging lane. Determine the lane change lane based on the above calculation results. The intersection point of the candidate diverging and merging lane line and the main lane line after filtering out the lane change lane line is the diverging and merging point; if the distance between the starting point of the candidate diverging and merging lane line and the main lane line is not less than the lane width threshold, then the trajectory between the starting point of the candidate diverging and merging lane line and the diverging and merging point is the diverging and merging lane line; if the distance between the end point of the candidate diverging and merging lane line and the main lane is not less than the lane width threshold, then the trajectory between the end point of the candidate diverging and merging lane line and the diverging and merging point is the diverging and merging lane line.

9. The method according to claim 7, characterized in that The virtual lanes in the intersection area and the lane line topology in the non-intersection area are merged to generate a final lane line topology map, including: Generate nodes at the start and end points of each lane line; Interrupting the main lane line at the diverging and merging position to add a node; Merging the nodes whose distances are close to the threshold; If a node only connects two lane lines, the node is deleted and the lane lines are merged to generate a final lane line topology map.

10. A device for obtaining lane topology based on crowdsourced trajectories, characterized in that: The device comprises: a crowdsourcing data acquisition module configured to acquire track point information collected by crowdsourcing vehicles, wherein the track point information includes timestamps and location information of a plurality of track points; An information recognition module configured to recognize an intersection area and a non-intersection area based on information of the trajectory points; a grouping module configured to group the plurality of trajectory points in the intersection area and the non-intersection area according to the timestamps of the trajectory points and the distance intervals of the trajectory points, so as to obtain a trajectory group in the intersection area and a trajectory group in the non-intersection area; an intersection area processing module, configured to perform distance clustering on the trajectory group of the intersection area to generate a virtual lane line of the intersection area; a non-intersection area processing module configured to perform distance clustering on the trajectory group in the non-intersection area to generate a lane line topology in the non-intersection area; a fusion module configured to merge the virtual lanes in the intersection area and the lane line topology of the non-intersection area to generate a final lane line topology map; The step of clustering the trajectory groups in the intersection area to generate virtual lane lines in the intersection area includes: Calculate the Hausdorff distance between each two trajectory groups in the intersection area. If the distances between the first and last parts are both less than the lane width threshold, the smaller distance is used as the Hausdorff distance between the two trajectory groups. Otherwise, the distance between the middle parts of the two trajectory groups is used as the Hausdorff distance. Clustering the trajectory groups in the intersection area according to the Hausdorff distance; From each clustered trajectory cluster, a smoother trajectory group is selected as a representative trajectory to generate a virtual lane line; Clustering the trajectory groups in the non-intersection area to generate a lane line topology in the non-intersection area includes: Clustering the trajectory groups in the non-intersection area according to travel direction and distance to obtain road segments; Taking the road segment as a unit, performing distance-based clustering on the trajectory groups in each of the road segments to generate a lane line topology in a non-intersection area; Clustering the trajectory groups in the non-intersection area according to the travel direction and distance to obtain road segments, including: calculating the distance between the starting points and the distance between the end points of each two trajectory groups, taking the smaller value of the two distances as the distance between the trajectory groups, and then preliminarily clustering the trajectory groups according to the distance; The Hausdorff distance between the road segments after the preliminary clustering is calculated. If the Hausdorff distance is less than the lane width, the two road segments are merged. Otherwise, the minimum distance between the trajectory groups in the road segment is calculated. If the minimum distance is less than the lane width, the two road segments are merged.

11. An electronic device, characterized in that: The electronic device comprises: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the electronic device to implement the method for obtaining lane line topology based on crowdsourcing trajectories as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed by a processor of a computer, the computer is caused to execute the method for obtaining lane line topology based on crowdsourcing trajectories according to any one of claims 1 to 9.

Images