Map updating methods and devices
By introducing road network topology information to match and filter trajectory data, and using road segment groups as units to identify U-turn points, the problem of low accuracy in U-turn point identification in existing technologies is solved, and the automatic updating and routine maintenance of high-precision maps are realized.
Patent Information
- Application Number
- CN202611248100.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-17
- Publication Date
- 2026-11-10
AI Technical Summary
Existing U-turn identification methods based on trajectory geometry features are susceptible to GPS drift, road curvature, and building obstruction, resulting in low identification accuracy and an inability to support automated updates of high-precision map data.
Road network topology information is introduced to perform map matching and filtering on the original trajectory data, generating matching trajectory strings. New U-turn results and redundant verification of the original U-turn points are identified using road segment groups as analysis units. The map is updated by combining the new and redundant identification results.
It effectively filters out trajectory noise interference, improves the accuracy of U-turn identification, enables automated high-frequency updates, reduces manual data collection costs, and provides reliable support for routine map maintenance.
Smart Images

Figure CN123068522A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of map updating technology, and in particular to a map updating method and apparatus. Background Technology
[0002] U-turn points are important semantic information about road traffic in standard electronic maps. The accuracy and timeliness of their data updates directly affect the quality of map services such as navigation route planning and driving safety tips.
[0003] In recent years, automatic U-turn intersection discovery technology based on crowdsourced vehicle GPS trajectory data has gradually developed. Existing discovery schemes mainly rely on the geometric features of the original trajectory for identification. However, automatic identification methods based on trajectory geometry features are easily affected by positioning drift, uneven sampling, and road curvature. The identification results are highly dependent on the original trajectory shape, resulting in low accuracy of U-turn intersection identification and failing to directly support high-precision automated map data updates. Summary of the Invention
[0004] The map updating method and apparatus provided in this application are used to improve the accuracy of U-turn identification.
[0005] In a first aspect, embodiments of this application provide a map updating method, including:
[0006] Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period; wherein, the road network topology information includes multiple road segment groups and original U-turn information, and each road segment group includes multiple road segments in the up and down directions of the same physical road;
[0007] Based on the road network topology information, the original trajectory data is subjected to map matching and trajectory filtering processing to obtain a matching trajectory string; wherein, the trajectory points in the matching trajectory string are associated with the matched road segments;
[0008] Based on the matching trajectory strings belonging to each road segment group, determine the newly added identification results of the U-turn entrances corresponding to each road segment group; and / or, perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the redundant identification results of the U-turn entrances.
[0009] Update the map based on the newly added identification results of the U-turn point and / or the redundant identification results of the U-turn point.
[0010] In one possible implementation, the step of performing map matching and trajectory filtering processing on the original trajectory data based on the road network topology information to obtain a matched trajectory string includes:
[0011] Based on the road network topology information, map matching is performed on the original trajectory data to generate an initial matching trajectory string; wherein, each trajectory point in the initial matching trajectory string is associated with the matched road segment;
[0012] According to preset reliability conditions, the initial matching trajectory string is filtered to obtain a matching trajectory string; wherein, the reliability conditions include at least one of road segment connectivity conditions, heading deviation threshold conditions, and road segment length deviation conditions.
[0013] In one possible implementation, the road segment connectivity condition is: the road segment sequence corresponding to the trajectory string contains consecutive entry road segments and exit road segments, and includes the preceding adjacent road segment of the entry road segment and the following adjacent road segment of the exit road segment;
[0014] The heading deviation threshold condition is as follows: the difference between the heading angle of the trajectory string corresponding to the previous adjacent road segment and the heading angle of the previous adjacent road segment is less than a preset angle threshold, and the difference between the heading angle of the trajectory string corresponding to the next adjacent road segment and the heading angle of the next adjacent road segment is less than a preset angle threshold.
[0015] The road segment length deviation condition is as follows: the difference between the length of the trajectory string corresponding to the previous adjacent road segment and the length of the previous adjacent road segment is less than a preset percentage threshold, and the difference between the length of the trajectory string corresponding to the next adjacent road segment and the length of the next adjacent road segment is less than a preset percentage threshold.
[0016] In one possible implementation, determining the newly added identification result of the U-turn intersection corresponding to each road segment group based on the matching trajectory string belonging to each road segment group includes:
[0017] For each road segment group, the matching trajectory string belonging to that road segment group is determined from the matching trajectory string;
[0018] For each matching trajectory string belonging to the road segment group, if it is determined that the matching trajectory string has U-turn characteristics based on the heading time sequence change information of the matching trajectory string, then a candidate point for the U-turn is determined based on the matching trajectory string.
[0019] Clustering is performed on the candidate points of the U-turn to obtain the location information of the newly added U-turn points corresponding to the road segment group.
[0020] In one possible implementation, determining the candidate points for the U-turn based on the matching trajectory string includes:
[0021] Project each trajectory point in the matched trajectory string onto the corresponding road segment to obtain each projection point;
[0022] The projection point furthest from the starting point of the corresponding entry segment of the matched trajectory string is selected as the candidate point for the U-turn.
[0023] In one possible implementation, the road network topology information further includes the location information of intersection nodes corresponding to the road segment group; the step of clustering the candidate U-turn points to obtain the location information of newly added U-turn points corresponding to the road segment group includes:
[0024] Clustering is performed on the candidate U-turn points to obtain the initial location information of newly added U-turn points corresponding to the road segment group;
[0025] If the distance deviation between the initial newly added U-turn location information and the intersection node location information corresponding to the road segment group is determined to be greater than or equal to a preset distance threshold, then the initial newly added U-turn location information corresponding to the road segment group is taken as the newly added U-turn location information corresponding to the road segment group.
[0026] In one possible implementation, the step of performing redundancy verification on the original U-turn information based on the matched trajectory string to obtain the U-turn redundancy identification result includes:
[0027] Based on the matching trajectory string, the traffic characteristics of each road segment in the current statistical period are determined; the traffic characteristics include at least one of the following: the transition probability from entering the road segment to exiting the road segment, the total traffic flow of the entering road segment, and the total traffic flow of the road segment pair between the entering and exiting road segments;
[0028] For a target original U-turn point, if the traffic characteristics of the entry and exit segments corresponding to the target original U-turn point meet preset conditions within multiple consecutive statistical periods, then the target original U-turn point is determined to be a redundant U-turn point; wherein, the preset conditions include at least one of the following: the transfer probability is less than a preset probability threshold, the total flow of the entry segment is less than a preset flow threshold, and the duration of the total flow of the segment pair being 0 is greater than a preset duration threshold.
[0029] In one possible implementation, updating the map based on the newly added identification results of the U-turn and / or the redundant identification results of the U-turn includes:
[0030] The location information of the newly added U-turn point corresponding to the newly identified U-turn point is entered into the map road network data; and / or,
[0031] The original U-turn information corresponding to the redundant identification results of the U-turn point is marked as invalid or removed from the map road network data.
[0032] Secondly, embodiments of this application provide a map updating apparatus, comprising:
[0033] The acquisition module is used to acquire the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period; wherein, the road network topology information includes multiple road segment groups and original U-turn information, and each road segment group includes multiple road segments in the up and down directions of the same physical road;
[0034] The processing module is used to perform map matching and trajectory filtering processing on the original trajectory data according to the road network topology information to obtain a matching trajectory string; wherein, the trajectory points in the matching trajectory string are associated with the matched road segments;
[0035] The identification module is used to determine the newly added identification results of the U-turn entrances corresponding to each road segment group based on the matching trajectory strings belonging to each road segment group; and / or to perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the U-turn entrance redundancy identification results.
[0036] The update module is used to update the map based on the newly added identification results of the U-turn and / or the redundant identification results of the U-turn.
[0037] Thirdly, embodiments of this application provide an electronic device, including: a memory and a processor;
[0038] The memory stores computer-executed instructions;
[0039] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.
[0040] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.
[0041] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.
[0042] The map updating method and apparatus provided in this application first acquire the original trajectory data of crowdsourced vehicles in the current statistical period and the road network topology information containing road segment groups. Based on the road network topology, map matching and filtering processing is performed on the original trajectories to obtain matched trajectory strings with road segment affiliations. Then, new U-turn point identification is carried out using road segment groups as analysis units, and / or redundancy verification is performed on existing original U-turn points based on the matched trajectory strings. Finally, the semantic data of map U-turn points is iteratively updated by combining the new identification results and / or the redundancy verification results. This method introduces the road network topology divided into road segment groups as a structured constraint, transforming the original geometric recognition based on discrete original trajectories into structured analysis based on road units. This effectively filters trajectory noise interference caused by GPS drift, road curvature, and building obstruction, distinguishes between real U-turn behavior and regular turning and noise trajectories, improves the accuracy of U-turn point identification, and simultaneously achieves automated high-frequency updates of U-turn point data, reducing the cost and cycle of manual collection and providing reliable technical support for the routine maintenance of maps. Attached Figure Description
[0043] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.
[0044] Figure 1 Flowchart of the map update method provided in this application Figure 1 ;
[0045] Figure 2 A schematic diagram of the road segment group provided in this application;
[0046] Figure 3 Flowchart of the map update method provided in this application Figure 2 ;
[0047] Figure 4 A schematic diagram of the matching trajectory string provided in this application;
[0048] Figure 5 A schematic diagram of a candidate point for a U-turn is provided in this application;
[0049] Figure 6 Flowchart of the map update method provided in this application Figure 3 ;
[0050] Figure 7 A schematic diagram of a trajectory transfer provided for this application;
[0051] Figure 8 Flowchart of the map update method provided in this application Figure 4 ;
[0052] Figure 9Flowchart of the map update method provided in this application Figure 5 ;
[0053] Figure 10 A schematic diagram of the structure of the map updating device provided in this application;
[0054] Figure 11 A schematic diagram of the structure of the electronic device provided in this application.
[0055] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation
[0056] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application.
[0057] U-turn intersections are important semantic information in Standard Definition (SD) maps. Current methods for collecting and updating U-turn intersection data mainly rely on manual on-site collection or specialized data collection vehicles, which are inefficient and costly. In recent years, technologies for automatically mining U-turn intersections based on GPS trajectory data have gradually developed, such as extreme value detection methods based on trajectory geometric features and continuous turning angle cumulative difference detection methods.
[0058] Among them, the core idea of the extreme value detection method based on trajectory geometric features is to take the straight distance from each trajectory point to the starting point of the yaw trajectory record as the analysis object, obtain the trajectory points whose straight distance reaches the extreme value by traversing, and then filter out the extreme value points that meet the U-turn intersection rules as the U-turn position by combining the preset U-turn angle range (1°~10°).
[0059] The core idea of the method based on continuous cumulative difference of turning angles is to calculate the continuous cumulative difference of turning angles for each GPS trajectory, identify trajectory segments exceeding a set threshold as U-turn trajectory segments, and then construct a database of potential illegal U-turn trajectories. It then combines deep neural networks such as object detection models to identify illegal elements (fences, traffic lights, U-turn signs, etc.) in street scene images, distinguishes between legitimate and illegal U-turns through scene analysis, and finally extracts hotspot areas for illegal U-turns.
[0060] However, the above solutions all rely on the original trajectory features. The noise of the trajectory itself (GPS drift, uneven sampling), road curvature, building obstruction and other factors will directly interfere with the detection results, resulting in low accuracy of U-turn recognition and failing to directly support high-precision automated map data updates.
[0061] In view of this, the inventors of this application realized that optimizing only the geometric features of the trajectory itself is insufficient to fundamentally eliminate misjudgments caused by external interference. It is necessary to introduce structured constraints at the road network level to provide spatial anchoring and boundary limits for U-turn behavior recognition. Therefore, this application proposes to rely on road network topology information divided into road segment groups, firstly perform map matching and filtering on the original trajectory data of crowdsourced vehicles to obtain matching trajectory strings that are associated with each road segment. Then, taking each road segment group consisting of up and down road segments of the same physical road as the basic analysis unit, on the one hand, new U-turn results are identified based on the matching trajectory strings belonging to the corresponding road segment group, and on the other hand, redundancy verification is performed on the existing original U-turn information based on the matching trajectory strings. Finally, the map data is updated synchronously by combining the new recognition results and the redundancy verification results, thereby effectively filtering out false detections caused by trajectory noise and road morphology interference, significantly improving the accuracy of U-turn recognition, and providing reliable support for the automated updating of high-precision maps.
[0062] The execution subject of this application embodiment can be an electronic device with processing capabilities, such as a computer, server, etc., and this application embodiment is not limited thereto.
[0063] It should be noted that there are no limitations on the application areas of this application; for example, it can be applied to intelligent transportation, electronic map maintenance and dynamic updates.
[0064] It should be noted that the trajectory data involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.
[0065] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.
[0066] Figure 1 Flowchart of the map update method provided in this application Figure 1 ,like Figure 1 As shown, the method includes:
[0067] S101. Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period.
[0068] For example, the current statistical period refers to a pre-set time window for batch statistical analysis of vehicle driving behavior, which can be flexibly configured by day, week or month according to update requirements. This application embodiment does not limit this.
[0069] Crowdsourced vehicles refer to vehicles that are connected to the map crowdsourcing data collection system to provide full-area, routine driving trajectory samples. Specifically, these can be vehicles equipped with in-vehicle navigation devices, mobile communication terminals, or other positioning data collection units. These vehicles upload vehicle location, timestamp, heading angle, speed, and positioning information during actual road travel.
[0070] Raw trajectory data refers to the sequence of GPS trajectory points collected and reported by vehicle terminals that has not undergone map matching processing, and includes basic information such as positioning coordinates, collection timestamps, and driving heading angles.
[0071] Road network topology information refers to the basic map data that describes the spatial structure and connectivity of roads. It includes multiple link groups and original U-turn information. A link is the smallest unidirectional road unit that makes up a road network. Each link group consists of multiple links in the up and down directions of the same physical road. It is used to constrain discrete trajectories to the corresponding road units for analysis. In other words, it can determine whether a vehicle has switched from one direction to another within the same road entity.
[0072] Original U-turn information refers to the existing semantic data of U-turns that have been included in the current map and are set in the middle of road segments, including attribute information such as the location of the U-turn and the road segment to which it belongs.
[0073] In one example, an electronic device can pull raw trajectory data packets reported by all crowdsourced vehicles in the target area within the current statistical period from a cloud-based crowdsourced data platform, and at the same time retrieve the road network topology information of the corresponding area from the map's basic road network database.
[0074] In this application, to facilitate subsequent aggregation by road entity, the electronic device can perform road segment assembly processing after acquiring road network topology information. Specifically, multiple road segments representing the up and down directions of the same physical road can be merged into the same road segment group (anchor point selected) according to information such as road spatial location relationships, traffic direction, and node connection structure. In other words, this application takes roads with separated up and down lanes as the research object and constructs a linkgroup for the same group of roads with separated up and down lanes. All links in the same group of roads with separated up and down lanes constitute a linkgroup, which serves as the anchor unit for newly excavated U-turn points. For example, Figure 2 A schematic diagram of the road segment group provided in this application, such as Figure 2As shown, let the links that are separated into upper and lower lines be link1, link2, link3 and link4. Then link1+link2+link3+link4 constitutes a link group. link1+link2+link3 and link4 are two links in opposite directions in the link group.
[0075] By uniformly acquiring raw trajectory data and road network topology information within the same statistical period, and based on the segment groups that include multiple road segments in both directions of the same physical road and the original U-turn information, subsequent processing can be performed within the road entity range constrained by the topology. This provides a data foundation for the identification of new U-turns and the redundancy verification of original U-turns.
[0076] S102. Based on the road network topology information, perform map matching and trajectory filtering on the original trajectory data to obtain a matching trajectory string.
[0077] Among them, the trajectory points in the matching trajectory string are associated with the matched road segments.
[0078] For example, map matching refers to the process of mapping discrete positioning points in the original trajectory to road segments in the road network topology information, so that discrete sampling points are converted into a travel sequence with a clear road affiliation and the road entity to which the trajectory belongs is clear.
[0079] Trajectory filtering is used to remove trajectories that do not meet reliability constraints from map matching results, so that the remaining trajectories can reflect the actual travel paths of vehicles in the road network.
[0080] The matched trajectory string represents the trajectory result after map matching and trajectory filtering. Each trajectory point in the result has a corresponding matching road segment identifier, projection position, cumulative distance along the road segment, and adjacent relationships between trajectory points. Multiple trajectory points form a continuous road passage string in chronological order.
[0081] In one example, an electronic device can use a hidden Markov map matching algorithm, combined with road segment coordinates and connectivity information from the road network topology, to map the original trajectory points one by one to the matching road segment with the highest confidence, generating an initial matching trajectory string. Then, based on preset reliability conditions such as trajectory continuity and positioning accuracy rules, invalid trajectory strings with positioning jumps or road segment matching interruptions are eliminated, finally obtaining a matching trajectory string that can be used for subsequent U-turn identification and analysis.
[0082] By associating the original trajectory points with specific road segments and their topological relationships, and by performing reliability screening on the matching results, the resulting matching trajectory string can retain the key traffic behaviors of vehicles in the road network.
[0083] S103. Based on the matching trajectory strings belonging to each road segment group, determine the newly added identification results of the U-turn entrances corresponding to each road segment group; and / or, perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the redundant identification results of the U-turn entrances.
[0084] For example, a matching trajectory string belonging to a road segment group refers to a matching trajectory string in which the travel path falls within the coverage area of the corresponding road segment group and the matching road segment belongs to that road segment group.
[0085] The newly added U-turn identification results refer to the location and attribute information of U-turns in the middle of road segments that are not yet included in the current map and are identified from the matched trajectory strings within the current statistical period.
[0086] Redundancy verification refers to the process of verifying whether existing U-turn points still possess genuine passage attributes based on actual trajectory passage behavior.
[0087] The identification result of redundant U-turn points refers to the information of existing U-turn points that are determined to be invalid and have no real traffic support.
[0088] In one example, the electronic device can first aggregate all matching trajectory strings into corresponding road segment groups according to their respective road segments. For each road segment group, it can extract the heading time sequence change features of the matching trajectory strings within the group, identify trajectory segments with U-turn behavior characteristics, and obtain the identification result of the new U-turn exit for that road segment group after clustering and intersection filtering; and / or, using the original U-turn exit information as the verification object, extract the passage records related to the original U-turn exit from the matching trajectory strings of the current statistical period, such as passage frequency, turning probability, etc., and judge the validity of the passage records of the original U-turn exit, thereby obtaining the identification result of U-turn exit redundancy.
[0089] In this step, the addition of new identification and the redundancy verification of the original turnaround point can be performed independently or jointly. By analyzing within the same topology framework, the electronic device can simultaneously complete the turnaround point replenishment and failure verification within a statistical period.
[0090] By using road segment groups as analysis units and combining them with matching trajectory strings, the newly identified U-turn points are determined, and the original U-turn points are redundantly verified by matching trajectory strings. This ensures that both the newly identified U-turn points and the redundant U-turn points are based on traffic evidence.
[0091] S104. Update the map based on the newly added identification results of the U-turn and / or the redundant identification results of the U-turn.
[0092] For example, updating the map refers to the incremental iterative operation of the semantic data of U-turns in the map road network. In other words, it involves writing the newly identified U-turns and / or redundant U-turns generated in the current statistical period into the map database, and simultaneously correcting the data status related to U-turns. The updated map data includes at least newly added U-turn records, invalid U-turn records, replacement relationship records, update timestamps, and binding information with road segment groups.
[0093] In one example, the electronic device can input the semantic information such as the location and road segment corresponding to the newly added U-turn identification result into the map road network semantic database, and / or mark the original U-turn that is determined to be invalid by redundant identification results as invalid or remove it directly from the road network data, thus completing the incremental update of map data for this statistical period.
[0094] By mapping the newly added and redundant U-turn identification results to the update operation in the map database, the updated map can reflect the actual U-turn traffic status of crowdsourced vehicles within the current statistical period.
[0095] The map update method provided in this application first obtains the original trajectory data of crowdsourced vehicles in the current statistical period and the road network topology information containing road segment groups. Based on the road network topology, the original trajectories are matched and filtered to obtain matching trajectory strings with road segment affiliations. Then, new U-turn point identification is carried out using road segment groups as analysis units, and / or redundancy verification is performed on existing original U-turn points based on the matching trajectory strings. Finally, the semantic data of map U-turn points is iteratively updated by combining the new identification results and / or the redundancy verification results. This method introduces the road network topology divided into road segment groups as a structured constraint, transforming the original geometric recognition based on discrete original trajectories into structured analysis based on road units. It can effectively filter trajectory noise interference caused by GPS drift, road curvature, and building obstruction, distinguish between real U-turn behavior and regular turning and noise trajectories, improve the accuracy of U-turn point identification, and realize automated high-frequency updates of U-turn point data, reducing the cost and cycle of manual collection and providing reliable technical support for the routine maintenance of maps.
[0096] Figure 3 Flowchart of the map update method provided in this application Figure 2 ,like Figure 3 As shown, in this embodiment... Figure 1 Based on the embodiments, the map updating method is described in detail, which includes:
[0097] S201. Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period.
[0098] It should be noted that this step is similar to the aforementioned step S101, and will not be repeated here.
[0099] S202. Based on the road network topology information, perform map matching on the original trajectory data to generate an initial matching trajectory string.
[0100] For example, the initial matching trajectory string refers to the trajectory sequence obtained after the initial map matching is completed. Each trajectory point in the sequence is bound to the matching road segment with the highest confidence. That is, each trajectory point in the initial matching trajectory string is associated with the matched road segment and has a preliminary road affiliation relationship, but has not yet undergone validity verification and noise filtering.
[0101] In one example, the electronic device can use a hidden Markov map matching algorithm, which combines the road segment geometric coordinates and connectivity rules of the road network topology to calculate the matching probability of each candidate road segment for each original trajectory point, map the trajectory points to the road segment with the highest matching probability, and concatenate them in the order of collection time to form an initial matching trajectory string.
[0102] S203. According to the preset reliability conditions, the initial matching trajectory string is filtered to obtain the matching trajectory string.
[0103] For example, the preset reliability conditions are a set of judgment rules used to verify the validity of the initial matched trajectory string, and are used to filter invalid trajectory samples that do not meet reliability constraints or have severe noise interference. The reliability conditions include at least one of the following: road segment connectivity conditions, heading deviation threshold conditions, and road segment length deviation conditions. Specifically, the road segment connectivity condition verifies the topological continuity of the road segment sequence corresponding to the trajectory, ensuring that the trajectory travel path conforms to the true connectivity logic of the road network; the heading deviation threshold condition limits the maximum allowable difference between the actual heading of the trajectory and the heading of the matched road segment, avoiding trajectory mismatch to road segments with inconsistent travel directions; and the road segment length deviation condition limits the proportional deviation range between the actual travel length of the trajectory segment and the length of the corresponding road segment, excluding abnormal trajectories with severe positioning drift. The matched trajectory string refers to a high-quality trajectory sequence that has passed all reliability checks and can be used for subsequent U-turn identification and analysis.
[0104] In one example, the electronic device can sequentially verify the road segment sequence connectivity, the heading difference between the trajectory segment and the matching road segment, and the length deviation ratio between the trajectory segment and the corresponding road segment for each initial matched trajectory string. Abnormal trajectory strings that do not meet the preset threshold requirements are eliminated, and the trajectories that meet the conditions are retained as the final matched trajectory strings.
[0105] By incorporating road network topology information into data preprocessing steps such as map matching and trajectory filtering, the matched trajectory strings can retain trajectory segments consistent with real road connectivity, driving direction, and road segment scale. This provides stable input for subsequent determination of new and / or redundant U-turn identification results based on the matched trajectory strings. Using this method, the sensitivity of trajectory matching results to positioning drift and local noise is reduced, the reliability of the matched trajectory strings is improved, and the map update results based on crowdsourced trajectories are more consistent with actual traffic conditions.
[0106] Optionally, the road segment connectivity condition is: the road segment sequence corresponding to the trajectory string contains consecutive entry road segments and exit road segments, and includes the previous adjacent road segment of the entry road segment and the next adjacent road segment of the exit road segment.
[0107] The entry and exit segments are defined as associated road segments for a single U-turn operation. The road segment a vehicle enters before performing a U-turn is called the entry segment, and the road segment it exits after completing the U-turn is called the exit segment. The preceding adjacent road segment is the next-level road segment in the road network topology that is directly connected to the entry end of the entry segment, and the following adjacent road segment is the next-level road segment that is directly connected to the exit end of the exit segment. The road segment connectivity condition is used to verify whether the road segment sequence corresponding to the trajectory has a complete U-turn passage link, ensuring that the trajectory travel path conforms to the complete U-turn passage logic.
[0108] The heading deviation threshold condition is: the difference between the heading angle of the trajectory string corresponding to the previous adjacent road segment and the heading angle of the previous adjacent road segment is less than a preset angle threshold, and the difference between the heading angle of the trajectory string corresponding to the next adjacent road segment and the heading angle of the next adjacent road segment is less than a preset angle threshold.
[0109] The heading deviation threshold condition is used to verify the consistency between the actual travel direction of the trajectory and the reference direction of the matched road segment. The preset angle threshold is a pre-set maximum allowable heading deviation value to exclude abnormal matching results with mismatched directions.
[0110] The road segment length deviation condition is: the difference between the length of the trajectory string corresponding to the previous adjacent road segment and the length of the previous adjacent road segment is less than a preset percentage threshold, and the difference between the length of the trajectory string corresponding to the next adjacent road segment and the length of the next adjacent road segment is less than a preset percentage threshold.
[0111] The road segment length deviation condition is used to verify the degree of matching between the actual travel distance of the trajectory and the actual length of the corresponding road segment. The preset proportion threshold is a pre-set maximum allowable percentage of length deviation, which filters out abnormal samples with severe positioning drift and trajectory distortion.
[0112] Specifically, the entry and exit segments are first determined from the matched trajectory string, and the preceding and following adjacent segments are extracted based on the road network topology. Then, it is determined whether the segment sequence of the trajectory string simultaneously contains the aforementioned continuous connection relationship. Subsequently, the difference between the heading angle of the trajectory string corresponding to the preceding adjacent segment and the standard heading angle of that adjacent segment is calculated, and the difference between the heading angle of the trajectory string corresponding to the following adjacent segment and the standard heading angle of that adjacent segment is also calculated. These differences are compared with preset angle thresholds. Simultaneously, the percentage difference between the length of the trajectory string and the standard length of the segment on both the preceding and following adjacent segments is calculated and compared with preset proportion thresholds. Only when the constraints of continuous segment relationship, heading consistency, and length consistency are all satisfied is the trajectory string retained as a valid matching result, resulting in a matched trajectory string, which is used for subsequent U-turn point identification or original U-turn point redundancy verification.
[0113] For example, Figure 4 A schematic diagram of the matching trajectory string provided in this application, such as Figure 4 As shown, link2 is the entry segment, link4 is the exit segment, link1 is the preceding adjacent segment of the entry segment, and link5 is the following adjacent segment of the exit segment. A trajectory string that satisfies the above segment connectivity conditions must continuously pass through link1-link2-link4-link5.
[0114] A trajectory string that satisfies the above heading deviation threshold condition can be represented as follows: the difference between the heading angle of the trajectory string corresponding to link1 and the heading angle of link1 is less than 30°, and the difference between the heading angle of the trajectory string corresponding to link5 and the heading angle of link5 is less than 30°.
[0115] A trajectory string that satisfies the above-mentioned road segment length deviation condition can be represented as follows: the difference between the length of the trajectory string corresponding to link1 and the length of link1 is less than 20%, and the difference between the length of the trajectory string corresponding to link5 and the length of link5 is less than 20%.
[0116] Understandably, trajectory strings that do not meet the above-mentioned preset reliability conditions are considered unreliable, and reliable matching trajectory strings can be obtained after filtering.
[0117] This method allows consecutive entry and exit road segments to form a complete road segment connection on both sides of the U-turn, while the introduction of the preceding and following adjacent road segments further defines the actual traffic flow background before and after the U-turn. Heading deviation and length deviation together serve as dual consistency constraints, making it difficult for abnormal trajectories affected by positioning drift, short-term detours, or local sampling sparsity to pass the screening. The matched trajectory string obtained through this method has high topological reliability and geometric consistency, fundamentally filtering out GPS drift and invalid trajectories. This ensures that subsequent processing is based on the real road topology, directly supporting the generation of new U-turn identification results and the redundancy verification of original U-turn information, thereby improving the accuracy and stability of map update results.
[0118] S204. For each road segment group, determine the matching trajectory string belonging to that road segment group from the matching trajectory string.
[0119] For example, an electronic device can traverse the bound road segment information of all matching trajectory strings based on the road segment identifiers contained in each road segment group, and aggregate the trajectory strings whose matching road segments fall within the target road segment group into that road segment group, thus completing the trajectory sample division of all road segment groups one by one. For example, for each road segment group, the matching trajectory strings whose corresponding entry road segments belong to that road segment group can be taken as the matching trajectory strings belonging to that road segment group.
[0120] S205. For each matching trajectory string belonging to the road segment group, if it is determined that the matching trajectory string has U-turn characteristics based on the heading time sequence change information of the matching trajectory string, then a candidate point for the U-turn is determined based on the matching trajectory string.
[0121] For example, heading time-series change information refers to the continuous change data of the heading angle of each trajectory point in the matched trajectory string arranged in chronological order, which is the core basis for identifying changes in the vehicle's driving direction. U-turn characteristics refer to the heading change characteristics corresponding to a vehicle completing a U-turn in the middle of a road segment, typically manifested as a continuous deflection of the heading angle of approximately 180 degrees within a short driving distance. Candidate points for U-turn exits are candidate points extracted from a single U-turn trajectory used to locate the position of the U-turn exit.
[0122] In one example, the electronic device can calculate the difference in heading angle between adjacent trajectory points in the matching trajectory string in chronological order, and accumulate the total heading change of the trajectory segment as the heading time sequence change information. When the total heading change reaches the preset turning determination threshold, the trajectory string is determined to have turning characteristics, and then the point corresponding to the position with the largest heading change is extracted from the turning trajectory segment as the candidate point of the turning point.
[0123] In another example, starting from the first trajectory point matched with the entry link of the road segment group, the angular deviation between the actual heading and the corresponding link's direction of travel is calculated point by point in the matched trajectory string. The temporal information of the angular deviation is used as the heading temporal change information to determine whether the matched trajectory string has a U-turn characteristic. For example, if the heading angle deviation corresponding to the trajectory point matched to the entry link continuously changes from around 0° to around 180°, or shows a continuous reverse trend from 180° to 0°, then the trajectory string is considered to have a U-turn characteristic. If the angle deviation of the trajectory point corresponding to the entry link always remains around 0°, while the angle deviation between the trajectory point matched to the next link and the current entry link is around 180°, and the spatial deviation distance between the two trajectories is within 10 meters, then the matched trajectory string is considered to have a U-turn characteristic. Furthermore, the critical point of the heading angle deviation can be used as a candidate point for the U-turn.
[0124] Understandably, U-turn identification schemes in related technologies cannot distinguish between U-turns at intersections and U-turns in the middle of road segments. This application introduces a road segment group anchor point mechanism as a structured judgment benchmark. This mechanism uses roads that are separated in the up and down directions on a physical road as anchor analysis units, integrating all road segments that are separated in the up and down directions on the same physical road into a road segment group. Only when a complete U-turn behavior occurs within a single closed road segment group, that is, when a vehicle enters from a road segment included in the road segment group, completes the U-turn operation, and then enters a road segment within the same road segment group again, will the corresponding point be judged as a valid candidate point for a road segment U-turn. Relying on the unit boundary constraints of the road segment group, it is possible to naturally distinguish between U-turn behavior at intersections and U-turn behavior in the middle of road segments from a logical level, avoiding confusion between the two types of scenarios and causing false detections.
[0125] Specifically, each trajectory point in the matched trajectory string can be projected onto the corresponding road segment to obtain each projection point; the projection point farthest from the starting point of the corresponding entry road segment of the matched trajectory string is used as a candidate point for the U-turn.
[0126] Projection to the corresponding road segment refers to the spatial coordinate transformation operation that maps trajectory points in space vertically to the centerline of their matching road segment. This is used to eliminate lateral deviations in GPS positioning and uniformly anchor the points to the road baseline. The corresponding road segment refers to the target road segment bound to each trajectory point after map matching; it serves as the reference carrier for trajectory point projection. The projection point refers to the corresponding point on the centerline of the road segment after the trajectory point is projected, possessing mileage location attributes along the road segment's travel direction. The entry road segment refers to the target road segment that the vehicle enters before performing a U-turn; it is the road unit that carries the starting position of the U-turn. The starting position of the entry segment refers to the starting endpoint of the entry segment closest to the upstream travel direction, serving as the reference zero point for calculating the mileage distance of the projection point. The candidate U-turn point refers to the candidate point extracted from a single U-turn trajectory, used to characterize the location where the U-turn occurs.
[0127] For example, such as Figure 5 As shown, Figure 5 This is a schematic diagram of a U-turn candidate point provided in this application. The electronic device can first extract the centerline vector data of the entry road segment corresponding to the target matching trajectory string, take the starting point of the entry road segment as the zero point of mileage, and establish a mileage coordinate system along the driving direction of the road segment; then, project each trajectory point of the matching trajectory string onto the centerline in a direction perpendicular to the centerline of the road segment to obtain the projection point corresponding to each trajectory point, and calculate the mileage value corresponding to each projection point; finally, iterate through the mileage values of all projection points, and select the projection point with the largest mileage value, that is, the projection point farthest from the starting point of the entry road segment, as the U-turn candidate point (Turnpoint) corresponding to the matching trajectory string.
[0128] By first projecting trajectory points onto the corresponding road segment, and then selecting the farthest projection point with the starting position of the entry segment as a reference, the candidate U-turn position can directly reflect the farthest travel position of the vehicle on the entry segment, thus forming a stable output of candidate U-turn points. This method reduces the sensitivity to local jitter of the original trajectory, making the spatial reference used for subsequent clustering and updating of U-turn point positions more consistent, thereby improving the stability and usability of U-turn point candidate point positioning.
[0129] S206. Cluster the candidate points of the U-turn exit to obtain the location information of the newly added U-turn exit corresponding to the road segment group.
[0130] For example, clustering is a spatial aggregation operation performed on all candidate U-turn points within the same road segment group. It is used to aggregate common U-turn locations of multiple vehicle trajectories, filter discrete noise points, and eliminate positioning deviations of individual trajectories. The newly added U-turn location information refers to the spatial coordinates and attribute information of the U-turn points in the middle of road segments that are not yet included in the current map, obtained after multi-trajectory aggregation verification.
[0131] In one example, the electronic device can employ a clustering algorithm to spatially cluster all candidate U-turn points within the same road segment group that share the same U-turn direction. Discrete, noisy candidate points are eliminated, and the center of each cluster is used as the final location of the new U-turn point. Simultaneously, the corresponding road segment identifier and travel direction attribute are matched to generate the location information of the new U-turn point for that road segment group. The clustering algorithm used can be a density-based clustering algorithm or a hierarchical clustering algorithm; other clustering methods can also be selected in practical applications, and this application does not impose any limitations on this.
[0132] For example, taking a density-based clustering algorithm, spatial clustering is performed on all candidate points for U-turns within the same road segment group that have the same U-turn direction. A neighborhood distance threshold and a minimum number of samples threshold for clustering are pre-configured. After traversing all candidate points, the regions where the number of candidate points within the neighborhood meets the minimum number of samples threshold are divided into effective clusters. At the same time, discrete candidate points that cannot be classified into any effective cluster are judged as noise candidate points and removed. Then, for each effective cluster, the confidence level of the effective cluster is determined based on the number of candidate points in the effective cluster and the spatial dispersion of the candidate points. The spatial center coordinates of the effective cluster with the highest confidence level are used as the location of the new U-turn for the road segment group. The corresponding road segment identifier and U-turn direction attribute are matched for the new U-turn, and finally, the location information of the new U-turn for the road segment group is generated.
[0133] Taking hierarchical clustering algorithm as an example, spatial clustering is performed on all candidate points of U-turns within the same road segment group that have the same U-turn direction. A bottom-up agglomerative clustering strategy is adopted. Initially, each candidate point of a U-turn is regarded as an independent initial cluster. The spatial distance between all pairs of clusters is calculated iteratively. The two clusters with the closest spatial distance are continuously merged until the minimum spatial distance between any two clusters exceeds the preset distance threshold. After the clustering is completed, clusters with fewer than the preset minimum sample number threshold are identified as noise clusters and removed. After obtaining several effective clusters, the location confidence of each effective cluster is quantitatively evaluated by comprehensively considering the number of candidate points contained in the effective clusters and the spatial dispersion of the candidate points within the clusters. The spatial center coordinates of the effective cluster with the highest confidence are selected as the final location result of the new U-turn. At the same time, the corresponding road segment identifier and U-turn direction attribute are matched for the U-turn, generating the location information of the new U-turn for the road segment group.
[0134] By adopting this method, the U-turn identification process is combined with the road segment group topology, which can limit the identification range to the up and down connected areas of the same road, reducing the interference of ordinary turns and local noise on the results; and by clustering candidate points to form a unified location result, the dispersion caused by the deviation of a single trajectory can be reduced, making the newly added U-turn identification results more stable and convenient for subsequent map entry.
[0135] Specifically, the road network topology information also includes the location information of the intersection nodes corresponding to the road segment group; the candidate points of the U-turn are clustered to obtain the initial new U-turn location information corresponding to the road segment group; if it is determined that the distance deviation between the initial new U-turn location information and the intersection node location information corresponding to the road segment group is greater than or equal to a preset distance threshold, then the initial new U-turn location information corresponding to the road segment group is used as the new U-turn location information corresponding to the road segment group.
[0136] The intersection node location information corresponding to the road segment group is the basic data in the road network topology used to describe the spatial location and attributes of road intersection nodes. It corresponds to real-world road intersection nodes and has clear spatial coordinates and connected road segment attributes. The initially added U-turn location information is a preliminary U-turn location result obtained after clustering processing. It has not yet undergone scene legality verification and may contain abnormal points caused by misjudgments of intersection turning behavior. The preset distance threshold is a pre-set spatial distance threshold used to define the proximity judgment criteria between candidate points and intersection nodes. The purpose of this step is to eliminate misjudged points adjacent to intersections and retain valid points that meet the U-turn positioning characteristics in the middle of the road segment.
[0137] For example, electronic devices can use density clustering algorithms to spatially cluster all candidate U-turn points within the same road segment group that have the same U-turn direction, generating several clusters and using the center coordinates of the highest density cluster as the initial location information of the newly added U-turn point for that road segment group; then, the coordinate data of the intersection node corresponding to that road segment group is retrieved to determine the straight-line distance between the initial newly added U-turn point of that road segment group and the corresponding intersection node, and this distance is compared with a preset 50-meter distance threshold. If the distance is greater than or equal to the preset distance threshold, the location information of the initial newly added U-turn point is used as the location information of the newly added U-turn point for that road segment group; if the distance is less than the preset distance threshold, it means that the location information of the initial newly added U-turn point is an intersection node.
[0138] Optionally, the above steps can be used to obtain the location information of newly added U-turn points in the road segment group, as well as the corresponding entry and exit points of the road segment group.
[0139] This method can accurately distinguish between U-turn behavior in the middle of a road segment and inherent turning behavior at an intersection, avoiding misjudging the trajectory features generated by normal turning at an intersection as newly added U-turn points in the middle of the road segment. It further filters out false detection results from the scene dimension, improves the accuracy of U-turn point identification in the middle of the road segment, and ensures that the output data of newly added U-turn points all meet the inclusion standards of facilities in the middle of the road segment, providing reliable support for high-quality updates of map semantic data.
[0140] S207. Enter the location information of the newly added U-turn point corresponding to the newly added recognition result into the map road network data.
[0141] For example, entering map road network data refers to the operation of writing the newly added U-turn semantic information into the map road network semantic database. The entered content includes attributes such as the spatial coordinates of the U-turn, the road segment identifier, and the direction of travel. The entered data can directly support map services such as navigation route planning and driving safety prompts.
[0142] In one example, after receiving the recognition result of a newly added U-turn exit, the electronic device writes the corresponding location information of the new U-turn exit into the U-turn exit table or road network element layer in the map road network data, and establishes the association between it and its corresponding road segment group, associated entry road segment, and exit road segment. During the writing process, the spatial coordinates, road level, direction, and effective timestamp of the new U-turn exit can be recorded simultaneously for subsequent version management and incremental updates.
[0143] The map update method provided in this application first obtains the original trajectory data of crowdsourced vehicles and road network topology information for the current statistical period. Based on the road network topology, it performs map matching on the original trajectories to generate an initial matching trajectory string. Then, it filters for high-quality matching trajectory strings through preset reliability conditions. After grouping the matching trajectory strings by road segment, it identifies trajectories with U-turn characteristics based on the time-series changes in heading and extracts candidate points. Through clustering, it obtains the location information of newly added U-turn points, and finally, it records the information of newly added U-turn points into the map road network data. This method introduces road segment groups from the road network topology as structured analysis units, filters noisy trajectories using multi-dimensional reliability conditions, and identifies U-turn behavior using the time-series characteristics of heading. It also verifies the identification through multi-trajectory clustering, effectively avoiding misjudgments caused by GPS drift, road curvature, and positioning deviations. It accurately distinguishes between real U-turn behavior and regular driving trajectories, improving the accuracy of identifying U-turn points in the middle of road segments. Simultaneously, it achieves automated collection and updating of U-turn point data based on crowdsourced trajectories, reducing the cost and cycle of manual on-site surveys and ensuring the timeliness and accuracy of map U-turn point data.
[0144] Figure 6 Flowchart of the map update method provided in this application Figure 3 ,like Figure 6 As shown, in this embodiment... Figure 1 Based on the embodiments, the map updating method is described in detail, which includes:
[0145] S301. Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period.
[0146] It should be noted that this step is similar to the aforementioned step S101, and will not be repeated here.
[0147] S302. Based on the road network topology information, perform map matching on the original trajectory data to generate an initial matching trajectory string.
[0148] It should be noted that this step is similar to the aforementioned step S202, and will not be repeated here.
[0149] S303. According to the preset reliability conditions, the initial matching trajectory string is filtered to obtain the matching trajectory string.
[0150] It should be noted that this step is similar to the aforementioned step S203, and will not be repeated here.
[0151] S304. Based on the matched trajectory string, determine the road segment traffic characteristics of each road segment in the current statistical period.
[0152] For example, road segment traffic features are used to map the trajectory behavior of the road segment into quantifiable structured features. Road segment traffic features include at least one of the following: the transition probability from entering the road segment to exiting the road segment, the total traffic flow of the entering road segment, and the total traffic flow of the road segment pair between the entering and exiting road segments. The transition probability refers to the proportion of the number of trajectories that complete a U-turn from the entering road segment to the exiting road segment within the statistical period, relative to the total number of traffic trajectories on the entering road segment, used to characterize the frequency of U-turns on that road segment. The total traffic flow of the entering road segment refers to the total number of all valid matching trajectories entering the corresponding entering road segment within the statistical period, used to characterize the overall traffic volume of the road segment. The total traffic flow of the road segment pair between the entering and exiting road segments refers to the number of trajectories that complete a U-turn from the entering road segment to the exiting road segment within the statistical period.
[0153] In one example, the electronic device can first aggregate the full matching trajectory strings according to their respective road segments, count the total number of valid trajectories for each entry road segment in the current period as the total traffic flow, and then further filter out the number of trajectories that complete the U-turn behavior from the corresponding entry road segment to the exit road segment as the total traffic flow of the road segment pair of entry and exit road segments. The transition probability of the turning pair is calculated by the ratio of the number of U-turn trajectories to the total traffic flow, and the road segment traffic characteristic data of the current statistical period is output segment by segment.
[0154] In some embodiments, electronic devices can pre-construct complete road network topology information covering the entire road network, using this road network topology information as the data basis for subsequent traffic feature statistics and location identification calculations. The road network topology information includes, but is not limited to, the following four core elements:
[0155] 1) Road segment connection relationship: Record the start node and end node corresponding to each one-way road segment, and record all adjacent road segments that are directly connected to the two ends of the road segment. Based on the above connectivity information, construct a directed reachable topology graph between all road segments.
[0156] 2) Intersection identification information: Filter road nodes where three or more road segments intersect and mark them as intersection nodes. Simultaneously record all road segments belonging to the intersection, as well as the road form (link_form) attribute corresponding to each road segment. The link_form attribute can be, for example, a straight lane, a left-turn lane, a right-turn lane, a U-turn lane, a roundabout ramp, an interchange ramp, etc.
[0157] 3) U-turn point identification: Links that carry U-turn point related identification in the link_form attribute are marked and stored separately. The marking information is used for subsequent intersection type cross_type determination and existing redundant U-turn point discovery process.
[0158] 4) Up and down separation relationship: Based on rules such as road direction, link_form, and spatial distance, identify two groups of unidirectional links that belong to the same physical road but are separated by up and down directions. This relationship is used to construct a linkgroup.
[0159] Based on the fully constructed road network topology information, it can quickly retrieve the candidate set of all downstream exit road segments with the possibility of passage for any entry road segment, without having to traverse the entire road network to complete the matching calculation, which greatly reduces the computing power overhead of data retrieval and effectively improves the overall computing efficiency of various passage features.
[0160] Furthermore, based on the aforementioned road network topology information, for each road segment, the road segment traffic characteristics can be calculated using the matching trajectory string of the current statistical period. For example, the road segment traffic characteristics can be represented as shown in Table 1.
[0161] Table 1 Traffic Characteristics of Road Sections
[0162]
[0163] The steps for extracting the trajectory point locations are as follows:
[0164] 1) in_link_point: For each entry / exit link relationship pair, extract the position of the last matching point on the entry link (i.e., the trajectory point that is about to exit the entry link), which is used to locate the starting point of the turn-off when new excavation is added at the turn-off point;
[0165] 2) pass_link_point: If there is an intermediate link between the entry link and the exit link (i.e., a non-direct connection), extract the midpoint position of the trajectory matching point of the matching trajectory string corresponding to the intermediate link, which is used to analyze the traversal path;
[0166] 3) out_link_point: Extracts the position of the first matching point on the exit link of the matching trajectory string (i.e. the trajectory point that just entered the exit link), which is used to locate the endpoint of the turn-off when adding new excavation at the turn-off point.
[0167] The transition probability reflects the likelihood of moving from one trajectory point to the next. In this embodiment, the transition probability is defined as the proportion of trajectories starting from an entry link and ultimately reaching each exit link among all trajectories leaving that entry link. Let the number of trajectories starting from entry link i and reaching exit link j be... The total number of paths from entering linki to all possible exit links is: The transition probability is then defined as: .
[0168] refer to Figure 7 , Figure 7 The schematic diagram of a trajectory transfer provided in this application shows that the number of trajectories passing through link1 is 13, of which 7 are from link1 to link2 and 3 are from link1 to link3. Therefore, the transfer probability of link1→link2 is 7 / 13=0.538; the transfer probability of link1→link3 is 3 / 13=0.230.
[0169] It should be noted that the transition probability has the following properties: (1) The sum of the transition probabilities from the same entry link to all possible exit links is 1; (2) The higher the transition probability, the more statistically significant the travel path is at the road node, and it can be used as a strong discriminative feature of road semantics (such as prohibition of turning, forced U-turn, etc.); (3) The transition probability is counted independently according to the travel date (create_time), and can support multiple modes of multi-day accumulation or single-day analysis.
[0170] By designing a structured and scalable traffic feature system, the originally discrete and difficult-to-reuse trajectory behavior information is abstracted into a unified feature model. This allows new mining requirements to be met simply by combining existing features and adjusting thresholds, significantly reducing the development cost of subsequent mining tasks. Furthermore, the road segment traffic features of this embodiment can be applied to the mining of various road semantic information, including but not limited to traffic restriction (cross-traffic restriction) recognition, intersection crossing determination, link connection mining, and U-turn recognition.
[0171] The cross-segment type identifier `cross_type` reflects the connectivity path structure from entering the link to exiting the link. Based on whether it passes through special nodes (intersections, U-turns, etc.) or whether there are trajectory breakpoints, it is divided into the following four types:
[0172] 1) cross_type=0: The continuous trajectory passes through two connected links (direct connection);
[0173] 2) cross_type=1: The continuous trajectory passes through one or more intersection links, and the entry / exit link markers are the linkpids for entering and exiting the intersection;
[0174] 3) cross_type=2: The continuous trajectory passes through U-turn points or other similar intersections;
[0175] 4) cross_type=-1: After a continuous trajectory passes through a link, there is a segment in the middle that is not matched (which may be a new road or a change in road direction).
[0176] S305. For the original U-turn point, if the traffic characteristics of the entry and exit road segments corresponding to the original U-turn point meet the preset conditions in multiple consecutive statistical periods, then the original U-turn point is determined to be a redundant U-turn point.
[0177] For example, the target original U-turn point refers to the intermediate U-turn point of the existing road segment whose validity needs to be verified. Multiple consecutive statistical periods refer to multiple consecutive independent statistical time windows, used to exclude traffic fluctuations caused by accidental factors such as construction and temporary traffic control within a single period, thus avoiding misjudgments. Preset conditions include at least one of the following: the transfer probability is less than a preset probability threshold, the total flow entering the road segment is less than a preset flow threshold, and the duration of zero total flow for the road segment pair is greater than a preset duration threshold. The preset probability threshold and preset duration threshold are pre-set critical values used to determine insufficient U-turn activity. The preset flow threshold is a pre-set critical value used to determine insufficient U-turn traffic volume. Redundant U-turn points refer to existing U-turn points that lack sufficient normalized traffic support and have actually lost their traffic function.
[0178] In one example, the electronic device can first match the corresponding entry and exit road segments based on the attribute information of the target original U-turn point, retrieve the transfer probability of the turning pair, the total traffic flow data of the entry road segment, and the duration of the total traffic flow of the road segment pair being 0 within three consecutive statistical periods. If the corresponding transfer probability is lower than the preset probability threshold of 0.5% within three consecutive periods, and the U-turn traffic flow of the entry road segment is lower than the preset traffic flow threshold, and the duration of the total traffic flow of the road segment pair being 0 is greater than 3 days, then it is determined that the target original U-turn point does not have normal traffic attributes and is identified as a redundant U-turn point.
[0179] For example, referring to Table 1, all records with cross_type=2 are selected from the traffic characteristic table for each road segment and statistically analyzed by date. A U-turn intersection is considered redundant if all of the following conditions are met:
[0180] 1) The total flow rate entering the road segment is less than the preset flow rate threshold;
[0181] 2) The probability of transitioning from entering a road segment to exiting a road segment is less than a preset probability threshold;
[0182] 3) The entry → exit link has a duration of 0 traffic that is greater than the preset duration threshold, such as greater than the preset number of days.
[0183] This processing method jointly verifies the traffic characteristics of entering and exiting road segments over multiple consecutive statistical periods. This ensures that the redundancy determination of the original U-turn point does not rely on occasional fluctuations in a single period, but rather on stable traffic statistics. When the road segment corresponding to the target original U-turn point maintains a low transfer probability, low total traffic volume, or no traffic volume for a long period, it indicates that the original U-turn point lacks continuous demand, thus forming a conclusion of U-turn point redundancy identification.
[0184] By adopting the above method, the original U-turn point can be continuously verified based on the trajectory statistics results, and short-term noise, occasional detours or low-frequency traffic factors can be excluded, thereby improving the stability and accuracy of U-turn point redundancy identification, and providing a reliable basis for the dynamic maintenance of U-turn point information in the map.
[0185] S306. Mark the original U-turn information corresponding to the redundant identification results of the U-turn intersection as invalid in the map road network data or remove it from the map road network data.
[0186] For example, marking a U-turn as invalid means adding an invalidation flag to the corresponding original U-turn point in the map road network data. This U-turn point will no longer participate in navigation route planning and driving prompt calculations, but the data record will be retained for subsequent verification. Removing from the map road network data means directly deleting the semantic data of the corresponding original U-turn point from the map road network data. The two processing methods can be flexibly selected according to the map update strategy.
[0187] In one example, electronic devices can mark redundant U-turn points as temporarily invalid in the road network semantic database of a standard precision map and push them to a manual review queue. After review and confirmation, the U-turn point can be physically deleted. Alternatively, the U-turn point information can be directly removed from the effective road network data of the online service to avoid incorrect navigation routes planning for invalid U-turn facilities.
[0188] Optionally, if the expiration marking method is used, the historical version, identification source and expiration time of the original record are retained to support retrospective query and subsequent review; if the removal method is used, the original U-turn information is removed from the current valid road network data, and its reference relationship in the navigation index, lane-level guidance index and statistical index is cleaned up at the same time.
[0189] In some embodiments, when there are both newly added and redundant U-turn exit identification results, the electronic device can first complete the input of the location information of the newly added U-turn exit, and then perform invalidation or removal processing on the corresponding redundant original U-turn exit information, so as to ensure that the updated map road network data is consistent with the actual traffic status in the current statistical period. It should be noted that the embodiments of this application do not limit the order of inputting the location information of the newly added U-turn exit and performing invalidation or removal processing on the redundant original U-turn exit information.
[0190] This update mechanism transforms newly identified U-turn intersections into directly writable road network elements and redundant U-turn intersection identification results into validity status changes, enabling U-turn intersection information to be incrementally synchronized to the map database. By simultaneously supporting the addition of new U-turn intersections (based on heading angle deviation and clustering) and redundancy detection (based on traffic flow and transition probability), it can cover the entire lifecycle management of U-turn intersections. Therefore, the coverage of U-turn intersections in the map can continuously expand with the identification results, and previously invalid U-turn intersections can be promptly suppressed or removed, thus maintaining a correspondence between road network data and actual road traffic conditions and reducing the reliance of subsequent navigation planning on erroneous U-turn intersection information.
[0191] The map update method provided in this application first obtains the original trajectory data of crowdsourced vehicles in the current statistical period and the road network topology information including existing original U-turn points. Based on the road network topology, the original trajectories are matched and the reliability is screened to obtain high-quality matching trajectory strings. Then, based on the matching trajectory strings, the traffic characteristics of each road segment are statistically obtained. Combining the traffic data of multiple consecutive statistical periods, the validity of existing original U-turn points is checked one by one. Original U-turn points that do not meet the traffic conditions are judged as redundant U-turn points. Finally, the redundant U-turn points are marked as invalid or removed from the data to complete the iterative update of the map's existing data. This method relies on routine traffic data from crowdsourced trajectories to quantify the actual traffic activity of U-turn points from two dimensions: turning probability and road segment traffic flow. Combined with multi-period continuous verification to avoid misjudgments caused by accidental factors such as temporary control and construction, it can accurately identify redundant U-turn points that have actually failed, replacing the traditional mode of manual on-site verification. This improves the verification efficiency and accuracy of existing U-turn point data, promptly cleans up invalid road traffic semantic information in the map, and ensures the rationality and accuracy of navigation route planning. At the same time, it complements the new U-turn point discovery mechanism to achieve automated updates and maintenance of map U-turn point data throughout its entire lifecycle.
[0192] In some embodiments, Figure 8 Flowchart of the map update method provided in this application Figure 4 , Figure 9 Flowchart of the map update method provided in this application Figure 5 ,like Figure 8 and Figure 9As shown, the method includes the following steps:
[0193] S401. Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period.
[0194] S402. Based on the road network topology information, perform map matching on the original trajectory data to generate an initial matching trajectory string.
[0195] S403. According to the preset reliability conditions, the initial matching trajectory string is filtered to obtain the matching trajectory string.
[0196] S404. For each road segment group, determine the matching trajectory string belonging to that road segment group from the matching trajectory string.
[0197] S405. For each matching trajectory string belonging to the road segment group, if it is determined that the matching trajectory string has U-turn characteristics based on the heading time sequence change information of the matching trajectory string, then a candidate point for the U-turn is determined based on the matching trajectory string.
[0198] S406. Cluster the candidate points of the U-turn exits to obtain the location information of the newly added U-turn exits corresponding to the road segment group.
[0199] S407. Enter the location information of the newly added U-turn point corresponding to the newly added recognition result into the map road network data.
[0200] S408. Based on the matched trajectory string, determine the road segment traffic characteristics of each road segment in the current statistical period.
[0201] S409. For the original U-turn point, if the traffic characteristics of the entry and exit road segments corresponding to the original U-turn point meet the preset conditions in multiple consecutive statistical periods, then the original U-turn point is determined to be a redundant U-turn point.
[0202] S410. Mark the original U-turn information corresponding to the redundant identification results of the U-turn intersection as invalid in the map road network data or remove it from the map road network data.
[0203] Specifically, within the current statistical period, the raw trajectory data and road network topology information of crowdsourced vehicles are first acquired. The road network topology information includes multiple road segment groups, original U-turn information, and intersection node information. Each road segment group contains multiple road segments in the up and down directions of the same physical road. Based on the road network topology information, map matching is performed on the raw trajectory data to generate an initial matched trajectory string. Each trajectory point in the initial matched trajectory string is associated with a corresponding matched road segment. Subsequently, the initial matched trajectory string is filtered according to preset reliability conditions, retaining the trajectory strings that meet the conditions as the matched trajectory strings. The preset reliability conditions include at least one of the following: road segment connectivity conditions, heading deviation threshold conditions, and road segment length deviation conditions. The road segment connectivity condition requires that the road segment sequence corresponding to the trajectory string contains consecutive entry and exit road segments, and also includes the preceding adjacent road segment of the entry segment and the following adjacent road segment of the exit segment.
[0204] When using a heading deviation threshold for filtering, the difference between the heading angle of the trajectory string corresponding to the preceding adjacent road segment and the heading angle of the preceding adjacent road segment are calculated, as are the differences between the heading angles of the trajectory string corresponding to the following adjacent road segment. When both differences are less than a preset angle threshold, the trajectory string meets the heading consistency requirement. When using a road segment length deviation condition for filtering, the proportion of the difference between the length of the trajectory string corresponding to the preceding adjacent road segment and the length of the preceding adjacent road segment, as well as the proportion of the difference between the length of the trajectory string corresponding to the following adjacent road segment and the length of the following adjacent road segment, are calculated. When both proportions are less than a preset ratio threshold, the trajectory string meets the length matching requirement. The matched trajectory strings obtained after filtering are used for new U-turn point identification and redundancy verification of existing U-turn points.
[0205] In one possible implementation, for each road segment group, matching trajectory strings belonging to that road segment group are extracted from the matching trajectory strings, and the heading time-series change information of each matching trajectory string is analyzed one by one. When the heading time-series change information indicates that the vehicle has U-turn characteristics, candidate points for U-turns are determined based on the matching trajectory strings. In the determination process, each trajectory point in the matching trajectory string is first projected onto its corresponding road segment to obtain multiple projection points. Then, the projection point farthest from the starting position of the corresponding entry road segment of the matching trajectory string is selected from these projection points as the candidate point for the U-turn corresponding to the matching trajectory string. The multiple candidate points for U-turns obtained within the same road segment group are clustered to obtain the initial new U-turn location information corresponding to the road segment group. Then, a filtering process is performed in combination with the intersection node information to remove the initial new U-turns whose distance deviation from the intersection node position is less than a preset range, and the remaining positions are retained as the new U-turn location information corresponding to the road segment group.
[0206] Simultaneously, redundancy checks are performed on the original U-turn exit information using the matching trajectory strings. First, the traffic characteristics of each road segment within the current statistical period are statistically analyzed based on the matching trajectory strings. These characteristics include at least one of the following: the probability of transitioning from an entry segment to an exit segment, and the total traffic flow of the entry segment. Then, for each target original U-turn exit, the traffic characteristics of the entry and exit segments corresponding to that target original U-turn exit are continuously tracked over multiple statistical periods. When preset conditions are met for multiple consecutive statistical periods, the target original U-turn exit is identified as a redundant U-turn exit. These preset conditions include at least one of the following: the probability of transitioning is less than a preset probability threshold, and the total traffic flow is less than a preset traffic flow threshold. After obtaining the identification results for newly added U-turn exits and redundant U-turn exits, the location information of the newly added U-turn exits is entered into the map road network data, and the original U-turn exit information corresponding to the redundant U-turn exits is marked as invalid or directly removed from the map road network data.
[0207] The map update method provided in this application first obtains the original trajectory data of crowdsourced vehicles and road network topology information including road segment groups and original U-turn information within the current statistical period. Based on the road network topology, the original trajectories are matched with the map to generate an initial matching trajectory string. After screening with preset reliability conditions, high-quality matching trajectory strings are obtained. On the one hand, the matching trajectory strings are collected by road segment groups, and trajectories with U-turn characteristics are identified based on the time-series changes in heading, and candidate points are extracted. After clustering processing, the location information of newly added U-turn points is obtained and entered into the map road network data. On the other hand, the road segment traffic characteristics of the current statistical period are statistically obtained based on the matching trajectory strings. The validity of existing original U-turn points is checked one by one by combining the traffic data of multiple consecutive statistical periods. Original U-turn points that continuously meet the preset conditions are judged as redundant U-turn points, and failure marking or data removal operations are performed on them. This solution introduces a road network topology divided into road segment groups as a structured analysis constraint, transforming the original geometric recognition based on discrete raw trajectories into a structured analysis based on road units. This effectively filters trajectory noise interference caused by GPS drift, road curvature, and building obstruction, accurately distinguishes between U-turn behavior in the middle of a road segment and intersection turning and regular driving trajectories, and improves the accuracy of U-turn identification. At the same time, the newly added bidirectional parallel mechanism of mining and redundancy verification can simultaneously complete the addition of new semantic data and the cleanup of invalid semantic data, realizing automated high-frequency updates of U-turn data throughout its entire lifecycle. This significantly reduces the cost and cycle of manual on-site surveys and verifications, ensuring the timeliness and accuracy of map U-turn semantic data, and providing reliable data support for map services such as navigation route planning and driving safety prompts.
[0208] Figure 10 A schematic diagram of the structure of the map updating device provided in this application is shown below. Figure 10 As shown, the map updating device 500 provided in this embodiment includes:
[0209] The acquisition module 501 is used to acquire the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period; wherein, the road network topology information includes multiple road segment groups and original U-turn information, and each road segment group includes multiple road segments in the up and down directions of the same physical road;
[0210] The processing module 502 is used to perform map matching and trajectory filtering processing on the original trajectory data based on the road network topology information to obtain a matching trajectory string; wherein, the trajectory points in the matching trajectory string are associated with the matched road segments;
[0211] The identification module 503 is used to determine the newly added identification results of the U-turn entrances corresponding to each road segment group based on the matching trajectory strings belonging to each road segment group; and / or to perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the redundant identification results of the U-turn entrances.
[0212] The update module 504 is used to update the map based on the newly added recognition results of the U-turn and / or the redundant recognition results of the U-turn.
[0213] In one possible implementation, the processing module 502 is configured to:
[0214] Based on the road network topology information, map matching is performed on the original trajectory data to generate an initial matching trajectory string; where each trajectory point in the initial matching trajectory string is associated with the matched road segment;
[0215] According to preset reliability conditions, the initial matching trajectory string is filtered to obtain the matching trajectory string; wherein, the reliability conditions include at least one of the following: road segment connectivity condition, heading deviation threshold condition, and road segment length deviation condition.
[0216] In one possible implementation, the road segment connectivity condition is: the road segment sequence corresponding to the trajectory string contains consecutive entry road segments and exit road segments, and includes the preceding adjacent road segment of the entry road segment and the following adjacent road segment of the exit road segment.
[0217] The heading deviation threshold condition is: the difference between the heading angle of the trajectory string corresponding to the previous adjacent road segment and the heading angle of the previous adjacent road segment is less than the preset angle threshold, and the difference between the heading angle of the trajectory string corresponding to the next adjacent road segment and the heading angle of the next adjacent road segment is less than the preset angle threshold.
[0218] The road segment length deviation condition is: the difference between the length of the trajectory string corresponding to the previous adjacent road segment and the length of the previous adjacent road segment is less than a preset percentage threshold, and the difference between the length of the trajectory string corresponding to the next adjacent road segment and the length of the next adjacent road segment is less than a preset percentage threshold.
[0219] In one possible implementation, the identification module 503 is used for:
[0220] For each road segment group, determine the matching trajectory string belonging to that road segment group from the matching trajectory string;
[0221] For each matching trajectory string belonging to the road segment group, if it is determined that the matching trajectory string has U-turn characteristics based on the heading time sequence change information of the matching trajectory string, then a candidate point for the U-turn is determined based on the matching trajectory string.
[0222] Clustering is performed on the candidate points of the U-turn to obtain the location information of the new U-turn points corresponding to the road segment group.
[0223] In one possible implementation, the identification module 503 is used for:
[0224] Project each trajectory point in the matched trajectory string onto the corresponding road segment to obtain each projection point;
[0225] The projection point furthest from the starting point of the corresponding entry segment of the matched trajectory string is selected as the candidate point for the U-turn.
[0226] In one possible implementation, the road network topology information also includes the location information of intersection nodes corresponding to road segment groups; the identification module 503 is used for:
[0227] Clustering is performed on the candidate points of the U-turn to obtain the initial location information of the newly added U-turns corresponding to the road segment group;
[0228] If the distance deviation between the initial newly added U-turn location information and the intersection node location information corresponding to the road segment group is greater than or equal to a preset distance threshold, then the initial newly added U-turn location information corresponding to the road segment group is used as the newly added U-turn location information corresponding to the road segment group.
[0229] In one possible implementation, the identification module 503 is used for:
[0230] Based on the matched trajectory string, determine the traffic characteristics of each road segment in the current statistical period; the traffic characteristics include at least one of the following: the transition probability from entering the road segment to exiting the road segment, the total traffic flow of the entering road segment, and the total traffic flow of the road segment pair between the entering and exiting road segments;
[0231] For a target original U-turn point, if the traffic characteristics of the entry and exit road segments corresponding to the target original U-turn point meet preset conditions within multiple consecutive statistical periods, then the target original U-turn point is determined to be a redundant U-turn point. The preset conditions include at least one of the following: the transfer probability is less than a preset probability threshold, the total flow of the entry road segment is less than a preset flow threshold, and the duration of the total flow of the road segment pair being 0 is greater than a preset duration threshold.
[0232] In one possible implementation, the update module 504 is used to:
[0233] Enter the location information of the newly added U-turn point corresponding to the newly identified U-turn point into the map road network data; and / or,
[0234] The original U-turn information corresponding to the redundant identification results of the U-turn point will be marked as invalid in the map road network data or removed from the map road network data.
[0235] The map updating device provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.
[0236] Figure 11 A schematic diagram of the structure of the electronic device provided in this application. Figure 11 As shown, the electronic device 600 provided in this embodiment includes at least one processor 601 and a memory 602. Optionally, the electronic device 600 further includes a communication component 603. The processor 601, memory 602, and communication component 603 are connected via a bus.
[0237] In a specific implementation, at least one processor 601 executes computer execution instructions stored in memory 602, causing at least one processor 601 to perform the above-described method.
[0238] The specific implementation process of processor 601 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.
[0239] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.
[0240] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage device.
[0241] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.
[0242] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.
[0243] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-described method.
[0244] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.
[0245] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in the device.
[0246] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.
[0247] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.
[0248] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.
[0249] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0250] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.
[0251] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.
Claims
1. A method for updating a map, characterized in that, include: Obtain the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period; wherein, the road network topology information includes multiple road segment groups and original U-turn information, and each road segment group includes multiple road segments in the up and down directions of the same physical road; Based on the road network topology information, the original trajectory data is subjected to map matching and trajectory filtering processing to obtain a matching trajectory string; wherein, the trajectory points in the matching trajectory string are associated with the matched road segments; Based on the matching trajectory strings belonging to each road segment group, determine the newly added identification results of the U-turn entrances corresponding to each road segment group; and / or, perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the redundant identification results of the U-turn entrances. Update the map based on the newly added identification results of the U-turn point and / or the redundant identification results of the U-turn point.
2. The method according to claim 1, characterized in that, The step of performing map matching and trajectory filtering on the original trajectory data based on the road network topology information to obtain a matched trajectory string includes: Based on the road network topology information, map matching is performed on the original trajectory data to generate an initial matching trajectory string; wherein, each trajectory point in the initial matching trajectory string is associated with the matched road segment; According to preset reliability conditions, the initial matching trajectory string is filtered to obtain a matching trajectory string; wherein, the reliability conditions include at least one of road segment connectivity conditions, heading deviation threshold conditions, and road segment length deviation conditions.
3. The method according to claim 2, characterized in that, The road segment connectivity condition is: the road segment sequence corresponding to the trajectory string contains consecutive entry road segments and exit road segments, and includes the previous adjacent road segment of the entry road segment and the next adjacent road segment of the exit road segment; The heading deviation threshold condition is as follows: the difference between the heading angle of the trajectory string corresponding to the previous adjacent road segment and the heading angle of the previous adjacent road segment is less than a preset angle threshold, and the difference between the heading angle of the trajectory string corresponding to the next adjacent road segment and the heading angle of the next adjacent road segment is less than a preset angle threshold. The road segment length deviation condition is as follows: the difference between the length of the trajectory string corresponding to the previous adjacent road segment and the length of the previous adjacent road segment is less than a preset percentage threshold, and the difference between the length of the trajectory string corresponding to the next adjacent road segment and the length of the next adjacent road segment is less than a preset percentage threshold.
4. The method according to claim 1, characterized in that, The step of determining the newly added identification results of the U-turn entrances corresponding to each road segment group based on the matching trajectory strings belonging to each road segment group includes: For each road segment group, the matching trajectory string belonging to that road segment group is determined from the matching trajectory string; For each matching trajectory string belonging to the road segment group, if it is determined that the matching trajectory string has U-turn characteristics based on the heading time sequence change information of the matching trajectory string, then a candidate point for the U-turn is determined based on the matching trajectory string. Clustering is performed on the candidate points of the U-turn to obtain the location information of the newly added U-turn points corresponding to the road segment group.
5. The method according to claim 4, characterized in that, The step of determining candidate points for U-turns based on the matched trajectory string includes: Project each trajectory point in the matched trajectory string onto the corresponding road segment to obtain each projection point; The projection point furthest from the starting point of the corresponding entry segment of the matched trajectory string is selected as the candidate point for the U-turn.
6. The method according to claim 4, characterized in that, The road network topology information also includes the location information of intersection nodes corresponding to the road segment group; the clustering process of the candidate U-turn points to obtain the location information of newly added U-turn points corresponding to the road segment group includes: Clustering is performed on the candidate U-turn points to obtain the initial location information of newly added U-turn points corresponding to the road segment group; If the distance deviation between the initial newly added U-turn location information and the intersection node location information corresponding to the road segment group is determined to be greater than or equal to a preset distance threshold, then the initial newly added U-turn location information corresponding to the road segment group is taken as the newly added U-turn location information corresponding to the road segment group.
7. The method according to claim 1, characterized in that, The step of performing redundancy verification on the original U-turn information based on the matched trajectory string to obtain the U-turn redundancy identification result includes: Based on the matching trajectory string, the traffic characteristics of each road segment in the current statistical period are determined; the traffic characteristics include at least one of the following: the transition probability from entering the road segment to exiting the road segment, the total traffic flow of the entering road segment, and the total traffic flow of the road segment pair between the entering and exiting road segments; For a target original U-turn point, if the traffic characteristics of the entry and exit segments corresponding to the target original U-turn point meet preset conditions within multiple consecutive statistical periods, then the target original U-turn point is determined to be a redundant U-turn point; wherein, the preset conditions include at least one of the following: the transfer probability is less than a preset probability threshold, the total flow of the entry segment is less than a preset flow threshold, and the duration of the total flow of the segment pair being 0 is greater than a preset duration threshold.
8. The method according to any one of claims 1-7, characterized in that, The step of updating the map based on the newly added identification results of the U-turn and / or the redundant identification results of the U-turn includes: The location information of the newly added U-turn point corresponding to the newly identified U-turn point is entered into the map road network data; and / or, The original U-turn information corresponding to the redundant identification results of the U-turn point is marked as invalid or removed from the map road network data.
9. A map updating device, characterized in that, include: The acquisition module is used to acquire the original trajectory data and road network topology information of crowdsourced vehicles within the current statistical period; wherein, the road network topology information includes multiple road segment groups and original U-turn information, and each road segment group includes multiple road segments in the up and down directions of the same physical road; The processing module is used to perform map matching and trajectory filtering processing on the original trajectory data according to the road network topology information to obtain a matching trajectory string; wherein, the trajectory points in the matching trajectory string are associated with the matched road segments; The identification module is used to determine the newly added identification results of the U-turn entrances corresponding to each road segment group based on the matching trajectory strings belonging to each road segment group; and / or to perform redundancy verification on the original U-turn entrance information based on the matching trajectory strings to obtain the U-turn entrance redundancy identification results. The update module is used to update the map based on the newly added identification results of the U-turn and / or the redundant identification results of the U-turn.
10. An electronic device / computer-readable storage medium / computer program product, characterized in that, The electronic device includes: a memory and a processor; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-8; The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-8; The computer program product includes a computer program that, when executed by a processor, is used to implement the method as described in any one of claims 1-8.
