Method, device and medium for constructing electronic map

Abstract

The present disclosure provides a method and device for constructing an electronic map, and a medium, relating to the technical field of computers, and in particular to the technical fields of electronic maps, high-definition maps, intelligent transportation, and cloud services. The specific implementation scheme is as follows: determining a target road associated with a candidate trajectory point according to a candidate trajectory point position of the candidate trajectory point; determining at least one candidate trajectory line segment according to the candidate trajectory point, and determining a line segment direction of the candidate trajectory line segment and a road direction of the target road; determining a target trajectory line segment from the candidate trajectory line segment according to the line segment direction, the road direction, and a road feature of the target road; and constructing a pedestrian crossing in an electronic map corresponding to the target road according to the target trajectory line segment. The present disclosure achieves the effect of constructing a pedestrian crossing in an electronic map based on a trajectory point, thereby eliminating the need to completely rely on collected images of the pedestrian crossing, and improving the accuracy and redundancy of constructing a pedestrian crossing in an electronic map.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and more particularly to the fields of electronic maps, high-precision maps, intelligent transportation and cloud services, and especially to a method, apparatus, device and medium for constructing electronic maps. Background Technology

[0002] With the development of technology, electronic maps have gradually replaced paper maps and become a necessity for people's travel. Whether it is vehicle navigation or autonomous driving, electronic maps are required to achieve the desired results.

[0003] Currently, the construction of electronic maps usually relies on real-scene imagery. For example, the construction of pedestrian crossings in electronic maps requires real-scene images of pedestrian crossings. Summary of the Invention

[0004] This disclosure provides a method, apparatus, electronic device, and medium for constructing pedestrian crossings in an electronic map without the need for image acquisition.

[0005] According to one aspect of this disclosure, a method for constructing an electronic map is provided, comprising:

[0006] Based on the location of the candidate trajectory points, determine the target road associated with the candidate trajectory points;

[0007] Based on the candidate trajectory points, at least one candidate trajectory line segment is determined, and the line segment direction of the candidate trajectory line segment and the road direction of the target road are determined.

[0008] The target trajectory segment is determined from the candidate trajectory segments based on the line segment direction, the road direction, and the road characteristics of the target road.

[0009] Based on the target trajectory line segment, a pedestrian crossing is constructed on the electronic map corresponding to the target road.

[0010] According to another aspect of this disclosure, an apparatus for constructing an electronic map is provided, comprising:

[0011] The target road determination module is used to determine the target road associated with the candidate trajectory points based on the candidate trajectory point positions.

[0012] The direction determination module is used to determine at least one candidate trajectory line segment based on the candidate trajectory points, and to determine the line segment direction of the candidate trajectory line segment and the road direction of the target road;

[0013] The trajectory segment determination module is used to determine the target trajectory segment from the candidate trajectory segments based on the segment direction, the road direction, and the road characteristics of the target road.

[0014] The map building module is used to construct pedestrian crossings on the electronic map corresponding to the target road based on the target trajectory line segments.

[0015] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0016] At least one processor; and

[0017] A memory that is communicatively connected to at least one processor; wherein,

[0018] The memory stores instructions that can be executed by at least one processor to enable the at least one processor to perform any of the methods of this disclosure.

[0019] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause a computer to perform any of the methods of this disclosure.

[0020] According to another aspect of this disclosure, a computer program product is provided, including a computer program and a method for the computer program to be executed by a processor according to any of the methods disclosed herein.

[0021] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0022] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0023] Figure 1A This is a schematic diagram of some marked pedestrian crossings disclosed in the embodiments of this disclosure;

[0024] Figure 1B This is a schematic diagram of some unmarked pedestrian crossings disclosed in the embodiments of this disclosure;

[0025] Figure 1C This is a flowchart of some electronic map construction methods disclosed in embodiments of this disclosure;

[0026] Figure 1D This is a schematic diagram of a scenario where some pedestrian crossings are constructed according to embodiments of this disclosure;

[0027] Figure 2AThis is a flowchart of another method for constructing electronic maps according to embodiments of this disclosure;

[0028] Figure 2B This is a schematic diagram of some trajectory line segment types disclosed in the embodiments of this disclosure;

[0029] Figure 3 This is a flowchart of some model training methods disclosed in the embodiments of this disclosure;

[0030] Figure 4 This is a schematic diagram of the structure of some electronic map construction apparatuses disclosed in embodiments of this disclosure;

[0031] Figure 5 This is a block diagram of an electronic device used to implement the electronic map construction method disclosed in the embodiments of this disclosure. Detailed Implementation

[0032] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0033] Pedestrian crossings are a crucial piece of map data in electronic maps. For example, in vehicle navigation, when a pedestrian crossing is detected ahead on the map, the system can remind the driver to slow down in advance. Similarly, in autonomous driving, when a pedestrian crossing is detected ahead on the map, the system can actively slow down the vehicle. This ensures maximum safety for both vehicles and pedestrians when crossing pedestrian crossings.

[0034] Currently, the construction of pedestrian crossings in electronic maps usually relies on real-world images of pedestrian crossings. For example, images of marked pedestrian crossings are captured by cameras mounted on a data collection vehicle, and the pose information of the pedestrian crossings is determined based on data collected by the vehicle's GPS (Global Positioning System) and IMU (Inertial Measurement Unit). Thus, based on the pose information and images of the marked pedestrian crossings, the pedestrian crossings are constructed in the electronic map. Figure 1A This is a schematic diagram of some marked pedestrian crossings disclosed in the embodiments of this disclosure.

[0035] However, in the real world, there are many unmarked crosswalks that are actually used by a large number of pedestrians. Figure 1BThese are schematic diagrams of some unmarked pedestrian crossings disclosed in the embodiments of this disclosure, such as... Figure 1B As shown, although the pedestrian crossing is not marked, pedestrians are crossing it.

[0036] This demonstrates that existing methods cannot construct pedestrian crossings on electronic maps without markings, which undoubtedly reduces the accuracy and redundancy of electronic map construction. Furthermore, in vehicle navigation or autonomous driving scenarios, due to the low accuracy of electronic map construction, when a vehicle crosses an unmarked pedestrian crossing, it cannot be promptly alerted or its speed cannot be reduced, posing a significant safety hazard.

[0037] Figure 1C This is a flowchart of some electronic map construction methods disclosed in embodiments of this disclosure. This embodiment can be applied to the construction of pedestrian crossings in electronic maps. The method of this embodiment can be executed by the electronic map construction apparatus disclosed in embodiments of this disclosure. The apparatus can be implemented in software and / or hardware and can be integrated into any electronic device with computing capabilities.

[0038] like Figure 1C As shown, the method for constructing an electronic map disclosed in this embodiment may include:

[0039] S101. Based on the location of the candidate trajectory points, determine the target road associated with the candidate trajectory points.

[0040] In this context, candidate trajectory points refer to pedestrian trajectory points in the electronic map, which are collected by monitoring the location of pedestrians at preset time intervals. In this embodiment, pedestrian trajectory points collected within a preset time period are selected as candidate trajectory points, for example, pedestrian trajectory points collected within one month are selected as candidate trajectory points. Each candidate trajectory point is associated with, but is not limited to, collection time, collection location, and trajectory point identifier. The unit of collection time can be selected as seconds; the collection location can include the longitude and latitude information of the candidate trajectory point; the trajectory point identifier is used to indicate the type of candidate trajectory point, with different types of candidate trajectory points representing different pedestrians. It is easy to understand that the candidate trajectory points in this embodiment are all collected with the authorization of the pedestrians.

[0041] In one implementation, the information associated with each candidate trajectory point is traversed and parsed to determine the collection location associated with each candidate trajectory point, which is then used as the candidate trajectory point location. Furthermore, the candidate road locations for each candidate road are determined. Here, candidate roads represent roads included in the electronic map. The candidate road location can be represented by the set of locations of the candidate road's coverage points, or by the location of the center point of the candidate road.

[0042] Based on the candidate trajectory point positions of each candidate trajectory point and the candidate road positions of each candidate road, determine whether any candidate trajectory point is located on any candidate road or whether the distance between it and any candidate road is less than a preset threshold. If so, the candidate road is used as the target road associated with the candidate trajectory point; otherwise, the candidate trajectory point is removed.

[0043] For example, suppose the candidate trajectory point A has a candidate trajectory point position of "120,120" and the candidate road 1 has a candidate road position of "120,120", then it means that the candidate trajectory point A is on the candidate road 1, and the candidate road 1 is then used as the target road associated with the candidate trajectory point A.

[0044] For example, suppose the candidate trajectory point B has a candidate trajectory point position of "100,100", and no candidate road has a candidate road position including "100,100", then it means that the candidate trajectory point B is not on any candidate road. However, if the distance between candidate road 2 and candidate trajectory point B is 20 meters, which is less than the preset threshold of 50 meters, then candidate road 2 will be regarded as the target road associated with candidate trajectory point B.

[0045] By determining the target road associated with the candidate trajectory point based on the candidate trajectory point's location, the effect of determining the target road that is close to the candidate trajectory point is achieved, laying the foundation for subsequently determining the road direction of the target road.

[0046] S102. Determine at least one candidate trajectory line segment based on the candidate trajectory points, and determine the line segment direction of the candidate trajectory line segment and the road direction of the target road.

[0047] Each candidate trajectory segment consists of at least two candidate trajectory points. The direction of the candidate trajectory segment indicates its orientation on the electronic map. The direction of the target road indicates its orientation on the electronic map.

[0048] In one implementation, for any candidate trajectory point associated with a target road, the candidate trajectory points are segmented according to their positional order to obtain at least one candidate trajectory segment. Each candidate trajectory segment consists of a preset number of candidate trajectory points. The preset number can be selected as four, meaning each candidate trajectory segment consists of four candidate trajectory points.

[0049] Based on the orientation of each candidate trajectory line segment in the electronic map, the line segment direction of each candidate trajectory line segment is determined, and based on the orientation of the target road in the electronic map, the road direction of the target road is determined.

[0050] Optionally, both orientations of the candidate trajectory line segment in the electronic map can be used as the line segment direction of the candidate trajectory line segment.

[0051] Optionally, when the target road is a one-way street, the only direction of the target road is taken as the road direction, and when the target road is a two-way street, both directions of the target road are taken as the road direction.

[0052] For example, suppose target road A is associated with candidate trajectory points 1, 2, ..., 12. Using a preset number of four, candidate trajectory points 1, 2, ..., 12 are divided into candidate trajectory segments 1, 2, 3, and 4. Then, the line direction of each of the candidate trajectory segments 1, 2, 3, and 4, as well as the road direction of target road A, are determined.

[0053] By determining at least one candidate trajectory segment based on candidate trajectory points, and determining the segment direction of the candidate trajectory segment and the road direction of the target road, the data preparation effect was achieved, laying a data foundation for subsequently determining the target trajectory segment based on the segment direction and road direction.

[0054] S103. Determine the target trajectory line segment from the candidate trajectory line segments based on the line segment direction, road direction, and road characteristics of the target road.

[0055] Among them, road features represent the attribute characteristics of the target road, used to reflect whether the target road has pedestrian crossing conditions, or whether it is close to building entrances and exits, etc. Road features include, but are not limited to, road speed limits, road width, and whether there are entrances and exits, etc.

[0056] In one implementation, the road features of the target road are first obtained from the database of the electronic map, and then compared with preset standard road features. If it is determined that the road features of the target road do not belong to the standard road features, it means that the target road does not have pedestrian crossing conditions, and there is no need to construct a pedestrian crossing in the electronic map. The standard road features are constructed based on the road features of roads with pedestrian crossing conditions.

[0057] When the road features of the target road are determined to be standard road features, it means that the target road has pedestrian crossing conditions. Then, based on the line direction of each candidate trajectory line segment associated with the target road and the road direction of the target road, the angle information between each candidate trajectory line segment and the target road is determined.

[0058] The included angle information is compared with the preset standard included angle information. If the included angle information between any candidate trajectory segment and the target road is determined to be different from the standard included angle information, then the candidate trajectory segment is not the target trajectory segment. If the included angle information between any candidate trajectory segment and the target road is determined to be different from the standard included angle information, then the candidate trajectory segment is the target trajectory segment.

[0059] For example, assuming the standard included angle information is 70° to 110°, when the included angle information between any candidate trajectory segment and the target road is within 70° to 110°, it means that the candidate trajectory segment is likely the trajectory segment generated by a pedestrian crossing the target road, and thus the candidate trajectory segment is taken as the target trajectory segment; correspondingly, when the included angle information between any candidate trajectory segment and the target road is not within 70° to 110°, it means that the candidate trajectory segment is likely the trajectory segment of a pedestrian traveling along the target road.

[0060] In another implementation, the road features of the target road are acquired, and the angle information between each candidate trajectory segment and the target road is determined based on the segment direction of each candidate trajectory segment associated with the target road and the road direction of the target road. The angle information and road features are then input into a pre-trained binary classification model to predict whether each candidate trajectory segment is the target trajectory segment. For example, when the binary classification model outputs "1", it indicates that the candidate trajectory segment is the target trajectory segment, meaning the candidate trajectory segment is likely a trajectory segment generated by a pedestrian crossing the target road; when the binary classification model outputs "0", it indicates that the candidate trajectory segment is not the target trajectory segment, meaning the target road does not have the conditions for pedestrian crossing, or the candidate trajectory segment is likely a trajectory segment generated by a pedestrian traveling along the target road.

[0061] By determining the target trajectory segment from candidate trajectory segments based on the direction of the line segment, the direction of the road, and the road characteristics of the target road, the system achieves the effect of automatically selecting the target trajectory segment from the candidate trajectory segments without manual intervention, saving labor costs and improving efficiency. Furthermore, it lays a data foundation for the subsequent construction of electronic maps based on the target trajectory segments.

[0062] S104. Construct pedestrian crossings on the electronic map corresponding to the target road based on the target trajectory line segments.

[0063] In one implementation, since the target trajectory segment is likely to be the trajectory segment generated by a pedestrian crossing the target road, a pedestrian crossing is constructed in the electronic map corresponding to the target road based on the candidate trajectory point positions included in the target trajectory segment.

[0064] Figure 1DThese are schematic diagrams illustrating scenarios of pedestrian crossing construction based on embodiments of this disclosure, such as... Figure 1D As shown in the figure, the trajectory points in the figure are all candidate trajectory points included in the target trajectory line segment. Then, based on the area covered by the candidate trajectory point positions in the electronic map, the pedestrian crossing is constructed in the electronic map corresponding to the target road.

[0065] This disclosure determines the target road associated with the candidate trajectory points based on their locations, identifies at least one candidate trajectory segment based on the candidate trajectory points, determines the segment direction of the candidate trajectory segment and the road direction of the target road, and determines the target trajectory segment from the candidate trajectory segments based on the segment direction, road direction, and road characteristics of the target road. Pedestrian crossings are then constructed on the electronic map corresponding to the target road based on the target trajectory segment. This achieves the effect of constructing pedestrian crossings on electronic maps based on trajectory points, thus eliminating the need to rely entirely on images of the pedestrian crossings. This allows for the construction of both marked and unmarked pedestrian crossings in the real world on the electronic map, improving the accuracy and redundancy of pedestrian crossing construction on electronic maps. Furthermore, for vehicle navigation or autonomous driving scenarios, the improved accuracy of pedestrian crossing construction on electronic maps enables timely reminders or control of vehicle speed reduction, lowering the probability of traffic accidents near pedestrian crossings.

[0066] Based on the above embodiments, optionally, before S101, the method further includes:

[0067] Determine the initial trajectory point location and the target area location; wherein, the target area includes building areas and / or closed park areas; remove initial trajectory points whose initial trajectory point locations are within the target area location, and use the remaining initial trajectory points as candidate trajectory points.

[0068] The initial trajectory point represents the original pedestrian trajectory point without any data filtering.

[0069] In one implementation, the collection location of each initial trajectory point is determined as the initial trajectory point location, and the location of building areas and closed park areas in the electronic map is determined. The initial trajectory point locations are matched with the location of the area, and based on the matching result, initial trajectory points whose locations are within the location of the area are eliminated, and the remaining initial trajectory points are used as candidate trajectory points.

[0070] By determining the initial trajectory point location and the target area location (including building areas and / or closed park areas), initial trajectory points located within the target area are eliminated, and the remaining initial trajectory points are used as candidate trajectory points. This achieves the effect of data filtering for initial trajectory points located within building areas and / or closed park areas, ensuring the data reliability of the remaining candidate trajectory points.

[0071] Based on the above embodiments, optionally, before S101, the following steps A, B, and C are further included:

[0072] A. Determine the type and number of initial trajectory points contained in the candidate grid area based on the trajectory point identifiers of the initial trajectory points contained in the candidate grid area; wherein, the candidate grid area is obtained by dividing the electronic map into grids.

[0073] The trajectory point identifier is used to indicate the type of the initial trajectory point. Different types of initial trajectory points indicate that they belong to different pedestrians.

[0074] In one implementation, the electronic map is divided into equidistant grids to obtain at least two candidate grid regions in the electronic map. The trajectory point identifiers of the initial trajectory points contained in each candidate grid region are determined, and the type and number of initial trajectory points contained in each candidate grid region are determined based on the trajectory point identifiers.

[0075] For example, suppose candidate grid region 1 contains initial trajectory point 1, initial trajectory point 2, ..., initial trajectory point 10, wherein the trajectory point identifiers of initial trajectory points 1 to 4 are "00001", the trajectory point identifiers of initial trajectory points 5 to 8 are "00002", and the trajectory point identifiers of initial trajectory points 9 to 10 are "00003", then the number of types of initial trajectory points contained in candidate grid region 1 is determined to be 3.

[0076] B. Select candidate grid regions with fewer types than the type number threshold as target grid regions.

[0077] In one implementation, the number of types of initial trajectory points contained in each candidate grid region is compared with a preset type number threshold, and candidate grid regions with a type number less than the type number threshold are designated as target grid regions. The type number threshold can be adjusted based on the precision or recall of the target trajectory segment. It is easy to see that when the precision of the target trajectory segment is low, the type number threshold is increased to filter out more abnormal initial trajectory points. When the recall of the target trajectory segment is low, the type number threshold is decreased to ensure that a sufficient number of remaining candidate trajectory points are filtered out.

[0078] C. Remove the initial trajectory points contained in the target grid area and use the remaining initial trajectory points as candidate trajectory points.

[0079] In one implementation, the initial trajectory points contained in all target grid regions are removed, and the initial trajectory points of the remaining grid regions are used as candidate trajectory points.

[0080] The number of initial trajectory points of each type in a candidate grid region is determined based on the trajectory point identifiers of the initial trajectory points contained therein. The candidate grid regions are obtained by dividing the electronic map into grids, and candidate grid regions with a number of types less than a threshold are selected as target grid regions. The initial trajectory points contained in the target grid regions are then removed, and the remaining initial trajectory points are selected as candidate trajectory points. Since target grid regions with a small number of initial trajectory points have low reliability, the effect of removing initial trajectory points with low reliability is achieved, ensuring the data reliability of the remaining candidate trajectory points after data filtering.

[0081] Figure 2A This is a flowchart of another method for constructing electronic maps according to embodiments of this disclosure, which is further optimized and extended based on the above technical solutions, and can be combined with the above optional implementation methods.

[0082] like Figure 2A As shown, the method for constructing an electronic map disclosed in this embodiment may include:

[0083] S201. Determine the candidate road locations and use a road matching algorithm to match the candidate trajectory point locations with the candidate road locations.

[0084] In one implementation, the candidate trajectory point positions and candidate road positions for each candidate trajectory point are determined. The candidate trajectory point positions and candidate road positions are used as input to a road matching algorithm, which is then executed to determine if any candidate road position successfully matches a candidate trajectory point position. A successful match indicates that the candidate trajectory point is on any candidate road; a failed match indicates that the candidate trajectory point is not on any candidate road.

[0085] Among them, the road matching algorithm can be the standard map-match algorithm, and the preferred algorithm is the HMM (Hidden Markov Model) algorithm.

[0086] S202. If the match is successful, execute S2031; if the match fails, execute S2032.

[0087] S2031. The candidate road location that matches the candidate trajectory point location is taken as the target road location, and the candidate road to which the target road location belongs is taken as the target road.

[0088] In one implementation, when any candidate road location successfully matches a candidate trajectory point location, the candidate road location is taken as the target road location, and then the candidate road to which the target road location belongs is taken as the target road.

[0089] For example, suppose that the candidate road position A1 of candidate road A is successfully matched with the candidate trajectory point position B1 of candidate trajectory point B, then candidate road A is taken as the target road associated with candidate trajectory point B.

[0090] By determining the candidate road locations and using a road matching algorithm to match the candidate trajectory point locations with the candidate road locations, if a match is successful, the candidate road location that matches the candidate trajectory point location is taken as the target road location, and the candidate road to which the target road location belongs is taken as the target road. This achieves the effect of taking the candidate road where the candidate trajectory point is located as the target road, ensuring the accuracy of the target road.

[0091] S2032. Use the inverse geocoding algorithm to determine the distance between the candidate trajectory point location and the candidate road location, determine the target road location from the candidate road locations, and take the candidate road to which the target road location belongs as the target road; wherein, the target road location is the candidate road location with the smallest distance value from the candidate trajectory point location.

[0092] In one implementation, when any candidate road location fails to match a candidate trajectory point location, the candidate trajectory point locations of each candidate trajectory point and the candidate road locations of each candidate road are used as input to a reverse geocoding algorithm. The reverse geocoding algorithm is then executed to determine the distance between any candidate trajectory point and each candidate road location. The candidate road location with the smallest distance to the candidate trajectory point is selected as the target road location, and the candidate road to which the target road location belongs is then selected as the target road.

[0093] In the event of a matching failure, an inverse geocoding algorithm is used to determine the distance between the candidate trajectory point and the candidate road location. The target road location is then determined from the candidate road locations, and the candidate road to which the target road location belongs is taken as the target road. The target road location is the candidate road location with the smallest distance from the candidate trajectory point location. This achieves the effect of taking the closest candidate road as the target road when the candidate trajectory point is not on any candidate road, thereby improving the recall rate of the target road.

[0094] S204. Based on the direction of the line segment and the direction of the road, determine the angle information between the candidate trajectory line segment and the target road, and based on the angle information and the road characteristics of the target road, determine the target trajectory line segment from the candidate trajectory line segments.

[0095] In one implementation, the angle information between each candidate trajectory segment and its associated target road is determined based on the angle formed by the segment direction of each candidate trajectory segment and the road direction of the target road. The angle information corresponding to each candidate trajectory segment and the road features of the target road are input into the trained model, thereby determining the target trajectory segment from the candidate trajectory segments based on the model's output.

[0096] By determining the angle between the candidate trajectory segment and the target road based on the direction of the line segment and the direction of the road, and then determining the target trajectory segment from the candidate trajectory segments based on the angle information and the road characteristics of the target road, the accuracy of the target trajectory segment determination is improved.

[0097] Optionally, S204 includes:

[0098] The included angle information and road features are input into the classification model, and the trajectory segment type of the candidate trajectory segment is determined based on the output of the classification model; the target trajectory segment is determined from the candidate trajectory segments based on the trajectory segment type.

[0099] The types of classification models include, but are not limited to, logistic regression models, GDBT (Gradient Boosting Decision Tree) models, and LSTM (Long Short-Term Memory) models.

[0100] In one implementation, the included angle information corresponding to each candidate trajectory segment and the road features of the target road are input into a classification model. The classification model outputs a probability value corresponding to each trajectory segment type, and the trajectory segment type with the highest probability value is selected as the trajectory segment type of the candidate trajectory segment. Finally, the candidate trajectory segment that matches the target trajectory segment type is selected as the target trajectory segment.

[0101] By inputting the included angle information and road features into the classification model, and determining the trajectory segment type of the candidate trajectory segments based on the output of the classification model, the target trajectory segment is determined from the candidate trajectory segments according to the trajectory segment type. This achieves the effect of filtering candidate trajectory segments based on trajectory segment type, ensuring the accuracy of the trajectory segment type of the target trajectory segment, and further ensuring the accuracy of pedestrian crossing construction.

[0102] Optionally, the trajectory segment types include road crossing type and road following type.

[0103] Among them, the "crossing road type" represents the candidate trajectory segment that crosses the target road, while the "going-forward road type" represents the candidate trajectory segment that goes forward on the target road.

[0104] "Determine the target trajectory segment from the candidate trajectory segments based on the trajectory segment type" includes: taking the candidate trajectory segments with the trajectory segment type of crossing roads as the target trajectory segment.

[0105] Among them, since the candidate trajectory segments of the road crossing type represent the trajectory points corresponding to pedestrians crossing the target road, the candidate trajectory segments of the road crossing type are taken as the target trajectory segments.

[0106] Figure 2B This is a schematic diagram of some trajectory line segment types disclosed in the embodiments of this disclosure, such as... Figure 2B As shown, candidate trajectory segment 20 is a candidate trajectory segment that travels along the target road 21, therefore, the trajectory segment type of candidate trajectory segment 20 is the "traveling along road" type. Candidate trajectory segment 22 is a candidate trajectory segment that crosses the target road 21, therefore, the trajectory segment type of candidate trajectory segment 22 is the "crossing road" type.

[0107] By using candidate trajectory segments of the type that cross the road as the target trajectory segment, the candidate trajectory points included in the target trajectory segment are all pedestrian trajectory points that cross the target road, thus ensuring the accuracy of pedestrian crossing construction.

[0108] S205. Based on the position of the target trajectory line segments, cluster the target trajectory line segments and determine at least one candidate cluster based on the clustering results.

[0109] In one implementation, based on the position of each target trajectory segment, the target trajectory segments are clustered using DBSCAN (Density-Based Spatial Clustering of Applications with Noise) to determine at least one candidate cluster, wherein each candidate cluster includes at least one target trajectory segment.

[0110] S206. Determine the target cluster from the candidate clusters based on the number of candidate trajectory points contained in the candidate clusters.

[0111] In one implementation, the number of candidate trajectory points contained in the target trajectory segment of each candidate cluster is determined, and the number of candidate trajectory points is compared with a number threshold. Candidate clusters containing more than the number threshold are selected as target clusters.

[0112] S207. Based on the candidate trajectory points contained in the target cluster, construct pedestrian crossings in the electronic map corresponding to the target road.

[0113] In one implementation, pedestrian crossings are constructed in the electronic map corresponding to the target road based on the candidate trajectory point locations contained in the target cluster.

[0114] By clustering the target trajectory segments according to their positions, and determining at least one candidate cluster based on the clustering results, the target cluster is determined from the candidate clusters based on the number of candidate trajectory points contained in each candidate cluster. Based on the candidate trajectory points contained in the target cluster, a pedestrian crossing is constructed on the electronic map corresponding to the target road. This achieves the effect of filtering out abnormal target trajectory segments caused by positioning deviations, ensuring the accuracy of the final pedestrian crossing construction.

[0115] Based on the above embodiments, optionally, the road features include at least one of the following:

[0116] Road grade, speed limit, road width, presence of roadblocks, presence of entrances and exits, and presence of a community gate.

[0117] Among these, "Road Class" indicates the level of the target road, including but not limited to highways, national roads, provincial roads, and urban roads. "Speed ​​Limit" indicates the maximum and minimum speed limits of the target road. "Road Width" indicates the width between the two sides of the target road. "Barriers" indicates whether the target road has obstacles that affect traffic flow. "Entrances / Exits" indicates whether the target road has entrances / exits leading to other roads. "Residential Area Gate" indicates whether there is a residential area gate near the target road.

[0118] By setting road features including at least one of road grade, speed limit, width, presence of roadblocks, presence of entrances / exits, and presence of community gates, the road features have multiple data dimensions, ensuring the reliability and accuracy of determining the target trajectory line segment based on the road features.

[0119] Figure 3 This is a flowchart of some model training methods disclosed in the embodiments of this disclosure. This embodiment can be applied to the training of the classification model in this disclosure.

[0120] like Figure 3As shown, the model training method disclosed in this embodiment may include:

[0121] S301. Determine the sample trajectory points associated with the sample road, and determine at least one sample trajectory line segment based on the sample trajectory points.

[0122] The sample roads are those that include pedestrian crossings in the electronic map.

[0123] In one implementation, all roads with established pedestrian crossings are identified from the map data of an electronic map and designated as sample roads. Based on the location of the sample roads and the locations of candidate trajectory points in the electronic map, candidate trajectory points associated with the sample roads are identified as sample trajectory points. For any sample trajectory point associated with a sample road, the sample trajectory point is segmented according to its position to obtain at least one sample trajectory segment, where each sample trajectory segment consists of a preset number of sample trajectory points.

[0124] S302. The sample trajectory segments of the sample type that crosses the road are taken as positive sample trajectory segments, and the sample trajectory segments of the sample type that go in the same direction as the road are taken as negative sample trajectory segments.

[0125] The sample types include road crossing type and road following type. Road crossing type represents the sample trajectory line segment that crosses the sample road, while road following type represents the sample trajectory line segment that travels forward on the sample road.

[0126] Sample types can be pre-generated manually, whereby relevant personnel label each sample trajectory segment with its sample type based on experience. Sample types can also be automatically labeled based on the angle between the sample trajectory segment and the sample road. For example, assuming the standard angle is 70°–110°, if the angle between any sample trajectory segment and the sample road is within this range, the sample type is automatically labeled as "crossing road"; otherwise, it is automatically labeled as "traveling road".

[0127] In one implementation, for any sample road, the sample types of all sample trajectory segments associated with that sample road are traversed. Sample trajectory segments with the sample type of "crossing a road" are designated as positive sample trajectory segments.

[0128] The sample trajectory segments of the type "conforming to the road" and those within a preset distance are collectively designated as negative sample trajectory segments. For example, the sample trajectory segments of the type "conforming to the road" and those within 500 meters of each other are collectively designated as negative sample trajectory segments.

[0129] S303. Train the model to be trained based on the trajectory segments of positive samples and negative samples to obtain a classification model.

[0130] In one implementation, the line segment directions of positive sample trajectory segments and negative sample trajectory segments are obtained. Then, based on the line segment directions of positive sample trajectory segments, negative sample trajectory segments, sample road directions, and sample road features, the model to be trained is trained to obtain a classification model.

[0131] By identifying sample trajectory points associated with sample roads and determining at least one sample trajectory segment based on these points, the sample trajectory segment of the type crossing the road is designated as the positive sample trajectory segment, and the sample trajectory segment of the type going with the flow of traffic is designated as the negative sample trajectory segment. The model to be trained is then trained based on the positive and negative sample trajectory segments to obtain the classification model. Here, the sample roads are those in the electronic map that include pedestrian crossings. This achieves the effect of obtaining both positive and negative sample training data, thus ensuring the training accuracy of the classification model.

[0132] Optionally, S303 includes the following steps:

[0133] 1) Determine the direction of the first sample line segment of the positive sample trajectory line segment, the direction of the second sample line segment of the negative sample trajectory line segment, and the direction of the sample road.

[0134] In one implementation, the first sample line segment direction of each positive sample trajectory line segment is determined based on its orientation in the electronic map. The second sample line segment direction of each negative sample trajectory line segment is determined based on its orientation in the electronic map. The sample road direction of the sample road is determined based on its orientation in the electronic map.

[0135] 2) Determine the first sample angle information between the positive sample trajectory line segment and the sample road based on the direction of the first sample line segment and the direction of the sample road.

[0136] In one implementation, the first sample angle information between each positive sample trajectory segment and its associated sample road is determined based on the angle formed by the first sample segment direction of each positive sample trajectory segment and the sample road direction of its associated sample road.

[0137] 3) Determine the angle information between the negative sample trajectory line segment and the sample road based on the direction of the second sample line segment and the direction of the sample road.

[0138] In one implementation, the second sample angle information between each negative sample trajectory segment and its associated sample road is determined based on the angle formed by the second sample segment direction of each negative sample trajectory segment and the sample road direction of its associated sample road.

[0139] 4) Train the model to be trained based on the angle information of the first sample, the angle information of the second sample, and the sample road features of the sample road.

[0140] In one implementation, sample road features of sample roads are obtained from the database of electronic maps, and the first sample angle information corresponding to the positive sample trajectory line segment, the second sample angle information corresponding to the negative sample trajectory line segment, and the sample road features are used as training data. The training data is used to train the model to be trained and generate a classification model.

[0141] By determining the direction of the first sample line segment of the positive sample trajectory, the direction of the second sample line segment of the negative sample trajectory, and the direction of the sample road, the first sample angle between the positive sample trajectory and the sample road is determined based on the direction of the first sample line segment and the direction of the sample road. Similarly, the second sample angle between the negative sample trajectory and the sample road is determined based on the direction of the second sample line segment and the direction of the sample road. Then, the model to be trained is trained based on the first sample angle of the positive sample trajectory and the second sample angle of the negative sample trajectory. Since the model to be trained is based on the first sample angle of the positive sample trajectory and the second sample angle of the negative sample trajectory, the convergence speed of the model to be trained is accelerated, and the training accuracy is guaranteed.

[0142] Figure 4 This is a schematic diagram of the structure of some electronic map construction apparatuses disclosed in embodiments of this disclosure, which can be applied to the construction of pedestrian crossings in electronic maps. The apparatus of this embodiment can be implemented in software and / or hardware and can be integrated into any electronic device with computing capabilities.

[0143] like Figure 4 As shown, the electronic map construction apparatus 40 disclosed in this embodiment may include a target road determination module 41, a direction determination module 42, a trajectory line segment determination module 43, and a map construction module 44, wherein:

[0144] The target road determination module 41 is used to determine the target road associated with the candidate trajectory point based on the candidate trajectory point position;

[0145] The direction determination module 42 is used to determine at least one candidate trajectory line segment based on the candidate trajectory points, and to determine the line segment direction of the candidate trajectory line segment and the road direction of the target road.

[0146] The trajectory segment determination module 43 is used to determine the target trajectory segment from the candidate trajectory segments based on the segment direction, road direction, and road characteristics of the target road.

[0147] Map building module 44 is used to construct pedestrian crossings on the electronic map corresponding to the target road based on the target trajectory line segments.

[0148] Optionally, the trajectory segment determination module 43 is specifically used for:

[0149] Based on the direction of the line segment and the direction of the road, determine the angle information between the candidate trajectory line segment and the target road;

[0150] Based on the included angle information and the road characteristics of the target road, the target trajectory segment is determined from the candidate trajectory segments.

[0151] Optionally, the trajectory segment determination module 43 is also used for:

[0152] The included angle information and road features are input into the classification model, and the trajectory segment type of the candidate trajectory segment is determined based on the output of the classification model.

[0153] The target trajectory segment is determined from the candidate trajectory segments based on the trajectory segment type.

[0154] Optionally, the trajectory segment types include road crossing type and road following type;

[0155] The trajectory segment determination module 43 is also specifically used for:

[0156] Candidate trajectory segments with the type of "crossing roads" are selected as target trajectory segments.

[0157] Optionally, the device also includes a model training module, specifically used for:

[0158] The classification model is trained in the following way:

[0159] Identify the sample trajectory points associated with the sample road, and determine at least one sample trajectory line segment based on the sample trajectory points;

[0160] The sample trajectory segments with the sample type of crossing the road are regarded as positive sample trajectory segments, and the sample trajectory segments with the sample type of going in the same direction as the road are regarded as negative sample trajectory segments.

[0161] The model to be trained is obtained by training the model based on the trajectory segments of positive samples and negative samples;

[0162] The sample roads are those that include pedestrian crossings in the electronic map.

[0163] Optional, the model training module is also used for:

[0164] Determine the direction of the first sample line segment of the positive sample trajectory line segment, the direction of the second sample line segment of the negative sample trajectory line segment, and the direction of the sample road;

[0165] Based on the direction of the first sample line segment and the direction of the sample road, determine the first sample angle information between the positive sample trajectory line segment and the sample road;

[0166] Based on the direction of the second sample line segment and the direction of the sample road, determine the angle information between the negative sample trajectory line segment and the sample road for the second sample;

[0167] The model to be trained is trained based on the angle information of the first sample, the angle information of the second sample, and the sample road features.

[0168] Optional, map building module 44, specifically used for:

[0169] Based on the position of the target trajectory line segments, cluster the target trajectory line segments, and determine at least one candidate cluster based on the clustering results;

[0170] The target cluster is determined from the candidate clusters based on the number of candidate trajectory points contained in the candidate clusters;

[0171] Based on the candidate trajectory points contained in the target cluster, pedestrian crossings are constructed in the electronic map corresponding to the target road.

[0172] Optionally, the target road determination module 41 is specifically used for:

[0173] The candidate road locations are determined, and a road matching algorithm is used to match the candidate trajectory point locations with the candidate road locations;

[0174] If a match is successful, the candidate road location that matches the candidate trajectory point location will be used as the target road location, and the candidate road to which the target road location belongs will be used as the target road.

[0175] Optionally, the device also includes a distance value determination module, specifically used for:

[0176] In the event of a failed match, an inverse geocoding algorithm is used to determine the distance between the candidate trajectory point and the candidate road location.

[0177] The target road location is determined from the candidate road locations, and the candidate road to which the target road location belongs is taken as the target road; where the target road location is the candidate road location with the smallest distance value from the candidate trajectory point location.

[0178] Optionally, the device further includes a first candidate trajectory point determination module, specifically used for:

[0179] Determine the initial trajectory point location and the target area location; the target area includes building areas and / or enclosed park areas.

[0180] Initial trajectory points whose initial trajectory point positions are located in the region are removed, and the remaining initial trajectory points are used as candidate trajectory points.

[0181] Optionally, the device further includes a second candidate trajectory point determination module, specifically used for:

[0182] The number of initial trajectory points in a candidate grid region is determined based on the trajectory point identifiers of the initial trajectory points contained therein; wherein, the candidate grid region is obtained by dividing the electronic map into grids.

[0183] Candidate grid regions with fewer types than the type number threshold are selected as target grid regions;

[0184] The initial trajectory points contained in the target grid area are removed, and the remaining initial trajectory points are used as candidate trajectory points.

[0185] Optionally, the road features include at least one of the following:

[0186] Road grade, speed limit, road width, presence of roadblocks, presence of entrances and exits, and presence of a community gate.

[0187] The electronic map construction apparatus 40 disclosed in this embodiment can execute the electronic map construction method disclosed in this embodiment, and has the corresponding functional modules and beneficial effects of the method. Content not described in detail in this embodiment can be referred to the description in the method embodiments of this disclosure.

[0188] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0189] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0190] Figure 5A schematic block diagram of an example electronic device 500 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0191] like Figure 5 As shown, device 500 includes a computing unit 501, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 502 or a computer program loaded from storage unit 508 into random access memory (RAM) 503. RAM 503 may also store various programs and data required for the operation of device 500. The computing unit 501, ROM 502, and RAM 503 are interconnected via bus 504. Input / output (I / O) interface 505 is also connected to bus 504.

[0192] Multiple components in device 500 are connected to I / O interface 505, including: input unit 506, such as keyboard, mouse, etc.; output unit 507, such as various types of monitors, speakers, etc.; storage unit 508, such as disk, optical disk, etc.; and communication unit 509, such as network card, modem, wireless transceiver, etc. Communication unit 509 allows device 500 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0193] The computing unit 501 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above, such as the method of constructing an electronic map. For example, in some embodiments, the method of constructing an electronic map can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on device 500 via ROM 502 and / or communication unit 509. When the computer program is loaded into RAM 503 and executed by the computing unit 501, one or more steps of the method of constructing an electronic map described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the method of constructing an electronic map by any other suitable means (e.g., by means of firmware).

[0194] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0195] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0196] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0197] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0198] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0199] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0200] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0201] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for constructing an electronic map, comprising: determining a target road associated with a candidate trajectory point according to a candidate trajectory point position of the candidate trajectory point; determining at least one candidate trajectory segment according to the candidate trajectory point, and determining a segment direction of the candidate trajectory segment and a road direction of the target road; determining a target trajectory segment from the candidate trajectory segment according to the segment direction, the road direction, and a road feature of the target road; constructing a crosswalk in the target road in an electronic map corresponding to the target road according to the target trajectory segment; wherein the determining the target road associated with the candidate trajectory point according to the candidate trajectory point position of the candidate trajectory point comprises: determining a candidate road position of a candidate road, and matching the candidate trajectory point position with the candidate road position by using a road matching algorithm; in a case of a failed matching, determining a distance value between the candidate trajectory point position and the candidate road position by using a reverse geocoding algorithm; determining a target road position from the candidate road position, and taking a candidate road to which the target road position belongs as the target road; wherein the target road position is a candidate road position with a minimum distance value from the candidate trajectory point position.

2. The method of claim 1, wherein, the determining the target trajectory segment from the candidate trajectory segment according to the segment direction, the road direction, and the road feature of the target road comprises: determining angle information between the candidate trajectory segment and the target road according to the segment direction and the road direction; determining the target trajectory segment from the candidate trajectory segment according to the angle information and the road feature of the target road.

3. The method of claim 2, wherein, the determining the target trajectory segment from the candidate trajectory segment according to the angle information and the road feature of the target road comprises: inputting the angle information and the road feature into a classification model, and determining a trajectory segment type of the candidate trajectory segment according to an output result of the classification model; determining the target trajectory segment from the candidate trajectory segment according to the trajectory segment type.

4. The method of claim 3, wherein, the trajectory segment type comprises a crossing road type and a straight road type; the determining the target trajectory segment from the candidate trajectory segment according to the trajectory segment type comprises: taking a candidate trajectory segment with the trajectory segment type being the crossing road type as the target trajectory segment.

5. The method of claim 3, wherein, the classification model is obtained by training in the following manner: determining sample trajectory points associated with a sample road, and determining at least one sample trajectory segment according to the sample trajectory points; taking a sample trajectory segment with a sample type being the crossing road type as a positive sample trajectory segment, and taking a sample trajectory segment with the sample type being the straight road type as a negative sample trajectory segment; training a to-be-trained model according to the positive sample trajectory segment and the negative sample trajectory segment to obtain the classification model; wherein the sample road is a road including a crosswalk in an electronic map.

6. The method of claim 5, wherein, the training the to-be-trained model according to the positive sample trajectory segment and the negative sample trajectory segment comprises: determine a first sample line segment direction of the positive sample trajectory line segment, a second sample line segment direction of the negative sample trajectory line segment, and a sample road direction of the sample road; determine first sample angle information between the positive sample trajectory line segment and the sample road according to the first sample line segment direction and the sample road direction; determine second sample angle information between the negative sample trajectory line segment and the sample road according to the second sample line segment direction and the sample road direction; train the to-be-trained model according to the first sample angle information, the second sample angle information, and a sample road feature of the sample road.

7. The method of claim 1, wherein, The constructing the crosswalk in the electronic map corresponding to the target road according to the target trajectory line segment includes: clustering the target trajectory line segment according to a line segment position of the target trajectory line segment, and determining at least one candidate clustering cluster according to a clustering result; determining a target clustering cluster from the candidate clustering cluster according to a number of candidate trajectory points contained in the candidate clustering cluster; constructing the crosswalk in the electronic map corresponding to the target road according to a candidate trajectory point contained in the target clustering cluster.

8. The method of claim 1, wherein, The determining the target road associated with the candidate trajectory point according to the candidate trajectory point position of the candidate trajectory point includes: determining a candidate road position of a candidate road, and matching the candidate trajectory point position with the candidate road position by using a road matching algorithm; in a case of successful matching, taking the candidate road position matched with the candidate trajectory point position as a target road position, and taking a candidate road to which the target road position belongs as the target road. 9.The method of claim 1, before the determining the target road associated with the candidate trajectory point according to the candidate trajectory point position of the candidate trajectory point, further comprising: determining an initial trajectory point position of an initial trajectory point and a region position of a target region; wherein the target region includes a building region and / or a closed park region; eliminating the initial trajectory point whose initial trajectory point position is at the region position, and taking the remaining initial trajectory points as the candidate trajectory points. 10.The method of claim 1, before the determining the target road associated with the candidate trajectory point according to the candidate trajectory point position of the candidate trajectory point, further comprising: determining a type number of initial trajectory points contained in a candidate grid region according to a trajectory point identifier of the initial trajectory points contained in the candidate grid region; wherein the candidate grid region is obtained by grid division on the electronic map; taking a candidate grid region with a type number less than a type number threshold as a target grid region; eliminating the initial trajectory points contained in the target grid region, and taking the remaining initial trajectory points as the candidate trajectory points.

11. The method of any one of claims 1-10, wherein, The road feature includes at least one of the following: a road level, a road speed limit, a road width, whether having a roadblock, whether having an entrance and exit, and whether having a cell gate. 12.An electronic map construction device, comprising: The target road determining module is configured to determine a target road associated with the candidate trajectory point according to a candidate trajectory point position of the candidate trajectory point. The direction determining module is configured to determine at least one candidate trajectory line segment according to the candidate trajectory point, and determine a line segment direction of the candidate trajectory line segment and a road direction of the target road. The trajectory line segment determining module is configured to determine a target trajectory line segment from the candidate trajectory line segment according to the line segment direction, the road direction, and a road feature of the target road. The map constructing module is configured to construct a crosswalk in an electronic map corresponding to the target road according to the target trajectory line segment. The target road determining module is specifically configured to: determine a candidate road position of a candidate road, and match the candidate trajectory point position with the candidate road position by using a road matching algorithm; in a case of a failed matching, determine a distance value between the candidate trajectory point position and the candidate road position by using a reverse geocoding algorithm; determine a target road position from the candidate road position, and take a candidate road to which the target road position belongs as the target road; wherein the target road position is a candidate road position with a minimum distance value from the candidate trajectory point position.

13. The apparatus of claim 12, wherein, The trajectory line segment determining module is specifically configured to: determine angle information between the candidate trajectory line segment and the target road according to the line segment direction and the road direction; determine a target trajectory line segment from the candidate trajectory line segment according to the angle information and the road feature of the target road.

14. The apparatus of claim 13, wherein, The trajectory line segment determining module is further configured to: input the angle information and the road feature into a classification model, and determine a trajectory line segment type of the candidate trajectory line segment according to an output result of the classification model; determine a target trajectory line segment from the candidate trajectory line segment according to the trajectory line segment type.

15. The apparatus of claim 14, wherein, The trajectory line segment type includes a cross road type and a go straight road type. The trajectory line segment determining module is further configured to: take a candidate trajectory line segment with the cross road type as the target trajectory line segment.

16. The apparatus of claim 14, wherein, The device further includes a model training module configured to: The classification model is obtained by training in the following manner: determine a sample trajectory point associated with a sample road, and determine at least one sample trajectory line segment according to the sample trajectory point; take a sample trajectory line segment with a cross road type as a positive sample trajectory line segment, and take a sample trajectory line segment with a go straight road type as a negative sample trajectory line segment; train a to-be-trained model according to the positive sample trajectory line segment and the negative sample trajectory line segment to obtain the classification model. The sample road is a road including a crosswalk in an electronic map.

17. The apparatus of claim 16, wherein, The model training module is further configured to: determine a first sample line segment direction of the positive sample trajectory line segment, a second sample line segment direction of the negative sample trajectory line segment, and a sample road direction of the sample road; determine first sample angle information between the positive sample trajectory line segment and the sample road according to the first sample line segment direction and the sample road direction; and determine second sample angle information between the negative sample trajectory line segment and the sample road according to the second sample line segment direction and the sample road direction. determine second sample angle information between the negative sample trajectory line segment and the sample road according to the second sample line segment direction and the sample road direction; train the to-be-trained model according to the first sample angle information, the second sample angle information, and sample road characteristics of the sample road.

18. The apparatus of claim 12, wherein, The map construction module is specifically configured to: cluster the target trajectory line segment according to a line segment position of the target trajectory line segment, and determine at least one candidate clustering cluster according to a clustering result; determine a target clustering cluster from the candidate clustering cluster according to a quantity of candidate trajectory points contained in the candidate clustering cluster; construct a pedestrian crossing in an electronic map corresponding to the target road according to the candidate trajectory points contained in the target clustering cluster.

19. The apparatus of claim 12, wherein, The target road determination module is specifically configured to: determine a candidate road position of a candidate road, and match the candidate trajectory point position and the candidate road position by using a road matching algorithm; in a case of successful matching, take the candidate road position matched with the candidate trajectory point position as a target road position, and take a candidate road to which the target road position belongs as the target road.

20. The apparatus of claim 12, wherein, The device further includes a first candidate trajectory point determination module, which is specifically configured to: determine an initial trajectory point position of an initial trajectory point and a region position of a target region; wherein the target region includes a building region and / or a closed park region; eliminate the initial trajectory point whose initial trajectory point position is at the region position, and take the remaining initial trajectory points as the candidate trajectory points.

21. The apparatus of claim 12, wherein, The device further includes a second candidate trajectory point determination module, which is specifically configured to: determine a type quantity of initial trajectory points contained in a candidate grid region according to trajectory point identifiers of the initial trajectory points contained in the candidate grid region; wherein the candidate grid region is obtained by grid division on an electronic map; take a candidate grid region with a type quantity less than a type quantity threshold as a target grid region; eliminate the initial trajectory points contained in the target grid region, and take the remaining initial trajectory points as the candidate trajectory points.

22. The apparatus of any of claims 12-21, wherein, The road characteristics include at least one of the following: a road level, a road speed limit, a road width, whether having a roadblock, whether having an entrance and exit, and whether having a cell gate. 23.An electronic device, comprising: at least one processor; and a memory connected with the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-11.

24. A non-transitory computer readable storage medium having stored thereon computer instructions, wherein, The computer instructions are used to enable the computer to perform the method of any one of claims 1-11. 25.A computer program product comprising a computer program which, when executed by a processor, implements the method of any one of claims 1-11.

Images