Overpass multilayer structure generation method considering directed topological connection relation

Abstract

The invention belongs to the technical field of road network generation, and discloses an overpass multilayer structure generation method considering a directed topological connection relationship, which comprises the following steps: obtaining overpass track data, generating a density grid map and a skeleton line based on the data, and obtaining a road geometric shape; defining a direction domain and center line rule, matching the direction cluster density grid map to the direction domain, and vectorizing according to the rule to obtain a directed center line in each direction; connecting the direction lines by the integral track center line to obtain an initial overpass map; setting an entrance and exit range, and screening a high-confidence track. And in combination with the high-confidence trajectory, matching the initial map with a map to obtain trajectory point road matching information. And carrying out topological check and correction on the matched information, and finally generating the directed three-dimensional vector flyover road. According to the method, the problem of complexity of overpass generation in road network construction is solved, information such as geometric morphology, topological connection and road direction of the overpass is extracted, and powerful support is provided for automatic generation of the road network.

Description

Technical Field

[0001] The invention belongs to the technical field of road network generation, and in particular relates to a method for generating a multi-layer structure of an overpass taking into account directed topological connection relations. Background Art

[0002] Compared with intersections, although both have the functions of connecting road sections and carrying turning points, the former covers a larger area and is composed of roads with diverse geometric linear structures; compared with roads in the road network, the road sections contained in the interchange are denser and intertwined, and the topological connection relationship is more complex. Considering that the road sections in the interchange are dense, the topological relationship is complex, and the non-planar structure is an important and complex structure in the road network, it is difficult for existing intersection recognition or road extraction technologies to extract its multi-layer structure. Summary of the invention

[0003] The purpose of the present invention is to provide a method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships, so as to solve the problems existing in the above-mentioned prior art.

[0004] To achieve the above object, the present invention provides a method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships, comprising:

[0005] Obtain trajectory data within the overpass area;

[0006] Based on the trajectory data, a density grid map and corresponding skeleton lines of each direction cluster and the overall trajectory are obtained to obtain the geometric shape data of the overpass road;

[0007] Define several directional domains and set the vector centerline rule of the centerline for each directional domain;

[0008] Determine the direction domain corresponding to the density grid map of each direction cluster, and perform directed vectorization on each density grid map based on the corresponding centerline vectorization rule to obtain the directed vector centerline of each direction;

[0009] Based on the directed vector centerline corresponding to the overall trajectory, the directed vector centerlines in each direction are connected to obtain an initial directed vector overpass geometric road map;

[0010] Setting an entrance and exit range of the overpass, and selecting a high-confidence trajectory in the trajectory data based on the set entrance and exit range of the overpass;

[0011] Based on the selected high-confidence trajectory and the initial directed vector overpass geometric road map, the matching information between each trajectory point and the road is obtained by combining the map matching method;

[0012] The matching information is topologically checked and corrected to generate a directed three-dimensional vector overpass road.

[0013] Optionally, the obtaining of density grid images and corresponding skeleton lines of each direction cluster and the overall trajectory based on the trajectory data specifically includes:

[0014] According to the orientation of the trajectory points, the trajectory data is split into different direction clusters;

[0015] A density grid map is constructed for each direction cluster and the overall trajectory, and the skeleton lines of each density grid map are extracted.

[0016] Optionally, the defining of several direction domains specifically includes:

[0017] The directions are divided into four direction domains of "north, south, east, and west", and a vector centerline rule is set for each direction domain; each direction domain contains 180°, and there is overlap between the direction domains.

[0018] Optionally, the process of obtaining the center line of the directed vector in each direction specifically includes:

[0019] Determine the direction domain to which the density grid map of each direction cluster belongs. If the direction cluster range of the density grid map is completely contained in one direction domain, the center line is processed in a directed vector according to the vector center line rule of the corresponding direction domain; if the density grid map belongs to two direction domains, the center line is processed in a directed vector based on the vector center line rules of the two direction domains.

[0020] Optionally, the connecting of the directed vector center lines in various directions based on the directed vector center line corresponding to the overall trajectory specifically includes:

[0021] The binary images corresponding to the center lines of the directional vectors in various directions are superimposed, and pixels containing information of more than two directions are extracted, and the extracted pixels are vectorized to obtain a number of vector polygons;

[0022] Each vector polygon is superimposed and analyzed with the directed vector center lines in each direction to determine the interrupted road pairs; based on the skeleton line corresponding to the overall trajectory, the interrupted road pairs and the intermediate connecting parts are merged into a road with only two endpoints, and the redundant parts are deleted to complete the construction of the initial directed vector overpass geometric road map.

[0023] Optionally, the screening of high-confidence trajectories in the trajectory data specifically includes:

[0024] The trajectory data is processed by removing stop points, processing sampling point loss, and eliminating abnormal sampling frequencies to obtain preprocessed trajectory data;

[0025] The entrance and exit range of the overpass is set, and based on the set entrance and exit range of the overpass, the trajectory passing through the start and end exit range is selected in the preprocessed trajectory data;

[0026] Set a buffer zone for each road and calculate the ratio of the length of the trajectory passing through the buffer zone to the total length of the trajectory;

[0027] Calculate the ratio of each trajectory, retain the trajectories with a ratio greater than the preset threshold, and remove the trajectories with a ratio less than the preset threshold.

[0028] Optionally, the method of combining a map matching method to obtain matching information between each track point and the road specifically includes:

[0029] When there is a problem with the road network that causes the matching to be interrupted, the track is segmented with the interruption point as the segmentation point, the unmatched road segments are marked, and the subsequent track segments are matched again until the entire track is matched and the matching road segment sequence is output;

[0030] Calculate the projection point of the trajectory point on the road section and the corresponding front and rear adjacent node positions to determine the direction of the road section; compare the direction of the trajectory point with the direction of the road section. If the difference between the direction of the trajectory point and the direction of the road section is within the preset threshold range, retain and construct a candidate road section set; if the difference between the direction of the trajectory point and the direction of the road section is not within the preset threshold range, remove it;

[0031] The HMM algorithm is improved to allow all projection points of a point to participate in subsequent matching calculations.

[0032] Optionally, the performing topological checking and correction on the matching information specifically includes:

[0033] The trajectories containing unmatched points are divided into multiple categories according to the sequence numbers of the sections passed by the trajectories;

[0034] Apply the best representative merging method to select a representative trajectory from each category as the classic trajectory;

[0035] Based on the classic trajectory, the intersection of the adjacent road sections before and after the mismatched trajectory point is intercepted as the new road, and the road network nodes are supplemented to ensure that the direction of the new road is consistent with the trajectory travel direction;

[0036] Eliminate the false road sections without matching information generated during the raster-to-vector conversion process;

[0037] Based on the road connectivity sequence of the long trajectory, a topological tree structure with the entrance as the root node and the road section as the node is constructed; the trajectory traffic of each section is used as the pruning basis, the sections with the same traffic sequence are merged, and the redundant connections are deleted to obtain the correct topological connection relationship; according to the obtained correct topological connection relationship, the overpass road network is corrected.

[0038] The technical effects of the present invention are:

[0039] The present invention uses trajectory data to study the most complex overpass generation problem in road network construction, and extracts information such as its geometric shape, topological connection, and road direction, laying the foundation for the automatic generation of large-scale road networks. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0042] Figure 1 It is a flow chart of an implementation of a method for generating a multi-layer structure of an overpass taking into account a directed extension connection relationship in an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of the calculation of the symmetric path distance of two trajectories in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as limiting the present invention, but should be understood as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0045] It should be understood that the terms described in the present invention are only for describing special embodiments and are not intended to limit the present invention. In addition, for the numerical range in the present invention, it should be understood that each intermediate value between the upper and lower limits of the scope is also specifically disclosed. Each smaller range between the intermediate value in any stated value or stated range and any other stated value or intermediate value in the described range is also included in the present invention. The upper and lower limits of these smaller ranges can be independently included or excluded in the scope.

[0046] It will be apparent to those skilled in the art that various modifications and variations may be made to the specific embodiments of the present invention description without departing from the scope or spirit of the present invention. Other embodiments derived from the present invention description will be apparent to those skilled in the art. The present application description and examples are exemplary only.

[0047] The words “include,” “including,” “have,” “contain,” etc. used in this article are open-ended terms, meaning including but not limited to.

[0048] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0049] like Figure 1 - Figure 2 As shown, in this embodiment, a method for generating a multi-layer structure of an overpass taking into account a directed topological connection relationship is provided, including: obtaining trajectory data in an overpass area; obtaining a density grid map and a corresponding skeleton line of each directional cluster and the overall trajectory based on the trajectory data to obtain geometric data of the overpass road; defining a plurality of directional domains, and setting a vector centerline rule of the centerline for each directional domain; determining the directional domain corresponding to the density grid map of each directional cluster, and performing directed vectorization on each density grid map based on the corresponding centerline vectorization rule to obtain a directed vector centerline in each direction; connecting the directed vector centerlines in each direction based on the directed vector centerline corresponding to the overall trajectory to obtain an initial directed vector overpass geometric road map; setting an overpass entrance and exit range, and filtering high-confidence trajectories in the trajectory data based on the set overpass entrance and exit range; obtaining matching information between each trajectory point and the road in combination with a map matching method based on the filtered high-confidence trajectory and the initial directed vector overpass geometric road map; performing topological inspection and correction on the matching information to generate a directed three-dimensional vector overpass road.

[0050] This embodiment first obtains the trajectory data in the overpass area, and generates the density grid map of each direction cluster and the overall trajectory and its corresponding skeleton line based on these data, so as to extract the road geometry data of the overpass. Then, multiple direction domains are defined, and the vector centerline rule of the centerline is set for each direction domain. According to the direction domain attribution of the density grid map, the corresponding centerline vectorization rule is used to perform directed vectorization to obtain the directed vector centerline of each direction. Then, based on the directed vector centerline of the overall trajectory, the directed vector centerlines of each direction are connected to form an initial directed vector overpass geometric road map. Subsequently, the overpass entrance and exit range is set, high-confidence trajectories are screened out, and the matching information of the trajectory points and the road is obtained in combination with the map matching method. Finally, the matching information is topologically checked and corrected to ensure the accuracy and connectivity of the road network, and finally a directed three-dimensional vector overpass road is generated. The present invention successfully solves the complex problem of overpass generation in road network construction, extracts information such as the geometric form, topological connection and road direction of the overpass, and provides strong support for the automatic generation of large-scale road networks.

[0051] The specific implementation process of this embodiment includes:

[0052] Step 1, extracting the initial overpass directed vector geometry: taking into account the non-planar structure of the overpass that overlaps up and down, integrating the trajectory grid graphs of different direction clusters to obtain the complete geometry of the overpass including the overlapping but not intersecting sections;

[0053] Step 2, directed vectorization method: construct a mapping relationship between the road skeleton lines of different direction clusters and the vector road directions obtained, and realize the directed vectorization of the road;

[0054] Step 3, high confidence trajectory screening: screening high confidence trajectories from aspects such as trajectory preprocessing, overpass entrance and exit range, and the ratio of trajectory sequences falling on road sections;

[0055] Step 4, topological connection relationship inspection and reconstruction: First, based on the improved map matching method, all road information is used as much as possible on the basis of the road network containing errors to obtain the matching information of each trajectory point and the road. Then, based on the matching results, inspection and reconstruction are carried out. First, roads are supplemented based on unmatched trajectory points, second, road sections with unmatched trajectories are eliminated, and third, based on the correct information after iterative matching, redundant information is deleted, so as to obtain a directed three-dimensional vector overpass road with complete geometric form and correct topological connection.

[0056] The extraction of the initial overpass directed vector geometry in step 1 specifically includes:

[0057] Step 1.1, direction cluster splitting, based on the split direction cluster threshold range, according to the direction of the trajectory point, respectively obtain the trajectory direction density grid map of each direction cluster and extract the skeleton line respectively;

[0058] Step 1.2, obtain the density grid map of the whole trajectory and extract the skeleton line.

[0059] The method for constructing the density grid map described in step 1.1 is as follows:

[0060] In the density map, the pixel value of each grid represents the number of trajectory points falling into the grid. For the convenience of display, after removing individual outliers, the density map under each direction cluster is globally normalized using Formula 1, so that the pixel values ​​of the image are between [0,1], and then multiplied by 255 to obtain the direction density map.

[0061]

[0062] Where P norm is the pixel value after trajectory; p is the number of trajectory points in the grid itself; p max is the maximum number of trajectory points in the grid; p min It is the minimum number of track points in the grid.

[0063] The directed vectorization method in step 2 may specifically include:

[0064] Step 2.1, divide the direction into four direction domains of "north, south, east, west", each direction domain covers 180°, and there is overlap between the direction domains, as shown in Table 1.

[0065] Step 2.2, setting the center line vectorization rules of each direction domain to assign the information of each direction domain to the vector line.

[0066] Table 1 Direction domain range and vector line rules

[0067]

[0068] Step 2.3, determine the direction domain to which the grid map of each direction cluster in step 1.1 belongs. If the direction cluster range of the direction density map is completely contained in a certain direction domain, the center line is vectorized in a directed manner according to the vector center line rule of the direction domain. If the direction density map belongs to two direction domains, the center line vectorization needs to meet the rules of the two direction domains.

[0069] Step 2.4, split the road and merge it. Use the center line obtained after the vector in step 1.2 to connect the center lines of the directed vectors in multiple directions obtained in step 2.3 to obtain the initial directed vector overpass geometric road map.

[0070] The split road merging method described in step 2.4 may specifically include:

[0071] (2.4.1) Split area detection, after superimposing the binary images of each direction, extract the pixels containing more than two direction information, and vectorize the area composed of such pixels to obtain multiple vector polygons;

[0072] (2.4.2) Road interruption extraction: Superimpose the vector polygon of the split area with the vector center lines of the roads in each direction in step 2.3. If the polygon contains the starting point of one road segment and the end point of another road segment, the two road segments are divided into interrupted road pairs and need to be merged; if the polygon does not contain endpoints, the area is regarded as the overlapping area of ​​the upper and lower roads and no processing is performed; if there are more than two endpoints or only one endpoint in the polygon, the area is marked as a pending inspection area and needs to be manually checked.

[0073] (2.4.3) Road merging: Use the road centerline in step 1.2 as the basis for geometric connection of the interrupted road segments, then merge the road pairs and the intermediate connecting parts into a road with only two endpoints, and delete the redundant parts. In addition, the direction of the connecting part should be consistent with the connecting sections before and after it.

[0074] The high confidence trajectory screening described in step 3 may specifically include:

[0075] Step 3.1 Initial trajectory filtering: including removal of stop points, interruption of trajectories where sampling points are lost, and elimination of trajectories with abnormal sampling frequencies.

[0076] Step 3.2 Interchange entrance and exit filtering: Set the interchange entrance and exit range, and only keep the tracks within the exit range of the starting point and the end point.

[0077] Step 3.3 Road segment filtering: Calculate the ratio T of the length of the trajectory sequence passing through the road buffer to the total length of the trajectory, and set a threshold. If T < threshold, the trajectory is removed, otherwise it is deleted.

[0078] The topology connection relationship inspection and reconstruction described in step 4 may specifically include:

[0079] Step 4.1, improve the hidden Markov map matching model to be applicable to the problem road network with missing roads and unconnected road topology, and also take into account the road directionality and non-planarity;

[0080] Step 4.2, topology check and correction. First, if the trajectory points are not matched, the mismatched trajectories are processed to supplement the missing roads and missing topological connections in the road network; second, the roads without matching trajectory points are removed; third, when all trajectory points are correctly matched, the redundant road connection relationships are removed.

[0081] The specific implementation method of step 4.1 to improve the hidden Markov map matching model is:

[0082] (4.1.1) In the map matching process of HMM, if there are problems in the road network, such as missing roads or topological disconnection errors, the trajectory points cannot be effectively matched, which will cause the matching algorithm to be interrupted. The improved method is to use the trajectory points that cause the matching interruption as the segmentation points of the complete trajectory, and mark "NULL" at the matchable road segment identifier, and then map match the subsequent trajectory segments again until the last trajectory point is matched, and finally output the road segment sequence matched by the complete trajectory sequence.

[0083] (4.1.2) Map matching that takes direction into account, calculate the projection point of the trajectory point on the road section, obtain the position of the road section nodes before and after it, and then calculate the direction of the road section. Compare the direction of the trajectory point with the direction of the road section. If the difference between the direction of the trajectory point and the direction of the road section is not within the threshold range, it will be eliminated from the candidate road section set, otherwise it will be retained.

[0084] (4.1.3) Improvements to take into account non-planar roads. The HMM algorithm can only select one projection point for a road section to participate in the subsequent calculation, and will mistakenly match similar but unconnected roads at the intersection of roads. The improvement method is to improve the constraint that HMM can only select one projection point, that is, to add all the projection points of a point to the subsequent calculation process.

[0085] The specific implementation method of step 4.2 topology check and correction is feasible as follows:

[0086] (4.2.1) Road supplement based on mismatched trajectories: First, the trajectory sequence containing unmatched points is divided into multiple categories according to the serial number of the road sections passed by the trajectory; then, the optimal representative merging method is applied to select a trajectory that is most similar to other trajectories as a whole from each cluster as the classic trajectory, and the classic trajectory is used to represent the generated new road information; finally, the intersection of the classic trajectory and the adjacent road sections before and after the mismatched trajectory point is intercepted as the newly added road, and the intersection nodes are supplemented as road network nodes, while ensuring that the direction of the newly added road is consistent with the direction of the trajectory.

[0087] It is feasible to calculate the trajectory similarity in (4.2.1) using the symmetric segment path distance, that is, to calculate the minimum distance from each trajectory point on the trajectory sequence of A to the line segment between two consecutive points on trajectory B, and then accumulate the minimum distance of all trajectory points of A, and divide it by the number of trajectory points of A to obtain the distance from trajectory A to trajectory B as the basis for similarity measurement ( Figure 2 ).

[0088] (4.2.2) Eliminate road sections without matching information. There are glitches when the trajectory points are converted from raster to vector. Such pseudo road sections have no trajectory points matching them during trajectory matching, so they are eliminated.

[0089] (4.2.3) Redundant connection deletion based on correct matching results. After iterative map matching and processing, although there are no mismatched trajectories or roads without matching information, it is still impossible to guarantee that there are no redundant road segment connections in the road network. To this end, the road connectivity sequence of long trajectories is taken into account, the trajectory road segment traffic sequence of the same entrance is merged, and a topological tree structure with the entrance as the root node and the road segment as the node is constructed. The trajectory traffic volume of each road segment is used as the basis for tree structure pruning. In this way, the correct topological connection relationship is obtained, and the overpass road network is corrected accordingly.

[0090] This embodiment uses trajectory data to study the most complex overpass generation problem in road network construction, and extracts information such as its geometric shape, topological connection, and road direction, laying the foundation for the automatic generation of large-scale road networks.

[0091] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships, characterized in that: include: Obtain trajectory data within the overpass area; Based on the trajectory data, a density grid map and corresponding skeleton lines of each direction cluster and the overall trajectory are obtained to obtain the geometric shape data of the overpass road; Define several directional domains and set the vector centerline rule of the centerline for each directional domain; Determine the direction domain corresponding to the density grid map of each direction cluster, and perform directed vectorization on each density grid map based on the corresponding centerline vectorization rule to obtain the directed vector centerline of each direction; Based on the directed vector centerline corresponding to the overall trajectory, the directed vector centerlines in each direction are connected to obtain an initial directed vector overpass geometric road map; Setting an entrance and exit range of the overpass, and selecting a high-confidence trajectory in the trajectory data based on the set entrance and exit range of the overpass; Based on the selected high-confidence trajectory and the initial directed vector overpass geometric road map, the matching information between each trajectory point and the road is obtained by combining the map matching method; The matching information is topologically checked and corrected to generate a directed three-dimensional vector overpass road.

2. A method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The step of obtaining the density grid map and the corresponding skeleton line of each direction cluster and the overall trajectory based on the trajectory data specifically includes: According to the orientation of the trajectory points, the trajectory data is split into different direction clusters; A density grid map is constructed for each direction cluster and the overall trajectory, and the skeleton lines of each density grid map are extracted.

3. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The definition of several direction domains specifically includes: The directions are divided into four direction domains, namely "north, south, east, and west", and a vector centerline rule is set for each direction domain; each direction domain contains 180°, and there is overlap between the direction domains.

4. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The process of obtaining the center line of the directed vector in each direction specifically includes: Determine the direction domain to which the density grid map of each direction cluster belongs. If the direction cluster range of the density grid map is completely contained in one direction domain, the center line is processed in a directed vector according to the vector center line rule of the corresponding direction domain; if the density grid map belongs to two direction domains, the center line is processed in a directed vector based on the vector center line rules of the two direction domains.

5. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The method of connecting the directed vector center lines in various directions based on the directed vector center lines corresponding to the overall trajectory specifically includes: The binary images corresponding to the center lines of the directional vectors in various directions are superimposed, and pixels containing information of more than two directions are extracted, and the extracted pixels are vectorized to obtain a number of vector polygons; Each vector polygon is superimposed and analyzed with the directed vector center lines in each direction to determine the interrupted road pairs; based on the skeleton line corresponding to the overall trajectory, the interrupted road pairs and the intermediate connecting parts are merged into a road with only two endpoints, and the redundant parts are deleted to complete the construction of the initial directed vector overpass geometric road map.

6. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The screening of high confidence trajectories in the trajectory data specifically includes: The trajectory data is processed by removing stop points, processing sampling point loss, and eliminating abnormal sampling frequencies to obtain preprocessed trajectory data; The entrance and exit range of the overpass is set, and based on the set entrance and exit range of the overpass, the trajectory passing through the exit range of the starting point and the end point is selected in the preprocessed trajectory data; Set a buffer zone for each road and calculate the ratio of the length of the trajectory passing through the buffer zone to the total length of the trajectory; Calculate the ratio of each trajectory, retain the trajectories with a ratio greater than the preset threshold, and remove the trajectories with a ratio less than the preset threshold.

7. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The method of combining the map matching method to obtain the matching information between each track point and the road specifically includes: When there is a problem with the road network that causes the matching to be interrupted, the track is segmented with the interruption point as the segmentation point, the unmatched road segments are marked, and the subsequent track segments are matched again until the entire track is matched and the matching road segment sequence is output; Calculate the projection point of the trajectory point on the road section and the corresponding front and rear adjacent node positions to determine the direction of the road section; compare the direction of the trajectory point with the direction of the road section. If the difference between the direction of the trajectory point and the direction of the road section is within the preset threshold range, retain and construct a candidate road section set; if the difference between the direction of the trajectory point and the direction of the road section is not within the preset threshold range, remove it; The HMM algorithm is improved to allow all projection points of a point to participate in subsequent matching calculations.

8. The method for generating a multi-layer structure of an overpass taking into account directed topological connection relationships according to claim 1, characterized in that: The topological checking and correction of the matching information specifically includes: The trajectories containing unmatched points are divided into multiple categories according to the serial numbers of the sections passed by the trajectories; Apply the best representative merging method to select a representative trajectory from each category as the classic trajectory; Based on the classic trajectory, the intersection of the adjacent road sections before and after the mismatched trajectory point is intercepted as the new road, and the road network nodes are supplemented to ensure that the direction of the new road is consistent with the trajectory travel direction; Eliminate the false road sections without matching information generated during the raster-to-vector conversion process; Based on the road connectivity sequence of the long trajectory, a topological tree structure with the entrance as the root node and the road section as the node is constructed; the trajectory traffic of each section is used as the pruning basis, the sections with the same traffic sequence are merged, and the redundant connections are deleted to obtain the correct topological connection relationship; according to the obtained correct topological connection relationship, the overpass road network is corrected.

Images