Community refined road network extraction method and device, and storage medium

By acquiring and filtering crowdsourcing trajectory data, identifying trajectory flow patterns, and using multi-level confidence slack constraints and pedestrian trajectory flow semantics, the problems of complexity and inaccuracy of community road network extraction are solved, achieving more accurate and high coverage Community road network generation.

CN119089308AActive Publication Date: 2024-12-06CHINA UNIV OF GEOSCIENCES (WUHAN)
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
CN202411218752.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-02
Publication Date
2024-12-06
Estimated Expiration
2044-09-02

AI Technical Summary

Technical Problem

The prior art faces GPS trajectory interference, trajectory data sparseness and non-homogeneity caused by low-frequency sampling, and the increased complexity of pedestrian free passage characteristics when acquiring community-level roads, resulting in complex and inaccurate extraction of community road networks.

Method used

By obtaining crowdsourcing trajectory data, building an initial trajectory flow set, filtering high confidence trajectory flow, identifying trajectory flow patterns, segmenting the road and free pass areas, using multi-level confidence slack constraint strategy and pedestrian trajectory flow semantics, to generate a community Road network.

Benefits of technology

Accurate extraction of community road networks is achieved, noise interference is suppressed, road structure depiction is enhanced, pedestrian movement characteristics are adapted to community-scale pedestrian movement characteristics, and higher accuracy and coverage are achieved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089308A_ABST
    Figure CN119089308A_ABST
Patent Text Reader

Abstract

The invention relates to the field of road extraction, and discloses a community refined road network extraction method and device and a storage medium, and the method comprises the steps: data acquisition: obtaining a crowdsourcing trajectory data source; initial trajectory flow set construction: constructing an initial trajectory flow set according to the crowdsourcing trajectory data source; track flow screening: determining a confidence level of a track flow set; according to the speed value of the trajectory flow, establishing speed confidence hierarchical setting in a segmented manner, and extracting a trajectory flow set with high confidence; identifying a track flow set mode: classifying a track flow set into a road mode and a free passing mode by setting a pedestrian group correlation index threshold; and extracting community road network information: analyzing forms of classification results of the trajectory flow sets in different modes, extracting community linear road elements and planar road area elements, and organizing into community road sections and free passage area elements. The method has the advantages that the structure and the connection relation of the road network can be described more accurately, and the method is not affected by individual track data.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of road extraction, and in particular to a method, device and storage medium for extracting a community refined road network. Background Art

[0002] Urban navigation is a basic cognitive task in people's daily lives, and its accuracy and completeness are crucial to the travel or navigation experience. In the navigation process, large and medium-scale urban road network data facilities provide reliable navigation information support. However, at the micro-map scale, there is still a huge gap in navigation map data and services in urban community spaces, which are the "nerve endings" of the last mile of navigation activities. This has become a prominent and urgent weak link in seamless urban navigation.

[0003] Many mapping companies, such as Google, Here, TomTom, Baidu, and AutoNavi, have built detailed MVR networks, but the community microscale road network for navigation services has not been considered in detail. At present, domestic urban community roads are not mapped, and the existing maps are poorly consistent with the real community navigation scenes, resulting in a serious lack of navigation applications and service capabilities in urban community spaces and navigation cognition confusion in community microspace scenes. As a result, the difficulty of pedestrian navigation in community spaces is significantly higher than that in spaces outside the community, which hinders the in-depth development of community map-related applications.

[0004] Existing road extraction methods face the following challenges when acquiring community-level roads: (1) The environmental characteristics of the community space (such as dense high-rise buildings and trees) cause interference with GPS trajectories, which leads to unavoidable noise or drift in the GPS trajectories during sampling, exacerbating the difficulty of generating potential trails. The interference of these trajectory drifts and noises results in a large number of unnecessary over-estimations of bends in the final generated roads.

[0005] (2) The micro-scale characteristics of community space (short road length, close road spacing, and dense arrangement) make it impossible for existing methods to deal with the sparsity and inhomogeneity of trajectory data caused by low-frequency sampling. For example, when facing the community-scale low-frequency trajectory data source, due to low-frequency sampling, the difference in the motion characteristics of pedestrian trajectories at road segments and intersections will be greatly reduced, resulting in recognition errors of road segments and intersections. In addition, relying solely on individual trajectories is not enough to characterize the intermediate clues of the characteristics of small paths in the community, which exacerbates the difficulty of capturing the movement characteristics of people under low-frequency sampling, making the task of generating road networks at the community level more difficult.

[0006] (3) The free passage of pedestrians in community space, that is, there are usually a large number of non-road scenes within the community where pedestrians can freely pass and stay, which increases the complexity of the community road extraction task. At the method level, existing methods fail to specifically distinguish between roads in the community and other areas where pedestrians can move, so it is easy to mistakenly generate roads that do not actually exist in non-road scenes. This undoubtedly increases the difficulty of generating accurate community-level road networks. Summary of the invention

[0007] The purpose of the present invention is to propose a community refined road network extraction method, device and storage medium to solve the technical problems that the existing community road network extraction process is complicated and the extraction results are inaccurate.

[0008] The present invention provides a community refined road network extraction method, comprising the following steps: S1. Data acquisition: obtaining crowdsourcing trajectory data source; S2, initial trajectory flow set construction: construct the initial trajectory flow set based on the crowdsourced trajectory data source; S3, track flow screening: determine the confidence level of the track flow set; establish a speed confidence layer setting according to the speed value of the track flow segment by segment, and extract the track flow set with high confidence; S4, identifying trajectory flow set modes: by setting the threshold of pedestrian group correlation index, the trajectory flow set is classified into road mode and free passage mode; S5. Extract community road network information: Analyze the morphology of the trajectory flow set classification results under different modes, extract the community linear road elements and surface road area elements, and organize them into community road segments and free passage area elements. A storage medium storing instructions and data for implementing a community refined road network extraction method.

[0009] A community refined road network extraction device comprises: a processor and the storage medium; the processor loads and executes instructions and data in the storage medium to implement a community refined road network extraction method.

[0010] The beneficial effects provided by the present invention are: 1. The present invention establishes a multi-level confidence relaxation constraint strategy to flexibly handle the interference of trajectory noise. By constraining the pedestrian trajectory data by high-confidence division, the interference of low-confidence trajectory noise is suppressed, so that the trajectory flow clustering can highlight the potential road object set with significant linear characteristics; based on the clustering range, low-confidence level data relaxation is performed to more accurately quantify the calculation mode classification features; by modulating the input trajectory data source, the trajectory data is layered to adapt to tasks of different semantic scales, so that the distinction between the trajectories of two different patterns of people is more accurate, and the optimal balance between accuracy and coverage is achieved; 2. The present invention introduces pedestrian trajectory flow semantics for the generation of road networks to overcome the semantic ambiguity problem of low-frequency trajectory data. By enhancing the representation of pedestrian movement features through trajectory flow, the structure and connection relationship of the road network can be more accurately depicted than traditional methods without being affected by individual trajectory data, and it can better adapt to the complexity brought by the random movement of pedestrians in special scenarios at the community scale. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a schematic flow chart of the method of the present invention; Figure 2 This is a schematic diagram of community refined linear road extraction according to an embodiment of the present invention; Figure 3 This is a schematic diagram of community refined planar road extraction according to an embodiment of the present invention; Figure 4 It is a working schematic diagram of the hardware device of the present invention. DETAILED DESCRIPTION

[0012] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0013] Before formally describing the present invention, the scheme of the present invention is first generally described for easy understanding.

[0014] Please refer to Figure 1 , Figure 1 It is a schematic flow diagram of the method of the present invention; The present invention provides a community refined road network extraction method, comprising the following steps: S1. Data acquisition: obtaining crowdsourcing trajectory data source; It should be noted that the crowdsourcing trajectory data source consists of pedestrian trajectory data, including: pedestrian ID and pedestrian position coordinate information (x, y) at different times. The pedestrian trajectory data is expressed as: pti (ID, x, y, t), where t represents the time corresponding to the i-th record.

[0015] As an example, the crowdsourced trajectory data source uses different users' GPS trajectory data collected by mobile phones.

[0016] S2. Construction of the initial trajectory stream set: Construct an initial trajectory stream set according to the crowdsourced trajectory data source; It should be noted that in step S2, the pedestrian trajectory data is converted into trajectory stream records, and the trajectory stream records with abnormal time intervals are removed to form an initial trajectory stream set.

[0017] As an example, step S2 is specifically as follows: S21: Extract all pedestrian ID sets from the pedestrian trajectories and perform a deduplication operation; S22: For each pedestrian ID, obtain the time and position coordinate information of all trajectory points belonging to this ID, and organize them into the pedestrian trajectory tr = {pt1, pt2,..., ptn} of this ID, where pti (1 ≤ i < n) represents the spatio-temporal coordinates of the i-th trajectory point, and all trajectory points in tr are arranged in chronological order to obtain tr ’ = {pt1, pt2,..., ptn}, where when i < j, pti (t) ≤ ptj (t); S23: Take out all adjacent trajectory point pairs {<pti, pti+1>| 1 ≤ i < n} in tr ’ where the spatio-temporal coordinates of the previous trajectory are pti, and the spatio-temporal coordinates of the next trajectory are pti+1, and obtain all trajectory streams {f (pti,pti+1) | 1 ≤ i < n} of this ID; S24: Repeat the operations of S22 and S23 until all pedestrian IDs are processed, and the construction of the initial trajectory stream set is completed.

[0018] S3. Trajectory stream screening: Determine the confidence level of the trajectory stream set; Establish a speed confidence level stratification setting by segmenting according to the speed values of the trajectory streams, and extract the trajectory stream set with high confidence; As an example, step S3 is specifically as follows: S31: Establish a speed confidence calculation formula and normalize the confidence value to the range of 0 to 1; Among them, in step S31, the speed confidence calculation formula is as follows:

[0019] where v represents the speed of the trajectory stream, and The values of are determined according to the concentrated interval of the moving speed of the crowd in the community.

[0020] S32: Calculate the speed value of the trajectory flow, substitute it into the speed confidence formula to calculate and obtain a speed confidence value set of all trajectory flows; S33: Extract speed confidence value Greater than As the high confidence trajectory flow, The speed confidence threshold is set.

[0021] S4, identifying trajectory flow set modes: by setting the threshold of pedestrian group correlation index, the trajectory flow set is classified into road mode and free passage mode; As an embodiment, step S4 is specifically as follows: S41: Calculate the spatial geometric similarity of high-confidence trajectory flows, cluster and merge trajectory flows according to the similarity. Use the KNN clustering method to spatially cluster the trajectory flows, and classify high-confidence trajectory sets with adjacent positions and similar geometric features into the same trajectory cluster; Specifically, the trajectory flow is abstracted as a common line segment, and the trajectory flow distance is the shortest Euclidean distance between the line segment centers; the K nearest neighbor (KNN) method is used to select the flows with adjacent distances, and the similarity is measured according to their parallelism and length; Convert any two flows belonging to the same neighborhood to the same origin O. The target point Vj must be located within a short distance near Vi. Set a boundary circle to identify similar flows whose target points are within the circle. To ensure symmetry, the circle is centered on the target point of the longer flow. Since using the same threshold r for flows of different lengths will lead to incorrect judgments, it is necessary to dynamically adjust r according to the length of the flow. By setting the threshold r, the dissimilarity sd between a group of trajectory flows f1 and f2 belonging to different trajectory clusters can be obtained. 12 , the dissimilarity determines whether the clusters of f1 and f2 can be merged; S42: Select a certain number of trajectory flows that are spatially clustered and group them together to ensure that the trajectory flows in this group are concentrated in the same section of community road space or place space; Specifically, the confidence relaxation process of the trajectory set is performed using the buffer zone. This process requires the set expansion of each trajectory flow with a radius of R. If the expanded buffer zones are adjacent to each other, they are fused, and the low-confidence trajectory sets contained in the fused buffer zones are merged into the same group. S43: Calculate the similarity of trajectory behaviors of different individuals in the group, only consider the first K individuals closest to the target individual i, and consider these K individuals to be in the neighborhood of i. By setting the group correlation threshold γ, they are divided into two types of trajectory patterns; Specifically, the speed similarity between individuals in the neighborhood is calculated, and the speed similarity calculation expression is:

[0022] Where C(i,j) represents the velocity correlation between individuals i and j in the neighborhood; The similarity matrix W is constructed using the speed correlation between different individuals to calculate the group similarity Φ. The group similarity calculation expression is:

[0023] Where z is the regularization coefficient, which is used to converge the final result; Set z to 0.5 / K so that the range of Φ converges to between 0 and 1. Set the group correlation threshold γ to distinguish between two types of trajectory sets with different modes. The trajectory flow sets with group similarity greater than γ are classified as road modes, and the remaining trajectory flow sets are classified as free passage modes.

[0024] S5. Extract community road network information: Analyze the morphology of the trajectory flow set classification results under different modes, extract the community linear road elements and surface road area elements, and organize them into community road segments and free passage area elements. As an embodiment, step S5 is specifically as follows: According to the classification result of step S4, the pedestrian trajectory flow set of the road mode is used to generate the community linear road according to steps S51 to S31, and the trajectory flow set of the free passage mode is used to generate the community surface road according to step S53; S51: Generate a grayscale image with density features using the density of the trajectory flow, process the GPS data set into trajectory flows, and perform preliminary outlier filtering on these trajectory flows based on the building footprint data, remove those unreasonable trajectory flows by filtering outliers, such as those that appear to cross buildings, extract the parts of the filtered trajectory flow divided into road modes using S4, and convert them into density-based discretized images through kernel density estimation, that is, compress the continuous GPS trajectory data into a two-dimensional density map; Specifically, the GPS data set processed into trajectory flow will first be filtered for outliers according to the building footprint data to eliminate abnormal trajectory flows that pass through buildings. The filtered trajectory flow will be converted into a density-based discretized image through kernel density estimation. In the kernel density discretization process, the complete GPS trajectory is compressed into a two-dimensional density estimate. This process divides the experimental area into 1×1 meter cells and calculates the number of trajectory flows passing through each cell to generate a two-dimensional histogram. This histogram is then convolved with a normal distribution function N(0; σ2) to simulate the expected GPS error distribution. S52: Binarize the image after Gaussian kernel smoothing, use a set of grayscale thresholds to generate a binary mask of the road, the algorithm repeatedly performs binary skeletonization operations with different density thresholds, and uses a thinning algorithm to calculate the skeleton of the road network for the binary graph, adding the newly generated road segments to the skeleton each time. In the topological thinning step, the initial map generated by density processing is updated according to the trajectory matched by the map, and all edges that are crossed zero or only once are discarded. When there are multiple crossings of the edge generated based on the trajectory, the two intersections are merged, and each pixel is associated with an edge. The Douglas-Peucker algorithm is used to generate the edges that constitute the shape of each road segment until all grayscale thresholds are executed. Finally, a complete linear road network is obtained; S53: Extract the OD vertices of each trajectory in the free passage mode, and construct a vertex index for each vertex. Calculate a simple set of boundaries based on the vertex position and alpha value, obtain its circumscribed circle radius r from the vertex set, set the alpha-shaped edge within the range of r, loop through the vertex index and circumscribed circle radius, and use the radius filter to add the edge set to the parameter edge. If the number of vertices is greater than 3, return the set of surrounding edges and use the edge set to construct a polygon. Otherwise, use the vertices to directly create a polygon, and the area covered by the polygon is the free passage area.

[0025] Finally, see Figure 2 and Figure 3 , Figure 2 This is a schematic diagram of community refined linear road extraction according to an embodiment of the present invention; Figure 3 This is a schematic diagram of community refined planar road extraction according to an embodiment of the present invention; See also Figure 4 , Figure 4 It is a schematic diagram of the working of the hardware device of an embodiment of the present invention, and the hardware device specifically includes: a community refined road network extraction device 401, a processor 402 and a storage medium 403.

[0026] A community refined road network extraction device 401: The community refined road network extraction device 401 implements the community refined road network extraction method.

[0027] Processor 402: The processor 402 loads and executes the instructions and data in the storage medium 403 to implement the community refined road network extraction method.

[0028] Storage medium 403: The storage medium 403 stores instructions and data; the storage medium 403 is used to implement the community refined road network extraction method.

[0029] The beneficial effects of the present invention are: 1. The present invention establishes a multi-level confidence relaxation constraint strategy to flexibly handle the interference of trajectory noise. By constraining the pedestrian trajectory data by high-confidence division, the interference of low-confidence trajectory noise is suppressed, so that the trajectory flow clustering can highlight the potential road object set with significant linear characteristics; based on the clustering range, low-confidence level data relaxation is performed to more accurately quantify the calculation mode classification features; by modulating the input trajectory data source, the trajectory data is layered to adapt to tasks of different semantic scales, so that the distinction between the trajectories of two different patterns of people is more accurate, and the optimal balance between accuracy and coverage is achieved; 2. The present invention introduces pedestrian trajectory flow semantics for the generation of road networks to overcome the semantic ambiguity problem of low-frequency trajectory data. By enhancing the representation of pedestrian movement features through trajectory flow, the structure and connection relationship of the road network can be more accurately depicted than traditional methods without being affected by individual trajectory data, and it can better adapt to the complexity brought by the random movement of pedestrians in special scenarios at the community scale.

[0030] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A community refined road network extraction method, characterized by: The method comprises the following steps: S1. Data acquisition: obtaining crowdsourcing trajectory data source; S2, initial trajectory flow set construction: construct the initial trajectory flow set based on the crowdsourced trajectory data source; S3, track flow screening: determine the confidence level of the track flow set; establish a speed confidence layer setting according to the speed value of the track flow segment by segment, and extract the track flow set with high confidence; S4, identifying trajectory flow set modes: by setting the threshold of pedestrian group correlation index, the trajectory flow set is classified into road mode and free passage mode; S5. Extract community road network information: Analyze the morphology of the trajectory flow set classification results under different modes, extract the community linear road elements and surface road area elements, and organize them into community road segments and free passage area elements.

2. A community refined road network extraction method as claimed in claim 1, characterized in that: The crowdsourcing trajectory data source consists of pedestrian trajectory data, including pedestrian ID and pedestrian position coordinate information (x, y) at different times. The pedestrian trajectory data is expressed as: pti (ID, x, y, t), where t represents the time corresponding to the i-th record.

3. A community refined road network extraction method as claimed in claim 2, characterized in that: In step S2, the pedestrian trajectory data is converted into trajectory flow records, and the trajectory flow records with abnormal time intervals are removed to form an initial trajectory flow set.

4. A community refined road network extraction method as claimed in claim 3, characterized in that: Step S2 is specifically as follows: S21: extract all pedestrian ID sets from pedestrian trajectories and perform a deduplication operation; S22: For each pedestrian ID, obtain the time and position coordinate information of all trajectory points belonging to the ID, organize them into the pedestrian trajectory tr = {pt1, pt2,…, ptn} of the ID, where pti (1 ≤ i < n) represents the time and space coordinates of the i-th trajectory point, and arrange all trajectory points in tr in time sequence to obtain tr ’ = {pt1, pt2,…, ptn}, where pti (t) ≤ ptj (t) when i < j; S23: Remove tr ’ All adjacent trajectory point pairs {< pti, pti+1> | 1 ≤ i < n} in , where the spatiotemporal coordinates of the previous trajectory are pti and the spatiotemporal coordinates of the next trajectory are pti+1, obtain all trajectory flows of this ID {f (pti,pti+1) | 1 ≤ i < n}; S24: Repeat S22 and S23 until all pedestrian IDs are processed and the initial trajectory flow set is constructed.

5. The method for extracting a community refined road network according to claim 1, characterized in that: Step S3 is specifically as follows: S31: Establish a speed confidence calculation formula and normalize the confidence value to a range of 0 to 1; S32: Calculate the speed value of the trajectory flow, substitute it into the speed confidence formula to calculate and obtain a speed confidence value set of all trajectory flows; S33: Extract speed confidence value Greater than As the high confidence trajectory flow, The speed confidence threshold is set.

6. A community refined road network extraction method as claimed in claim 1, characterized in that: In step S31, the speed confidence calculation formula is as follows: Where v represents the velocity of the trajectory flow, and The value of is determined according to the concentrated range of the speed of people moving in the community.

7. A community refined road network extraction method as claimed in claim 6, characterized in that: Step S4 is specifically as follows: S41: Calculate the spatial geometric similarity of high-confidence trajectory flows, cluster and merge trajectory flows according to the similarity; use the KNN clustering method to spatially cluster the trajectory flows, and classify the high-confidence trajectory sets with adjacent positions and similar geometric features into the same trajectory cluster; S42: Select a certain number of trajectory flows that are spatially clustered and group them together to ensure that the trajectory flows in this group are concentrated in the same section of community road space or place space; S43: Calculate the similarity of trajectory behaviors of different individuals in the group, only consider the first K individuals closest to the target individual i, and assume that these K individuals are in the neighborhood of i. By setting the group correlation threshold γ, they are divided into two types of trajectory patterns.

8. The method for extracting a community refined road network according to claim 1, characterized in that: Step S5 is specifically as follows: S51: Generate a density feature grayscale image according to the density of the trajectory flow; S52: The density feature grayscale image is smoothed and binarized using the Gaussian kernel image processing algorithm. A set of grayscale thresholds are used to generate a binary mask of the road. The algorithm repeatedly performs the binary skeletonization operation, iteratively adds the detected road segments to the skeleton, associates each pixel with an edge, and uses the Douglas-Peucker algorithm to simplify and reconstruct each road segment until all grayscale thresholds are executed, and finally a linear road network is obtained; S53: Extract the OD vertices of each trajectory in the free passage mode, and construct a vertex index for each vertex. Calculate a simple set of boundaries based on the vertex position and alpha value, obtain its circumscribed circle radius r from the vertex set, set the alpha-shaped edge within the range of r, loop through the vertex index and circumscribed circle radius, and use the radius filter to add the edge set to the parameter edge. If the number of vertices is greater than 3, return the set of surrounding edges and use the edge set to construct a polygon. Otherwise, use the vertices to directly create a polygon, and the area covered by the polygon is the free passage area.

9. A storage medium, characterized in that: The storage medium stores instructions and data for implementing a community refined road network extraction method as described in any one of claims 1 to 8.

10. A community refined road network extraction device, characterized by: include: Processor and storage medium; the processor loads and executes instructions and data in the storage medium to implement a community refined road network extraction method as described in any one of claims 1 to 8.

Citation Information

Patent Citations

  • Community-level vector road network extraction method based on mobile phone positioning data

    CN113052084A

  • Community road network extraction method combining track position and neighborhood structure supervision

    CN115620140A

  • Information visualization method and intelligent visual analysis system based on text curriculum vitae information

    US20170200125A1