Multi-source trajectory data filtering method, device and equipment

Abstract

The invention discloses a multi-source track data filtering method, device and equipment, and the method comprises the steps: building a tile route set of a tile based on a plurality of continuous track segments in the tile; based on the similarity between the tile routes in the tile route set, grouping the tile routes in the tile set to obtain at least one group of tile route subsets; and filtering the tile routes in the tile route subset based on the acquisition precision of the tile routes and the similarity density of the tile routes. On the premise that the map making precision effect is guaranteed, original multi-source trajectory data are filtered to delete low-quality data with too low collection precision or repeated data with concentrated spatial density, so that it is guaranteed that semantic objects generated during subsequent data clustering and aggregation have completeness, the computing power during subsequent data processing can be improved, and the data processing efficiency is improved. The data processing efficiency is improved, and the storage space is saved.

Description

Technical Field

[0001] The present invention relates to the field of geographic information technology, and more particularly to a method, device and equipment for filtering multi-source trajectory data. Background Art

[0002] High-precision maps can be used for autonomous driving, and the recognition and production of road signs are important components of generating high-precision maps. Road signs (semantic objects) are generally identified and extracted through road image information collected by cameras mounted on vehicles.

[0003] In the process of making high-precision maps, in order to solve the drawback of large resource investment in single-trip closed-loop collection (closed-loop single-trip collection) and optimize application scenarios, the inventor previously proposed a solution for making high-precision maps based on multi-source crowdsourcing data, that is, by using private cars as collection equipment, data collection is performed without interfering with the stable and fast driving of the vehicle. In the process of using private cars to collect data, since it is impossible to stipulate that a certain vehicle strictly follows a closed-loop trajectory, most vehicle trajectories are only one-way trajectories from the starting point to the destination, so multiple single-track data is used in the later production of high-precision maps, that is, multiple / multiple image collections are required for the same location, and then the road sign information in each collected image is extracted separately, and finally the same road sign collected in multiple trips is fused, and then a high-precision map is produced based on multi-source crowdsourcing data. Summary of the invention

[0004] In order to improve the ability to process multi-source data, and thereby improve computing power while ensuring the accuracy of high-precision maps, so as to enrich technical routes and increase selection space, a method, device and equipment for filtering multi-source trajectory data are provided in an embodiment of the present invention.

[0005] In a first aspect, an embodiment of the present invention provides a method for filtering multi-source trajectory data, which may include:

[0006] constructing a tile route set of the tile based on a number of continuous trajectory segments within the tile;

[0007] Based on the similarities between the tile routes in the tile route set, the tile routes in the tile route set are grouped to obtain at least one set of tile route subsets;

[0008] Based on the acquisition accuracy of the tile routes and the similarity density of the tile routes, the tile routes in the tile route subset are filtered.

[0009] Optionally, filtering the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes may include:

[0010] Determining whether the number of tile routes in the tile route subset is greater than a preset number threshold;

[0011] If not, no filtering is performed;

[0012] If yes, sorting the acquisition accuracy of the tile routes, and filtering the tile routes with low acquisition accuracy based on the sorting result;

[0013] Determine whether the number of tile routes in the tile route subset after the primary filtering is greater than the preset number threshold;

[0014] If so, a second filtering is performed on several tile routes with the lowest acquisition accuracy in the tile route subset after the first filtering based on the similarity density.

[0015] Optionally, performing secondary filtering on several tile routes with the lowest acquisition accuracy in the tile route subset after the primary filtering based on the similarity density may include:

[0016] Determine the similarity density values ​​of several tile routes with the lowest acquisition accuracy respectively;

[0017] The tile routes with the lowest similarity density value are filtered twice;

[0018] The above determination steps are repeated until the number of tile routes after the primary filtering and / or secondary filtering is less than the preset number threshold.

[0019] Optionally, the acquisition accuracy of the tile route is predetermined based on the following method:

[0020] Based on the quality of the acquisition accuracy of each trajectory point in the trajectory segment corresponding to the tile route, the number of trajectory points with good acquisition accuracy in the trajectory segment is counted;

[0021] The ratio of the number of trajectory points with good acquisition accuracy to the number of all trajectory points in the trajectory segment is used as the acquisition accuracy of the tile route.

[0022] Optionally, before performing secondary filtering on several tile routes with the lowest acquisition accuracy, the following may also be included:

[0023] Neighbor tile routes are determined based on the projected distances between the tile routes, so as to filter the tile routes based on the number of neighbor tile routes.

[0024] Optionally, after filtering the tile routes in the tile route subset, the following may also be included:

[0025] Based on the coverage of the tile routes in the tile route subset, the tile routes with deletion marks in the tile route subset are detected to update the filtering result.

[0026] Optionally, the detecting the tile routes with deletion marks in the tile route subset based on the coverage of the tile routes in the tile route subset to update the filtering result may include:

[0027] Splitting the tile to obtain a plurality of high-level tiles;

[0028] Traversing the tile route subset included in the tile, and determining whether there is a tile route with a deletion mark in the tile route subset;

[0029] If it exists, then count the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes;

[0030] Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes;

[0031] Determine the size of the coverage ratio and a preset coverage ratio threshold;

[0032] If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored;

[0033] If it does not exist, the filtering results will not be updated.

[0034] Optionally, the step of restoring the tile route with the deletion mark may include:

[0035] Restore the tile routes with the deletion mark and the ones with the highest increase in coverage ratio.

[0036] Optionally, the detecting the tile routes with deletion marks in the tile route subset based on the coverage of the tile routes in the tile route subset to update the filtering result may include:

[0037] Determine whether there is a tile route with a deletion mark in the tile route subset;

[0038] If yes, the tile is segmented to obtain a number of high-level tiles;

[0039] Counting the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes;

[0040] Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes;

[0041] Determine the size of the coverage ratio and a preset coverage ratio threshold;

[0042] If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored.

[0043] Optionally, constructing a tile route set of the tile based on map data collected from a plurality of continuous track segments in the tile may include:

[0044] Based on the vehicle ID and the timestamp of the track points, the track points of the tile are divided to determine the continuous track segments corresponding to the current vehicle ID;

[0045] The continuous trajectory segments are regarded as a tile route, and a tile route set of the tile is constructed.

[0046] Optionally, grouping the tile routes in the tile set based on the similarities between the tile routes in the tile route set to obtain at least one set of tile route subsets may include:

[0047] Taking the maximum value of the starting point distance and / or the ending point distance between two tile routes in the tile route set as the similarity between the two tile routes;

[0048] The maximum value of the starting point height and / or the ending point height between the two tile routes is used as the height difference between the two tile routes;

[0049] The tile routes in the tile route set whose similarity is less than a preset similarity threshold and whose height difference is less than a preset height difference threshold are taken as the tile route subset corresponding to the same actual road.

[0050] Optionally, after the tile routes in the tile set are grouped, the following steps may also be included:

[0051] Based on the projection distance, projection length and projection angle of the starting point and / or end point of the tile route included in the tile route subset, it is determined whether the tile route subsets are aggregated into the same tile route subset.

[0052] Optionally, the method may further include:

[0053] The filtering results are adjusted based on the topological relationship of tile routes between adjacent tiles.

[0054] Optionally, adjusting the filtering result based on the topological relationship of tile routes between adjacent tiles may include:

[0055] If it is determined that the filtering result of the tile route in the current tile is marked with a delete mark, and the filtering result of the tile route in the adjacent tile with a topological relationship is marked with a non-delete mark, the filtering result of the tile route in the current tile is adjusted to be marked with a non-delete mark;

[0056] or,

[0057] Based on the topological relationship between tile routes between adjacent tiles, the tile routes in adjacent tiles are filtered using the filtering result of the tile routes in the current tile.

[0058] In a second aspect, an embodiment of the present invention provides a filtering device for multi-source trajectory data, which may include:

[0059] A construction module, configured to construct a tile route set of the tile based on a plurality of continuous trajectory segments within the tile;

[0060] A grouping module, configured to group the tile routes in the tile set based on similarities between the tile routes in the tile route set to obtain at least one set of tile route subsets;

[0061] The filtering module is used to filter the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes.

[0062] In a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for filtering multi-source trajectory data as described in the first aspect.

[0063] In a fourth aspect, an embodiment of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for filtering multi-source trajectory data as described in the first aspect is implemented.

[0064] The beneficial effects of the above technical solution provided by the embodiment of the present invention include at least:

[0065] The embodiments of the present invention provide a method, device and apparatus for filtering multi-source trajectory data. The method filters the original multi-source trajectory data to delete low-quality data with too low collection accuracy or duplicate data with concentrated spatial density under the premise of ensuring the accuracy of map making. This ensures that the semantic objects generated during subsequent data clustering are complete, and can also improve the computing power during subsequent data processing, so as to improve data processing efficiency and save storage space.

[0066] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0067] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0069] Figure 1 A flowchart of a method for filtering multi-source trajectory data provided in an embodiment of the present invention;

[0070] Figure 2 A schematic diagram of the overall architecture provided in an embodiment of the present invention;

[0071] Figure 3 The specific execution flow chart of step S12;

[0072] Figure 4 The adjacent distance distribution of the boundary points entering the Tile is counted in the embodiment of the present invention;

[0073] Figure 5 The distribution of the adjacent distances of the boundary points leaving the Tile is counted in the embodiment of the present invention;

[0074] Figure 6 The distribution of adjacent height differences of boundary points entering / leaving a Tile is counted in the embodiment of the present invention;

[0075] Figure 7 The specific execution flow chart of step S13;

[0076] Figure 8 A box plot of the number of items when the quantity threshold is determined in an embodiment of the present invention;

[0077] Fig. 9 This is a schematic diagram of the effect after filtering in an embodiment of the present invention;

[0078] Fig.10 This is a tile route filtering effect diagram corresponding to different actual roads in an embodiment of the present invention;

[0079] Fig.11 One of the flowcharts for the specific execution of step S14;

[0080] Fig.12 The second flowchart specifically executes step S14;

[0081] Fig.13 A schematic diagram of adjusting filtering results based on Tile topological relationships in an embodiment of the present invention;

[0082] Fig.14 A flowchart of a detailed multi-source trajectory data filtering method provided in an embodiment of the present invention;

[0083] Fig.15 Schematic diagram of the structure of a filtering device for multi-source trajectory data provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0084] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0085] When making high-precision maps, the imported multi-source data needs to be filtered to reduce the amount of data and improve the overall execution efficiency. Since there is currently no direct filtering method for multi-source data (trajectory), and the existing trajectory segmentation and grouping statistics highly overlapped trajectories, it is necessary to consider the similarity of time, space, and even driving speed at the same time. If there are too many grouping dimensions, the data will be diluted to a certain extent, and the effect will not be good if the data base is not very large. In addition, although the similarity of each group is more stringent when there are more dimensions, the calculation complexity is also higher and the performance is slower.

[0086] In view of the above technical problems, an embodiment of the present invention provides a method for filtering multi-source trajectory data. The method can filter the imported data while ensuring that the collected road sections are covered as much as possible, and filter the road section data that is repeatedly collected and has a high density, so as to minimize the impact on subsequent module processing and improve the operating efficiency. Figure 1 As shown, the method may include the following steps:

[0087] Step S11: construct a tile route set of the tile based on a number of continuous trajectory segments in the tile.

[0088] Step S12: grouping the tile routes in the tile set based on the similarities between the tile routes in the tile route set to obtain at least one set of tile route subsets.

[0089] Step S13: filtering the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes.

[0090] Reference Figure 2 As shown, the method can be integrated into a filtering module, that is, a filtering module and a module for importing filtered data are added on the basis of the original architecture to realize filtering processing of the imported data. In order to simplify the data collected by the vehicle, the above method provided in the embodiment of the present invention deletes a large amount of duplicate data by means of trajectory segmentation, grouping, spatial similarity density calculation, etc., while ensuring the recall rate as much as possible. On the one hand, it ensures that the final data is complete and has a guaranteed effect, and on the other hand, it greatly improves the execution efficiency and computing storage space of the subsequent algorithm.

[0091] In an optional embodiment, the above step S11 is to pre-process the data to extract several continuous trajectory segments in each tile (Tile), and then form a tile route with the trajectory segments, and use the tile route as the smallest unit for subsequent steps. It should be noted that this step can process data from different Tiles in parallel, and of course there can be a sequence, which is not specifically limited in this embodiment. During the specific execution, first, based on the vehicle ID and the trajectory point timestamp, the trajectory points of the tile are divided to determine the continuous trajectory segments corresponding to the current vehicle ID; then, the continuous trajectory segments are used as a tile route to construct a tile route set for the tile.

[0092] In a specific example, the input data is: all track points corresponding to a single car ID in a single Tile. Then, when processing, all track points (single Tile & single car ID) are sorted in ascending order by the timestamp field; two adjacent track points of non-continuous travel are found, and the first point is used as the end point of the previous tile route, and the second point is used as the starting point of the next tile route. In this way, several continuous travel track segments corresponding to a single car ID in a single Tile are output, that is, several tile routes of a single car ID in a certain Tile.

[0093] In another optional embodiment, the above step S12 groups the tile routes of all car IDs in each Tile according to their driving similarities, and each group approximately represents a section of road in the tile. It should be noted that in this embodiment of the present invention, considering the computational cost, the grouping is only based on the road level, and is no longer subdivided into the lane level. Figure 3As shown, during execution, the following steps may be specifically included:

[0094] Step S31: taking the maximum value of the starting point distance and / or the ending point distance between any two tile routes in the tile route set as the similarity between any two tile routes.

[0095] Calculate the similarity of two tile routes in the tile route set corresponding to a single Tile, Max (starting point distance, end point distance) is as follows:

[0096]

[0097]

[0098]

[0099] Among them, R i is the i-th tile route in Tile;

[0100] is the starting point distance between the i-th and j-th tile routes;

[0101] is the end point distance between the i-th and j-th tile routes;

[0102] d ij is the similarity between the i-th and j-th tile routes.

[0103] Step S32: taking the maximum value of the starting point height and / or the ending point height between any two tile routes as the height difference between any two tile routes.

[0104] Calculate the height difference between two tile routes in the tile route set corresponding to a single Tile. Max (starting point height difference, end point height) is as follows:

[0105]

[0106]

[0107]

[0108] Among them, R i is the i-th tile route in Tile;

[0109] is the starting point height difference between the i-th and j-th tile routes;

[0110] is the height difference between the endpoints of the i-th and j-th tile routes;

[0111] h ijis the height difference between the i-th and j-th tile routes.

[0112] Step S33: The tile routes in the tile route set whose similarity is less than a preset similarity threshold and whose height difference is less than a preset height difference threshold are taken as the tile route subset corresponding to the same actual road.

[0113] The screening method in this step can be achieved by the following formula:

[0114]

[0115] Among them, D is the similarity threshold, which can be 12m; H is the height difference threshold, which can be 2m.

[0116] In the embodiment of the present invention, the tile routes that meet the above similarity threshold and height difference threshold can be considered that the two tile routes enter / leave the tile through the same road. In the embodiment of the present invention, the inventors counted the adjacent distance distribution of the boundary points of leaving / entering the tile when the continuous trajectory crossed the tile boundary (refer to Figure 4 and Figure 5 ). Since the verification data is relatively dense, all lanes of most roads have trajectories. 4m can be roughly regarded as the width of a single lane. Considering that there are cases where single lane trajectories on the same road are missing in actual situations, and there is a certain angle between the actual direction of the road and the tile boundary, the threshold of the boundary point of the same road trajectory is located at 12m, that is, the D value is taken as 12m. Similarly, refer to Figure 6 As shown, if the height difference between the start / end points of the two tile routes is within 2 m, they can be considered as the same road in the viaduct scenario. If it is greater than 2 m, they are considered as different roads in the viaduct scenario.

[0117] In another alternative embodiment, referring to Figure 3 As shown, after the above step S12 groups the tile routes in the tile set, the following steps may also be included, namely: step S34, based on the projection distance, projection length and projection angle of the starting point and / or end point of the tile route included in the tile route subset, determining whether the tile route subsets are aggregated into the same tile route subset.

[0118] In the specific implementation of this step, the starting point and the end point of a tile route subset are selected, and projected to another tile route subset to obtain two projection distances, and the maximum value is taken as the projection distance. Based on the projection vector constructed from the projection point of the starting point to the projection point of the end point, a trajectory vector is constructed between the starting point and the end point, and the angle between the projection vector and the trajectory vector is obtained as the projection angle. The tile route subset obtained in the above step S33 is specifically processed as follows:

[0119] Take the distance similarity of two tile route subsets, max(starting point projection distance, end point projection distance) as follows:

[0120]

[0121]

[0122]

[0123]

[0124]

[0125] Among them, Point i start : The starting point of the i-th tile route in Tile

[0126] Point i end : The end point of the i-th tile route in Tile

[0127] Line j : The jth tile route in Tile

[0128] PD ij start : The shortest distance from the starting point of the i-th tile route to the j-th tile route (projection distance)

[0129] PD ij end : The shortest distance between the end point of the i-th tile route and the j-th tile route (projection distance)

[0130] pd ij : The similarity of the projection distance between the i-th tile route and the j-th tile route

[0131] Projection length:

[0132]

[0133]

[0134] pl ij =max(PL ij ,PL ji )

[0135] Among them, Point i start : The starting point of the i-th tile route in Tile

[0136] Point iend : The end point of the i-th tile route in Tile

[0137] Line j : The jth tile route in Tile

[0138] PL ij start : The distance from the starting point of the i-th tile route to the projection point on the j-th tile route and the starting point of the j-th tile route

[0139] PL ij end : The distance from the projection point of the end point of the i-th tile route on the j-th tile route to the starting point of the j-th tile route

[0140] PL ij : Similarity of the projection length of the i-th tile route on the j-th tile route

[0141] L i : The length of the i-th tile route

[0142] pl ij : Similarity of projection length between the ith tile route and the ith tile route

[0143] Projection angle similarity, the angle between the projection vector and the trajectory vector:

[0144]

[0145]

[0146]

[0147] pa ij =min(PA ij ,PA ji )

[0148] Among them, Point ij tart : The starting point of the i-th tile route in Tile

[0149] Point i end : The end point of the i-th tile route in Tile

[0150] Line j : The jth tile route in Tile

[0151] Point ij start : The starting point of the i-th tile route in the Tile is projected onto the j-th tile route

[0152] Point ij end : The end point of the i-th tile route in the Tile is projected on the j-th tile route

[0153] PA ij : The angle between the vector formed by the starting point and end point of the i-th tile route in the tile and the vector formed by its projection point on the j-th tile route

[0154] pa ij : The similarity of the projection angle between the i-th tile route and the j-th tile route

[0155] Projection matching rule, if the distance similarity of two tile routes is less than or equal to PD, the projection length similarity is greater than or equal to PL, and the projection angle similarity is less than or equal to PA, then the two tile routes meet the aggregation requirements, where PD = 12, PL = 0.8, PA = 90.

[0156]

[0157] When executing the above step S34, it is preferred to merge from the subset with a small number of tile routes to other subsets, and judge from the one with the largest number of tiles. This judgment logic is to prevent the merge operation from being performed to prevent the group that should be merged into one group from being merged and missing. The specific process is: sort the divided groups (subsets) from small to large (the number of tile routes in the subset); the group with the least number of tile routes is preferentially merged into the group with more tile routes for judgment.

[0158] In a specific example, after the intra-tile route grouping in step S33, four groups are divided, namely, group A (containing 2 tile routes), group B (containing 3 tile routes), group C (containing 4 tile routes), and group D (containing 5 tile routes). At this time, the 2 tile routes in group A are first sent to the first-ranked group D (containing 5 tile routes) to perform the above projection matching rule judgment: if the 2 tile routes in group A and the tile route in group D are judged to meet the projection matching rule, group A is merged into group D; if there is a tile route that does not meet the projection matching rule when group A and group D are matched, group A and group C, which is ranked second, are judged again, and so on. If the number of tile routes in group C and group D is the same, and group A, group C, and group D all meet the projection matching rule, group A is randomly combined into group C or group D.

[0159] In another alternative embodiment, referring to Figure 7 As shown, the above step S13 may specifically include the following steps when executed:

[0160] Step S71, determine whether the number of tile routes in the tile route subset is greater than a preset number threshold; if not, execute step S72; if yes, execute step S73.

[0161] In the embodiment of the present invention, the preset number threshold T can be determined according to the actual road properties (for example, the highway and ordinary rural roads have different properties and different values), or it can be obtained by taking into account the computer operation capacity and counting the number of times similar roads are passed, referring to Figure 8 As shown, the preset quantity threshold T in this embodiment is 50, that is, the maximum number of tile routes in the same tile route subset is filtered to 50.

[0162] Step S72: No filtering is performed.

[0163] Step S73: sorting the acquisition accuracy of the tile routes, and filtering the tile routes with low acquisition accuracy based on the sorting result.

[0164] The acquisition accuracy of the tile route in this step is predetermined based on the following method: first, based on the acquisition accuracy of each trajectory point in the trajectory segment corresponding to the tile route, the number of trajectory points with good acquisition accuracy in the trajectory segment is counted; then, the ratio of the number of trajectory points with good acquisition accuracy to the number of all trajectory points in the trajectory segment is used as the acquisition accuracy of the tile route.

[0165] It should be noted here that when multiple bicycle trajectories are collected on the map roads, the signals of the trajectory points are good and bad. By counting the quality of the trajectory points included in the continuous trajectory segment corresponding to a tile route, the collection accuracy of each tile route can be obtained, and the collection accuracy can be expressed as a ratio.

[0166] In this step, one-time filtering can definitely filter out tile routes with low acquisition precision values. If the acquisition precision values ​​of all tile routes are relatively high, one-time filtering cannot be achieved.

[0167] In a specific example, if the number of tile routes included in a tile route subset G is N, if N>T, a filtering is performed, and the tile routes included in the tile subset are sorted in descending order according to the acquisition accuracy, and the acquisition accuracy of the Tth largest tile route is recorded as at, and a subset G1 is generated according to at, and the subset G1 is as follows:

[0168] G1={R i |R i ∈G,a i ≥at}

[0169] Among them, a i is the acquisition accuracy of the i-th tile route.

[0170] Step S74: determine whether the number of tile routes in the tile route subset after the primary filtering is greater than a preset number threshold. If yes, execute step S75. If no, end.

[0171] Step S75: performing secondary filtering on several tile routes with the lowest collection accuracy in the tile route subset after the primary filtering based on the similarity density.

[0172] According to the above example, the tile routes with precision equal to at are extracted from G1 to form a set G2 as follows:

[0173] G2={R i |R i ∈G1,a i =at}

[0174] In a specific embodiment, when the above step S75 is executed, the similarity density values ​​of several tile routes with the lowest acquisition accuracy are first determined respectively; then, the tile route with the lowest similarity density value is subjected to secondary filtering; finally, the above judgment steps are repeated until the number of tile routes after the first filtering and / or the second filtering is less than the preset number threshold.

[0175] In this step, the similarity density d of each tile route in the above set G2 is calculated. The specific calculation formula is as follows:

[0176] d i =avg(nsmallest(K,D i ))

[0177] Among them, d i is the similarity density value of the i-th tile route;

[0178] avg(S) is the average value of all values ​​in set S;

[0179] nsmallest(k,S) is the set consisting of the first k smallest values ​​in set S;

[0180]

[0181] D i ={d ij |R i ∈G2,R j ∈G1,j≠i}

[0182] After the above steps S73 to S75, the above filtered tile routes are marked with deletions, as follows:

[0183]

[0184] Among them, del i: The deletion flag of the i-th tile route, 1 means deletion, 0 means retention.

[0185] Reference Fig. 9 As shown, after the above filtering process, the data volume in each tile is reduced, and the mapping efficiency is improved without affecting the mapping accuracy requirements in subsequent steps.

[0186] In another optional embodiment, referring to Figure 7, before step S75 in the above step S13 performs secondary filtering on several tile routes with the lowest acquisition accuracy, the following steps may also be included, namely: step S76, determining neighbor tile routes based on the projection distance between tile routes, so as to filter the tile routes based on the number of neighbor tile routes.

[0187] In the specific implementation of this step, the neighbor track set of each tile route is extracted from the set G2 in the above example, and then the tile routes with fewer neighbor tracks are filtered out from the set G2 to form a set G3. The number of tile routes in the set G3 is counted to see whether it is greater than zero. If so, the similarity density of each tile route in the set G3 is calculated. The specific formula execution process is as follows:

[0188] neighbor(R i )={R j |R i ∈G2,R j ∈G2,pe ij ≤4}

[0189] G3={R i |R i ∈G2,|neighbor(R i )|≥5}

[0190] Among them, pe ij : Similarity of the projection distance from trajectory i to trajectory j

[0191] neighbor(Ri): the set of neighbor trajectories of trajectory i

[0192] nsmallest(K,D): a subset of the first K smallest values ​​in set D

[0193] avg(D): average value of set D

[0194] e i : Similarity density of trajectories

[0195] In another optional embodiment, the above method of grouping and filtering based on starting point and end point ignores the differences in the attributes of tile routes themselves, which may result in tile routes with the same starting point but traveling through different roads to the same end point being divided into one group (tile route set), and then the tile routes corresponding to a certain actual road may be completely deleted during the filtering process, which brings great difficulties to the subsequent production of high-precision maps. For example, referring to Fig.10 As shown, Fig.10 The tile routes in the lower part are actually data corresponding to other roads. If they are completely deleted, high-precision maps cannot be produced normally. Figure 1 As shown, after filtering the tile routes in the tile route subset in step S13, the following steps may also be included, namely: step S14, based on the coverage of the tile routes in the tile route subset, detecting the tile routes with deletion marks in the tile route subset to update the filtering result.

[0196] In a specific embodiment, referring to Fig.11 As shown, the above step S14 may specifically include the following steps when executed:

[0197] Step S111: split the tile to obtain a number of high-level tiles.

[0198] In this step, a single Tile is divided into higher-level tiles. For example, if the current Tile is level 13 (about 2127m*2435m), it is divided into higher-level tiles of level 20 (about 16.6m*19m) or level 21 (about 8.3m*9.5m).

[0199] Step S112, traverse the tile route subset included in the tile, and determine whether there is a tile route with a deletion mark in the tile route subset; if so, execute step S113.

[0200] Step S113: Count the number of high-level tiles passed by tile routes with a non-deletion mark and the number of high-level tiles passed by all tile routes in the tile route subset, so as to respectively determine the coverage of tile routes with a non-deletion mark and the coverage of all tile routes.

[0201] Step S114: Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes.

[0202] Step S115, determine the size of the coverage ratio and the preset coverage ratio threshold; if the coverage ratio is less than the preset coverage ratio threshold, execute step S116; if it does not exist, execute step S117.

[0203] Step S116: Restore the tile routes with deletion marks.

[0204] In the specific implementation, this step will restore the tile routes with a high increase in coverage ratio among the tile routes with deletion marks. In this step, for the group (subset of tile routes) with deletion mark = 1, calculate the ratio of advanced tiles passed by the tile routes with deletion mark = 0 in the group to the advanced tiles passed by all tile routes in the group. When the ratio is less than a certain percentage, select the one with the largest increase in ratio from the tile routes with deletion mark = 1 and update the deletion mark to 0 to perform the filtering recovery step.

[0205] Step S117: Do not update the filtering results.

[0206] In another alternative embodiment, referring to Fig.12 As shown, the above step S14 may further include the following steps when executed:

[0207] Step S121, determine whether there is a tile route with a deletion mark in the tile route subset; if so, execute step S122. The embodiment of the present invention is different from the above embodiment in that the judgment step is performed first, and the purpose is to reduce the amount of calculation. The practical significance is to ensure that at least a certain number of trajectories are retained on each lane through the number of neighbor tiles. High-level tiles are only used to calculate coverage, so in terms of saving calculation amount, these two steps can be reversed. That is, first determine whether there is a tile route indicated by deletion, and if so, calculate the high-level tiles.

[0208] Step S122: split the tile to obtain a number of high-level tiles.

[0209] Step S123: Count the number of high-level tiles passed by tile routes with a non-deletion mark and the number of high-level tiles passed by all tile routes in the tile route subset, so as to respectively determine the coverage of tile routes with a non-deletion mark and the coverage of all tile routes.

[0210] Step S124: Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes.

[0211] Step S125, determine the size of the coverage ratio and the preset coverage ratio threshold; if the coverage ratio is less than the preset coverage ratio threshold, execute step S126.

[0212] Step S126: Restore the tile route with the deletion mark.

[0213] In another alternative embodiment, referring to Figure 1As shown, after executing step S13 or step S14, the following step S15 may be executed, that is, step S15, adjusting the filtering result based on the topological relationship of the tile routes between adjacent tiles.

[0214] In a specific embodiment, the above step S15 may specifically include the following solutions, namely:

[0215] If it is determined that the filtering result of the tile route in the current tile is marked with a delete mark, and the filtering result of the tile route in the adjacent tile with a topological relationship is marked with a non-delete mark, the filtering result of the tile route in the current tile is adjusted to be marked with a non-delete mark. Fig.13 As shown, the lane in the Tile will have a complete route after filtering and a route to be filtered, and then the surrounding Tiles are associated, and the trajectories to be filtered that have a topological relationship with the complete route are retained according to the topological relationship between the trajectories, and the retained routes to be filtered are retained as trajectory routes within a certain range of the Tile boundary. The retention range is intercepted according to the time window T, and the trajectory data within the time range T is read backward / forward from the starting point / end point of the tile route for retention. Example: starting time = st, end time = et, then the retained time periods are [st, st+T] and [et-T, et]. T is set according to the downstream processing dependency. The above steps in this embodiment provide a reference for subsequent multi-source data alignment.

[0216] In another optional embodiment, the above step S15 can also be specifically as follows: based on the topological relationship of the tile routes between adjacent tiles, the tile routes in the adjacent tiles are filtered with the filtering result of the tile route in the current tile. For example, among three adjacent tiles, according to the topological relationship, the tile routes in the tiles before and after the same trajectory route are all retained, and the tile route in the middle tile is marked as a deletion mark, then the entire trajectory route is retained.

[0217] In a more specific example, refer to Fig.14 As shown, module 1 is a single tile, while module 2 relies on surrounding tiles. The solution can be roughly as follows:

[0218] 1. Get a single vehicle ID, sort all the track points in each tile by time, and get the track route;

[0219] 2. Cut the acquired trajectory route to obtain the cut tile route;

[0220] 3. Traverse the tile routes in each tile and obtain the similarity between any two tile routes;

[0221] 4. Group the tile routes within each tile according to the preset similarity threshold;

[0222] 5. Calculate the accuracy of the tile routes in each group, and calculate the similarity density of the tile routes that meet the accuracy requirements, and delete the tile routes in the group according to the preset quantity threshold;

[0223] 6. Calculate the road-level coverage within each tile, retain those that meet the coverage threshold, and restore the deleted roads that do not meet the threshold so that the coverage meets the requirements.

[0224] In a specific example, the data processing performance of the embodiment of the present invention is verified as follows:

[0225] Data preprocessing:

[0226] This module reads the trajectory.geojson downloaded from import (tile_id=556168550)

[0227] File size: 831085KB, number of track points: 760140

[0228] Reading + serialization + generating DataFrame object takes: 76 seconds

[0229] DataFrame coordinate system conversion (wgs84-utm) takes: 55 seconds

[0230] Tile route cutting: DataFrame is cut according to car_id, each group is sorted according to timestamp, and the sorted cutting data is cut again according to the front and rear track distance (more than 3m is considered as another group) / front and rear distance time (front and rear distance of 5 seconds is considered as another group). Total time consumption: 18 seconds.

[0231] process Execution time File read storage DataFrame 76 seconds Coordinate system conversion 55 seconds Route cutting 18 seconds

[0232] Route grouping, tile_id = 556168550, number of track points: 760140, number of "tile routes": 377. Execution time: 15.8 seconds. Route filtering, full test data:

[0233] Serial number Tile ID Group ID Number of routes (>50) 1 556168503 556168503_0 84 2 556168503 556168503_1 79 3 556168546 556168546_0 84 4 556168546 556168546_1 79 5 556168547 556168547_0 84 6 556168547 556168547_1 78 7 556168556 _0 68

[0234] Execution time: 8.7 seconds. The filtered data can ensure that the subsequent algorithms are executed within the required time range.

[0235] Based on the same inventive concept, an embodiment of the present invention further provides a filtering device for multi-source trajectory data, referring to Fig.15 As shown, the device may include: a construction module 151, a grouping module 152 and a filtering module 153, and its working principle is as follows:

[0236] The construction module 151 is used to construct a tile route set of a tile based on a number of continuous trajectory segments in the tile;

[0237] The grouping module 152 is used to group the tile routes in the tile set based on the similarity between the tile routes in the tile route set to obtain at least one set of tile route subsets;

[0238] The filtering module 153 is used to filter the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes.

[0239] In an optional embodiment, the filtering module 153 is specifically used for:

[0240] Determining whether the number of tile routes in the tile route subset is greater than a preset number threshold;

[0241] If not, no filtering is performed;

[0242] If yes, sorting the acquisition accuracy of the tile routes, and filtering the tile routes with low acquisition accuracy based on the sorting result;

[0243] Determine whether the number of tile routes in the tile route subset after the primary filtering is greater than the preset number threshold;

[0244] If so, a second filtering is performed on several tile routes with the lowest acquisition accuracy in the tile route subset after the first filtering based on the similarity density.

[0245] In another optional embodiment, the filtering module 153 is specifically used for:

[0246] Determine the similarity density values ​​of several tile routes with the lowest acquisition accuracy respectively;

[0247] The tile routes with the lowest similarity density value are filtered twice;

[0248] The above determination steps are repeated until the number of tile routes after the primary filtering and / or secondary filtering is less than the preset number threshold.

[0249] In an optional embodiment, the filtering module 153 is specifically configured to: determine neighbor tile routes based on the projection distance between the tile routes, so as to filter the tile routes based on the number of neighbor tile routes.

[0250] In another alternative embodiment, referring to Fig.15As shown, it may also include: a detection module 154, which is used to detect tile routes with deletion marks in the tile route subset based on the coverage of the tile routes in the tile route subset, so as to update the filtering result.

[0251] In another optional embodiment, the detection module 154 is specifically used for:

[0252] Splitting the tile to obtain a plurality of high-level tiles;

[0253] Traversing the tile route subset included in the tile, and determining whether there is a tile route with a deletion mark in the tile route subset;

[0254] If it exists, then count the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes;

[0255] Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes;

[0256] Determine the size of the coverage ratio and a preset coverage ratio threshold;

[0257] If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored;

[0258] If it does not exist, the filtering results will not be updated.

[0259] In another optional embodiment, the detection module 154 is further specifically configured to restore tile routes with a high increase in coverage ratio among tile routes marked with deletion marks.

[0260] In another optional embodiment, the detection module 154 is specifically used for:

[0261] Determine whether there is a tile route with a deletion mark in the tile route subset;

[0262] If yes, the tile is segmented to obtain a number of high-level tiles;

[0263] Counting the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes;

[0264] Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes;

[0265] Determine the size of the coverage ratio and a preset coverage ratio threshold;

[0266] If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored.

[0267] In another optional embodiment, the construction module 151 is specifically used for:

[0268] Based on the vehicle ID and the timestamp of the track points, the track points of the tile are divided to determine the continuous track segments corresponding to the current vehicle ID;

[0269] The continuous trajectory segments are regarded as a tile route, and a tile route set of the tile is constructed.

[0270] In another optional embodiment, the grouping module 152 is specifically configured to:

[0271] Taking the maximum value of the starting point distance and / or the ending point distance between two tile routes in the tile route set as the similarity between the two tile routes;

[0272] The maximum value of the starting point height and / or the ending point height between the two tile routes is used as the height difference between the two tile routes;

[0273] The tile routes in the tile route set whose similarity is less than a preset similarity threshold and whose height difference is less than a preset height difference threshold are taken as the tile route subset corresponding to the same actual road.

[0274] In another optional embodiment, the grouping module 152 is further specifically used to determine whether the tile route subsets are aggregated into the same tile route subset based on the projection distance, projection length and projection angle of the starting point and / or end point of the tile route included in the tile route subset.

[0275] In another alternative embodiment, referring to Fig.15 As shown, the device may further include: an adjustment module 155, and the adjustment module 155 is used to adjust the filtering result based on the topological relationship of the tile routes between adjacent tiles.

[0276] In another optional embodiment, the adjustment module 155 is specifically used for:

[0277] If it is determined that the filtering result of the tile route in the current tile is marked with a delete mark, and the filtering result of the tile route in the adjacent tile with a topological relationship is marked with a non-delete mark, the filtering result of the tile route in the current tile is adjusted to be marked with a non-delete mark;

[0278] or,

[0279] Based on the topological relationship between tile routes between adjacent tiles, the tile routes in adjacent tiles are filtered using the filtering result of the tile routes in the current tile.

[0280] Based on the same inventive concept, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the above-mentioned filtering method for multi-source trajectory data is implemented.

[0281] Based on the same inventive concept, an embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned multi-source trajectory data filtering method when executing the program.

[0282] The principles of the problems solved by the above-mentioned devices, media, and related equipment in the embodiments of the present invention are similar to those of the above-mentioned methods, so their implementation can refer to the implementation of the above-mentioned methods, and the repeated parts will not be repeated.

[0283] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0284] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0285] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0286] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0287] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A method for filtering multi-source trajectory data, characterized in that: include: constructing a tile route set of the tile based on a number of continuous trajectory segments within the tile; Based on the similarities between the tile routes in the tile route set, the tile routes in the tile route set are grouped to obtain at least one set of tile route subsets; Based on the acquisition accuracy of the tile routes and the similarity density of the tile routes, the tile routes in the tile route subset are filtered.

2. The method according to claim 1, characterized in that The filtering process on the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes includes: Determining whether the number of tile routes in the tile route subset is greater than a preset number threshold; If not, no filtering is performed; If yes, sorting the acquisition accuracy of the tile routes, and filtering the tile routes with low acquisition accuracy based on the sorting result; Determine whether the number of tile routes in the tile route subset after the primary filtering is greater than the preset number threshold; If so, a second filtering is performed on several tile routes with the lowest acquisition accuracy in the tile route subset after the first filtering based on the similarity density.

3. The method according to claim 2, wherein the second filtering of the tile routes with the lowest collection accuracy in the tile route subset after the first filtering based on the similarity density comprises: Determine the similarity density values ​​of several tile routes with the lowest acquisition accuracy respectively; The tile routes with the lowest similarity density value are filtered twice; The above determination steps are repeated until the number of tile routes after the primary filtering and / or secondary filtering is less than the preset number threshold.

4. The method according to claim 3, characterized in that The acquisition accuracy of the tile route is predetermined based on the following method: Based on the quality of the acquisition accuracy of each trajectory point in the trajectory segment corresponding to the tile route, the number of trajectory points with good acquisition accuracy in the trajectory segment is counted; The ratio of the number of trajectory points with good acquisition accuracy to the number of all trajectory points in the trajectory segment is used as the acquisition accuracy of the tile route.

5. The method according to claim 4, characterized in that Before the secondary filtering of several tile routes with the lowest acquisition accuracy, it also includes: Neighbor tile routes are determined based on the projected distances between the tile routes, so as to filter the tile routes based on the number of neighbor tile routes.

6. The method according to claim 1, characterized in that After filtering the tile routes in the tile route subset, the method further includes: Based on the coverage of the tile routes in the tile route subset, the tile routes with deletion marks in the tile route subset are detected to update the filtering result.

7. The method according to claim 4, characterized in that The detecting, based on the coverage of the tile routes in the tile route subset, the tile routes with deletion marks in the tile route subset to update the filtering result includes: Splitting the tile to obtain a plurality of high-level tiles; Traversing the tile route subset included in the tile, and determining whether there is a tile route with a deletion mark in the tile route subset; If it exists, then count the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes; Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes; Determine the size of the coverage ratio and a preset coverage ratio threshold; If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored; If it does not exist, the filtering results will not be updated.

8. The method according to claim 7, characterized in that The step of restoring the tile route with the deletion mark includes: Restore the tile routes with the deletion mark and the ones with the highest increase in coverage ratio.

9. The method according to claim 6, characterized in that The detecting, based on the coverage of the tile routes in the tile route subset, the tile routes with deletion marks in the tile route subset to update the filtering result includes: Determine whether there is a tile route with a deletion mark in the tile route subset; If yes, the tile is segmented to obtain a number of high-level tiles; Counting the number of tile routes with a non-deletion mark passing through the high-level tile and the number of all tile routes in the tile route subset passing through the high-level tile, so as to respectively determine the coverage rate of the tile routes with a non-deletion mark and the coverage rate of all tile routes; Determine the coverage ratio based on the coverage ratio of the tile routes with the non-deletion mark and the coverage ratio of all tile routes; Determine the size of the coverage ratio and a preset coverage ratio threshold; If the coverage ratio is less than a preset coverage ratio threshold, the tile route with the deletion mark is restored.

10. The method according to claim 1, characterized in that The step of constructing a tile route set of the tile based on the map data collected based on a plurality of continuous track segments in the tile comprises: Based on the vehicle ID and the timestamp of the track points, the track points of the tile are divided to determine the continuous track segments corresponding to the current vehicle ID; The continuous trajectory segments are regarded as a tile route, and a tile route set of the tile is constructed.

11. The method according to claim 1, characterized in that The step of grouping the tile routes in the tile set based on the similarities between the tile routes in the tile route set to obtain at least one set of tile route subsets includes: Taking the maximum value of the starting point distance and / or the ending point distance between two tile routes in the tile route set as the similarity between the two tile routes; The maximum value of the starting point height and / or the ending point height between the two tile routes is used as the height difference between the two tile routes; The tile routes in the tile route set whose similarity is less than a preset similarity threshold and whose height difference is less than a preset height difference threshold are taken as the tile route subset corresponding to the same actual road.

12. The method according to claim 11, characterized in that After the tile routes in the tile set are grouped, the method further includes: Based on the projection distance, projection length and projection angle of the starting point and / or end point of the tile route included in the tile route subset, it is determined whether the tile route subsets are aggregated into the same tile route subset.

13. The method according to any one of claims 1 to 11, characterized in that: Also includes: The filtering results are adjusted based on the topological relationship of tile routes between adjacent tiles.

14. The method according to claim 13, characterized in that The step of adjusting the filtering result based on the topological relationship of tile routes between adjacent tiles includes: If it is determined that the filtering result of the tile route in the current tile is marked with a delete mark, and the filtering result of the tile route in the adjacent tile with a topological relationship is marked with a non-delete mark, the filtering result of the tile route in the current tile is adjusted to be marked with a non-delete mark; or, Based on the topological relationship between tile routes between adjacent tiles, the tile routes in adjacent tiles are filtered using the filtering result of the tile routes in the current tile.

15. A filtering device for multi-source trajectory data, characterized in that: include: A construction module, configured to construct a tile route set of the tile based on a plurality of continuous trajectory segments within the tile; A grouping module, configured to group the tile routes in the tile set based on similarities between the tile routes in the tile route set to obtain at least one set of tile route subsets; The filtering module is used to filter the tile routes in the tile route subset based on the acquisition accuracy of the tile routes and the similarity density of the tile routes.

16. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for filtering multi-source trajectory data according to any one of claims 1 to 14 is implemented.

17. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for filtering multi-source trajectory data according to any one of claims 1 to 14 is implemented.

Images