Road generation method, device, equipment and storage medium

Abstract

Embodiments of the present application provide a road generation method, device, equipment and storage medium, the method comprises: obtaining at least one first trajectory segment which does not exist in the road network map in the original trajectory data, calculating the similarity degree between at least one first trajectory segment by using a preset distance index, clustering at least one first trajectory segment based on the similarity degree to obtain a segment clustering cluster, obtaining the segment vector of the second trajectory segment in the segment clustering cluster, and determining a new road in the road network map based on the segment vector. By clustering the first trajectory segment in the original trajectory data which cannot be matched with the road network map, and mining the new road based on the clustered cluster, the mining of unknown roads is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of map, in particular to a road generation method and device, equipment and storage medium. BACKGROUND

[0002] With the development and construction of the city, municipal road planning and construction has become a basic demand for urban development. Under such a background, more and more roads will be expanded or newly built. Therefore, it is necessary to effectively mine these new roads and update them to the map in time to provide a more accurate result for the later path planning, road navigation, etc.

[0003] The existing map updating scheme usually generates new roads by remote sensing satellite map changes, or collects data on site through a collection vehicle and then analyzes and processes the collected data in the data center. The road generation efficiency is low, therefore, it is urgent to propose an efficient road generation method. SUMMARY

[0004] The main purpose of the present application is to provide a road generation method, device, equipment and storage medium, which aims to mine unknown roads and improve the efficiency and effect of generating new roads. The technical solution is as follows:

[0005] In the first aspect, the present application provides a road generation method, comprising:

[0006] Obtaining at least one first trajectory segment that does not exist in a road network map from original trajectory data;

[0007] Using a preset distance index to calculate the similarity degree between the at least one first trajectory segment;

[0008] Based on the similarity degree, the at least one first trajectory segment is clustered to obtain a segment cluster;

[0009] Obtaining a segment vector of a second trajectory segment in the segment cluster, and determining a new road in the road network map based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment cluster.

[0010] In the second aspect, the present application provides a road generation device, comprising:

[0011] An acquisition module is configured to obtain at least one first trajectory segment that does not exist in a road network map from original trajectory data;

[0012] A similarity degree confirmation module is configured to use a preset distance index to calculate the similarity degree between the at least one first trajectory segment;

[0013] a clustering module configured to cluster the at least one first trajectory segment based on the similarity degree, to obtain a segment clustering cluster;

[0014] a new road generation module configured to obtain a segment vector of a second trajectory segment in the segment clustering cluster, and determine a new road in the road network map based on the segment vector.

[0015] In a third aspect, an electronic device is provided, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, the steps of the above method are implemented.

[0016] In a fourth aspect, a storage medium is provided, which stores a computer program. When the computer program is executed by a processor, the steps of the above method are implemented.

[0017] In the embodiments of the present application, at least one first trajectory segment not present in the road network map is obtained from the original trajectory data, a preset distance index is used to calculate the similarity degree between the at least one first trajectory segment, the at least one first trajectory segment is clustered based on the similarity degree, to obtain a segment clustering cluster, a segment vector of a second trajectory segment in the segment clustering cluster is obtained, and a new road is determined in the road network map based on the segment vector. By clustering the first trajectory segment that cannot be matched with the road network map in the original trajectory data, and mining a new road based on the clustered clustering cluster, the mining of unknown roads is realized, and the mining based on the original trajectory data is more convenient. By clustering processing based on the preset distance index, the first trajectory segments with similar positions are taken as a category, and a new road is obtained based on the second trajectory segments of the same category, thereby avoiding data redundancy. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0019] Figure 1 is an example schematic diagram of a road generation method provided by the embodiments of the present application;

[0020] Figure 2 is a flowchart of a road generation method provided by the embodiments of the present application;

[0021] Figure 3This is a schematic diagram illustrating the definition of the distance index in a road generation method provided in an embodiment of this application;

[0022] Figure 4 This is a detailed flowchart illustrating a road generation method provided in an embodiment of this application;

[0023] Figure 5 This is a further detailed flowchart of a road generation method provided in an embodiment of this application;

[0024] Figure 6 This is a schematic diagram of fragment clustering of a road generation method provided in an embodiment of this application;

[0025] Figure 7 This is a schematic diagram of the structure of a road generation device provided in an embodiment of this application;

[0026] Figure 8 This is a schematic diagram of the structure of a road generation device provided in an embodiment of this application. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0028] The road generation device can be a terminal device such as a mobile phone, computer, tablet, smartwatch, or vehicle-mounted device, or it can be a module in the terminal device used to implement the road generation method. The road generation device can obtain at least one first trajectory segment that does not exist in the road network map from the original trajectory data, calculate the similarity between at least one first trajectory segment using a preset distance index, perform clustering processing on at least one first trajectory segment based on the similarity to obtain segment clusters, obtain the segment vector of the second trajectory segment within the segment cluster, and determine the new road in the road network map based on the segment vector.

[0029] Please see also Figure 1 This is a schematic diagram illustrating an example of a road generation method provided in this application. The road generation device is based on the collected original trajectory data ( Figure 1 The black line in the left middle is used for road network map matching to obtain at least one first trajectory segment that does not exist in the road network map. Then, the first trajectory segments are clustered using a preset distance index to obtain segment clusters. Based on the segment vectors of each second trajectory segment in the segment cluster, new roads are generated for the segment clusters. Figure 1 right).

[0030] The road generation method provided in the specification will be described in detail below in conjunction with specific embodiments.

[0031] Please refer to Figure 2 A flowchart of a road generation method is provided for the embodiments of the present application. As shown in Figure 2 The method of the embodiments of the present application can include the following steps S101-S104.

[0032] S101, obtaining at least one first trajectory segment not existing in a road network map from original trajectory data;

[0033] S102, calculating the similarity degree between the at least one first trajectory segment by using a preset distance index;

[0034] S103, performing clustering processing on the at least one first trajectory segment based on the similarity degree to obtain a segment clustering cluster;

[0035] S104, obtaining a segment vector of a second trajectory segment in the segment clustering cluster, and determining a new road in the road network map based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment clustering cluster.

[0036] Since the existing map updating scheme is usually to generate a new road by remote sensing satellite map whether there is a change, or to collect data on site by a collection vehicle and then to bring the collected data back to a data center for analysis and processing by the collection vehicle, the road generation efficiency is low. Therefore, in the embodiments of the present application, the new road is mined based on the original trajectory data. It can be understood that the take-out rider as a "small cell phone" of the city often rides in the streets and alleys of the city, and their trajectory data reflects their path in the delivery process. With the application of the rider delivery full-link trajectory data collection, the rider trajectory data is growing explosively. Therefore, one problem to be solved in the embodiments is how to mine potential unknown roads from a large number of trajectory paths. Therefore, the embodiments propose a road generation method based on the collected original trajectory data to mine new roads.

[0037] The various steps will be described in detail below:

[0038] S101, obtaining at least one first trajectory segment not existing in a road network map from original trajectory data;

[0039] The original trajectory data can be a driving trajectory of a two-wheeled vehicle or a driving trajectory of a four-wheeled vehicle, or a combined trajectory of the two. It can be understood that, since some roads can be passed by two-wheeled vehicles but cannot be passed by four-wheeled vehicles, the original trajectory data required can be determined according to the purpose of the subsequent updated map, for example, the trajectory data of a two-wheeled vehicle can be collected as the original trajectory data when the updated map is used for real-time delivery path planning in a two-wheeled vehicle scenario. Specifically, the trajectory data can be collected by a positioning device installed on a rider and a vehicle, or can be collected by an application installed on a mobile terminal used by the rider.

[0040] Specifically, the original trajectory data can be matched to the road network map based on a road network matching technology. Since the original trajectory data collected is a plurality of point sequences, for example, the signal frequency is collected about once every 3-5 seconds, and a position coordinate is obtained each time. The purpose of map matching is to match the position coordinate to the road network, to confirm which road the trajectory point is on, that is, which road the rider is on, and correspondingly, if the trajectory point cannot be matched to a certain road in the road network, it means that the road can be a newly discovered road by the rider, and a new road can be generated based on the first trajectory point that is not matched.

[0041] The map road network can be a provincial road network, a municipal road network, or a partial regional road network, and the specific requirement can be determined according to the original trajectory data collected, that is, how much regional road network information is required.

[0042] S102, calculating the similarity between the at least one first trajectory segment based on a preset distance index;

[0043] Specifically, after obtaining the first trajectory segment that is not matched to the road network map, the at least one first trajectory segment is subjected to clustering processing. The process of dividing a collection of physical or abstract objects into a plurality of classes composed of similar objects is called clustering. The cluster generated by clustering is a collection of data objects, which are similar to each other in the same cluster and different from the objects in other clusters.

[0044] Before dividing the trajectory segment data into different clusters, an important problem that needs to be solved is the distance measurement of the trajectory, that is, comparing the similarity of the trajectory segments. The similarity between the first trajectory segments is calculated by a preset distance index, and clustering is performed according to the similarity. Specifically, there are many methods for measuring the similarity between trajectory segments. Therefore, the similarity comparison of the trajectory should follow a special strategy, not only to compare the differences between the trajectories comprehensively, but also to select different comparison strategies according to the purpose of clustering.

[0045] In one embodiment, a distance index compatible with the distance threshold between trajectories can be designed by referring to the method of the classic trajectory clustering algorithm Traclus. The index simultaneously fuses the angle distance, vertical distance and parallel distance three distance indexes, and uses the sum of the three measurements as the total measurement to calculate the similarity between two trajectories. The same category of first trajectory segments are finally aggregated together by DBSCAN algorithm (density-based clustering algorithm).

[0046] Specifically, referring to Figure 3 , Figure 3 is a definition diagram of distance index in a road generation method provided by the present application, L j and L j represent different first trajectory segments, s and e are the starting point and the ending point respectively, and the similarity between the two first trajectory segments can be obtained by calculating the vertical distance, the parallel distance and the included angle distance. The definitions of the three distances are as follows:

[0047] a. Vertical distance d ⊥ (L i , L j ): The vertical distance between L i and L j is defined as follows,

[0048]

[0049] If the projection points of the starting point s j and the ending point e j of the point segment L j on the line segment L i are p s and p e respectively, the distance between s j and p s is l ⊥1 , and the distance between e j and p e is l ⊥2 , then the two vertical distances are the Euclidean distances between s i and ei e j and p s .

[0050] b. Parallel distance d || (L i , L j ): The parallel distance between L i and L j is defined as follows,

[0051] d || (L i , L j ) = MIN(l||1 , l ||2 )

[0052] If the projection points of s j and e j on the line segment L i are p s and p e respectively, the distance between s i and p s is l ||1 , and the distance between e i and p e is l ||2 , then the distance between the two parallel lines is the minimum Euclidean distance between p s and the line segment s i , e i and the minimum Euclidean distance between p e and the line segment s i , e i .

[0053] c. The angle distance d θ (L i , L j ): The angle distance between L i and L j is defined as shown below, and θ is the smaller angle between L i and L j . When the angle is less than 90 degrees, it is represented by the norm of vector L j ||L j || multiplied by sin(θ), and when the angle is greater than 90 degrees, it is represented by the norm of vector L i ||L j ||.

[0054]

[0055] It should be noted that an assumption is made here that the angle deviation between similar trajectories will not be particularly large, so the angle distance calculation formula is further modified:

[0056]

[0057] By this operation, the punishment of the angle is enlarged, and finally the effect of differentiation is achieved.

[0058] Wherein, in order to facilitate calculation, all vector calculations are used:

[0059] The coordinates of points p s and p e can be expressed as:

[0060]

[0061] wherein:

[0062]

[0063] Finally, the three distances are added to obtain the final distance between the trajectory segments, as follows, to prepare for the next clustering. Wherein w ⊥ , w || , w θ are the corresponding penalty coefficients. The latter can be adjusted by itself.

[0064] dist(L i , L j ) = w ⊥ ·d ⊥ (L i , L j ) + w || ·d || (L i , L j ) + w θ ·d θ (L i , L j )

[0065] S103, performing clustering processing on the at least one first trajectory segment based on the similarity degree, to obtain a segment clustering cluster;

[0066] Specifically, after calculating the similarity degree based on the above distance index, the first trajectory segment is clustered. For example, a pre-set similarity degree threshold is used to determine which first trajectory segments can be classified into a category. Through clustering, the data volume can be reduced, and it is not necessary to generate multiple new roads for similar unmatched trajectory segments, which can reduce data redundancy to a certain extent.

[0067] S104, obtaining a segment vector of a second trajectory segment in the segment clustering cluster, and determining a new road in the road network map based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment clustering cluster.

[0068] Specifically, when the segment clustering cluster is obtained by clustering, at least one second trajectory segment is included in each segment clustering cluster, wherein the second trajectory segment is a first trajectory segment in the same clustering cluster. A segment vector corresponding to the second trajectory segment in the segment clustering cluster is obtained, and a first trajectory segment corresponding to the segment clustering cluster is generated based on the segment vector. The segment vector refers to the vector representation of the connection line from the starting point to the ending point of the trajectory. It can be understood that the second trajectory segments are similar trajectory segments, but they are usually not completely consistent, for example, in terms of direction, they can not be parallel, and can be intersected at an angle. Therefore, the direction of the new road needs to be determined based on the segment vectors of the second trajectory segments. Optionally, the mean value of the segment vectors can be taken, and the vector corresponding to the mean value is taken as the vector representation of the new road.

[0069] In the embodiment of the present application, at least one first trajectory segment not existing in the road network map is obtained from the original trajectory data, a preset distance index is used to calculate the similarity between the at least one first trajectory segment, and the at least one first trajectory segment is clustered based on the similarity to obtain a segment clustering cluster. The segment vector of the second trajectory segment in the segment clustering cluster is obtained, and the new road is determined in the road network map based on the segment vector. By collecting the first trajectory segment that cannot be matched with the road network map from the original trajectory data, based on the unknown road mining scene, the preset distance index is used to measure the similarity between the first trajectory segments, which can obtain better clustering effect, and then the segment vector of the second trajectory segment in the segment clustering cluster obtained after clustering is used to determine the new road, which can obtain more accurate road direction and improve the accuracy of the new road.

[0070] Please refer to Figure 4 A detailed flowchart of a road generation method provided by the embodiment of the present application. As Figure 4 shown, the method of the embodiment of the present application can include the following steps S201-S211.

[0071] S201, obtaining a matching road in a preset matching range of a first trajectory point in the original trajectory data from the road network map; the first trajectory point is any trajectory point in the plurality of trajectory points collected in the original trajectory data;

[0072] Specifically, in the present embodiment, the Hidden Markov Model (HMM) is used when performing map matching. However, in the traditional HMM, in the step of calculating the observation probability and the transition probability to select the most likely candidate road, since the amount of candidate roads that can be obtained is very large, there is a problem of large amount of calculation. Therefore, first, the road network in the preset matching range is obtained, which can reduce the search area of the road network and reduce the search cost and the amount of calculation.

[0073] S202, obtain a trajectory vector based on a line connecting a first trajectory point and a second trajectory point in the original trajectory data, calculate an included angle between the trajectory vector and a road vector of the matching road, the second trajectory point being a trajectory point obtained in a chronological order before the first trajectory point according to a collection time of the original trajectory data;

[0074] Further, the azimuth angle pruning method is used when recalling the surrounding roads, and the azimuth angle is used to recall only the road vector expressions in the specified direction. For example, all roads in the opposite direction of the trajectory vector are filtered out. Using this method, at least 50% of the road vector expressions are recalled, while the accuracy is improved, thereby reducing the recall of some incorrect paths.

[0075] Specifically, when recalling the matching roads in a predetermined matching range around a certain trajectory point, a part of the roads different from the direction of the trajectory are filtered out according to the direction of the road (the vector of the line connecting the current trajectory point and the previous trajectory point, and the vector of the road in the road network, the included angle between the two vectors is calculated), thereby reducing the matching roads recalled by a trajectory point. For example, if the direction of the current trajectory point is north, and both roads are north, the included angle between the two vectors is 0.

[0076] S203, confirming the matching road with the included angle less than an included angle threshold value as a first initial candidate road corresponding to the first trajectory point;

[0077] Specifically, the recalled roads are filtered according to a predetermined included angle threshold value. If the included angle is greater than the included angle threshold value, it is considered that the road is different from the direction of the trajectory, the difference is too large, and the trajectory point is unlikely to match the road. If the included angle is less than the included angle threshold value, it is considered that the direction of the matching road is close to the direction of the trajectory, and the trajectory point is likely to match the matching road, and the matching road is taken as the first candidate road. The included angle threshold value can be selected according to the actual situation, for example, 45°.

[0078] S204, calculating a road surface distance between the first initial candidate road and the first trajectory point;

[0079] It can be understood that, since the roads in the road network all look like a line, but in fact the road has a width. If the distance between the trajectory point and the road is too large, the trajectory point is unlikely to be on the road. Therefore, the first initial candidate road is filtered by calculating the road surface distance between the first initial candidate road and the first trajectory point.

[0080] S205, eliminating the first initial candidate road with the road surface distance greater than a distance threshold value;

[0081] Specifically, the first candidate road with a road surface distance greater than a distance threshold is eliminated. The distance threshold needs to be determined according to the road width of the candidate road, and a certain error range is allowed. For example, if the road width is 10 m and the error range is 2 m, the distance threshold is 12 m, and the distance between the trajectory point and the road is 30 m, which is obviously not the road currently being traveled, so it is filtered out.

[0082] S206, calculating an observation probability between the first trajectory point and the first initial candidate road;

[0083] S207, calculating a transition probability from the first initial candidate road to a second initial candidate road corresponding to the second trajectory point;

[0084] In the HMM-based map matching algorithm, there is one or more candidate road segments within a certain distance for a trajectory point, and the projection point of the trajectory point on the candidate road segment is regarded as a vertex in the Markov chain. The closer the trajectory point is to the position on the adjacent road segment, the greater the probability of the point on the road segment. The closer the distance between the two real position points, the greater the state transition probability, or the closer the distance between the two points on the real road segment to the distance between the two points observed by the trajectory point, the greater the state transition rate. Therefore, the observation probability between the first trajectory point and the first initial candidate road is calculated, and the transition probability from the first initial candidate road to the second initial candidate road is calculated. The specific calculation method can refer to the existing HMM algorithm, which is not described here.

[0085] S208, confirming whether there is a target candidate road matching the first trajectory point in the first initial candidate road based on the observation probability and the transition probability;

[0086] S209, if not, generating at least one first trajectory segment not existing in the road network map based on the first trajectory point.

[0087] Based on the calculated observation probability and transition probability, it is confirmed whether there is a target candidate road matching the first trajectory point in the first initial candidate road. Generally, the first initial candidate road with the largest sum of observation probability and transition probability is confirmed as the target candidate road. If the observation probability of the first trajectory point to the first initial candidate road, or the transition probability from the first initial candidate road to the second initial candidate road is very small, the matching target candidate road cannot be found, and it can be considered that the trajectory point cannot be matched with the existing road on the road network map. Then, based on the obtained first trajectory points that cannot be matched, the first trajectory segment is generated.

[0088] Further, in an embodiment, after the observation probability between the first trajectory point and the first initial candidate road is calculated, the method further comprises:

[0089] S210, if the maximum value of the observation probability is less than a probability threshold, calculating a trajectory point distance between the first trajectory point and the second trajectory point;

[0090] S211, if the trajectory point distance is greater than a detour threshold, generating a non-passable trajectory segment based on a line connecting the first trajectory point and the second trajectory point, and taking the non-passable trajectory segment as the at least one first trajectory segment that does not exist in the road network map.

[0091] In an embodiment, if the maximum probability of the trajectory point projected onto each first initial candidate road, i.e., the maximum value of the observation probability, is still small, it may be that there is a pass-through behavior, and it is necessary to determine whether a detour can be made. If a small amount of detour (detour threshold) can be made, it is considered that the matching is correct, and if it cannot be solved by detour, it is considered that there is no road matching, and the trajectory may be a missing road information, i.e., there is no first trajectory segment in the road network map. In some scenarios, such as instant delivery scenarios, the rider may cross the road, such as a two-way lane, and it is not possible for a motor vehicle to pass through if it needs to find a suitable U-turn location to U-turn to the other side of the road. However, in a cycling scenario, it may be possible to cross, and in this case, it is not actually because of a new road, but because of a pass-through situation.

[0092] Specifically, by calculating the trajectory point distance between the first trajectory point and the second trajectory point, if the trajectory point distance is greater than the detour threshold, it means that a detour cannot be made, i.e., a pass-through cannot be made, and it is determined as a non-passable trajectory segment, and the non-passable trajectory segment is taken as the first trajectory segment that does not exist in the road network map.

[0093] In this embodiment, map matching is performed by using a hidden Markov algorithm to obtain a first trajectory segment that is not matched. Compared with directly superimposing trajectory data on a road network topology and determining whether it is a new road according to whether it coincides with the road network topology, a more accurate new road can be obtained. Furthermore, the azimuth angle pruning and candidate road filtering methods are used to improve the performance of the map matching algorithm based on the hidden Markov model, and compared with the original algorithm, the efficiency and effectiveness are obviously improved. At the same time, the pass-through situation is considered, and the accuracy of trajectory point matching is further improved.

[0094] Please refer to Figure 5 , another detailed flowchart of a road generation method provided by the embodiment of the present application. As shown in Figure 5As shown, the method of the embodiment of the present application can include steps S301-S302.

[0095] S301, obtaining the midpoint longitude and latitude coordinates of the target trajectory segment in the first trajectory segment, and confirming the grid where the target trajectory segment is located based on the midpoint longitude and latitude point coordinates;

[0096] Specifically, the target trajectory segment is the first trajectory segment currently performing similarity degree calculation. Based on the coordinates of the starting point and the ending point of the target trajectory segment, the midpoint coordinates of the target trajectory segment are calculated. Specifically, the coordinates are expressed by longitude and latitude, and are kept to the last three digits of longitude and latitude, i.e. accurate to 100 meters. And the last three digits of the longitude and latitude of the midpoint are taken as a string index, and a grid is defined based on the range of 100 meters around the midpoint.

[0097] S302, based on the distance index, calculating the similarity degree between the target trajectory segment and the grid where it is located and all first trajectory segments in the neighborhood grid of the grid where it is located.

[0098] It can be understood that in the clustering process, since the original method in the trajectory clustering algorithm adopts the DBSCAN algorithm, the calculation efficiency is low when calculating globally, and global traversal is required every time the similarity degree is calculated to determine the category, which greatly reduces the performance of the algorithm. Therefore, a grid division method based on longitude and latitude index is adopted to optimize the performance of the clustering process, that is, the last three digits of longitude and latitude are taken as a string index for each trajectory segment, and when similar trajectory segments are searched, only all trajectory segments in the grid where the target trajectory segment is located and the eight neighborhood grids around it are calculated, which simplifies the global calculation to local calculation and greatly improves the performance. The method of "convolution" expansion clustering avoids invalid and redundant calculations, and by doing so, we overcome the challenge of high density of trajectory data.

[0099] Further, in an embodiment, the method comprises the following steps:

[0100] S303, confirming the minimum circumscribed rectangle of the segment clustering cluster;

[0101] S304, obtaining the average vector direction of the segment vector of the second trajectory segment, and confirming the center line of the minimum circumscribed rectangle consistent with the average vector direction, taking the center line as the corresponding new road of the segment clustering cluster in the road network map.

[0102] Specifically, after obtaining the corresponding segment cluster of the first trajectory segment, a minimum circumscribed rectangle of the segment cluster trajectory is calculated, and a center line consistent with the average vector direction of the second trajectory segment is taken as a new road vector representation. As shown in Figure 6 FIG. 6a is a segment cluster obtained by clustering a plurality of second trajectory segments, 6b is a minimum circumscribed rectangle of the segment cluster, and 6c is a center line of the minimum circumscribed rectangle of the segment cluster, which is used to determine a new road.

[0103] Further, in an embodiment, before obtaining the at least one first trajectory segment that does not exist in the road network map from the original trajectory data, the method further comprises:

[0104] S305, collecting position information of the trajectory points and sequentially saving the position information into a preset queue according to the time sequence of collection;

[0105] Specifically, the position information of the trajectory points is collected at a preset time, and the trajectory points are saved into a preset queue according to the time sequence of collection,

[0106] S306, calculating the distance proximity between a third trajectory point and a fourth trajectory point in the preset queue, the third trajectory point being any collected trajectory point in the preset queue, and the fourth trajectory point being two adjacent trajectory points before and after the third trajectory point in the preset queue;

[0107] S307, determining whether the third trajectory point is a drift trajectory point based on the distance proximity;

[0108] S308, if yes, filtering the third trajectory point from the preset queue to generate original trajectory data.

[0109] Drift is a situation that due to device abnormalities, the position of the GPS point has a sudden large deviation under continuous timestamps. Drift points have a great impact on subsequent analysis, so in this embodiment, drift points are obtained and removed.

[0110] Specifically, the distance proximity between the current trajectory point and the two adjacent trajectory points, i.e., the distance ratio, is used to determine whether the point is drift. This ratio is also used to define the characteristics of good and bad points. It can be understood that since the interval of collection time is fixed, if the distance difference is too large or too small within the same time interval, it means that the trajectory point may be a drift point. For example, three points A, B and C are collected in time sequence, and if BC / AB is approximately 1, it means that point B is normal, and if BC / AB is much greater than 1 or much less than 1, it is determined that the point is a drift point, and point B is removed from the preset queue.

[0111] Further, in an embodiment, the drift trajectory point is obtained based on the distance proximity, the drift trajectory point is filtered from the preset queue, and original trajectory data is generated, including:

[0112] S308, a drift trajectory point is obtained based on the distance proximity, the drift trajectory point is filtered from the preset queue, and an initial trajectory point sequence is generated;

[0113] S309, a stay point in the initial trajectory point sequence is identified, the stay point is filtered from the initial trajectory point sequence, and original trajectory data is generated.

[0114] After filtering the drift point, the stay point in the initial trajectory point sequence is further identified, and the stay point is filtered. The stay point mainly includes two types: a static stay point and a wandering stay point. The stay point is identified from the trajectory data, and the unimportant and redundant information in the trajectory data can be effectively removed. The algorithm for identifying the stay point is not limited, and the stay point detection algorithm based on density or the stay point detection algorithm based on speed can be used.

[0115] In addition, after filtering the stay point from the initial trajectory point sequence, a simple thinning operation can also be performed. The coordinate data with less influence on the route can be deleted under the condition of meeting the original route, so as to save the data processing time.

[0116] In the embodiment of the present application, by collecting the position information of the trajectory point and pre-processing the trajectory point in the preset queue collected, the redundant part and the part not concerned in the data are removed through drift point filtering and stay point filtering, so as to improve the data quality of the original trajectory data. The longitude and latitude point coordinates of the second trajectory segment are used as indexes to obtain similar segments, so as to reduce the calculation amount of the clustering process and improve the performance of the clustering algorithm. The average vector direction of the segment vector in the segment clustering cluster is obtained, so as to confirm the center line of the minimum circumscribed rectangle of the segment clustering cluster, the center line represents the unmatched trajectory in the segment clustering cluster, and the new road is generated based on the center line, so as to improve the accuracy of the position of the new road.

[0117] The embodiments of the present application will be described below with reference to the accompanying drawings. Figure 7 The road generation device provided by the embodiments of the present application will be described in detail. It should be noted that the road generation device in the embodiments of the present application is used to execute the method of the embodiments of the present application shown in the description. Figure 7 The road generation device in the embodiments of the present application is used to execute the method of the embodiments of the present application shown in the description. Figures 2-6 For the convenience of description, only the parts related to the embodiments of the present application are shown, and the specific technical details not disclosed are described with reference to the embodiments shown in the description. Figures 2-6 For the convenience of description, only the parts related to the embodiments of the present application are shown, and the specific technical details not disclosed are described with reference to the embodiments shown in the description.

[0118] Please refer to Figure 7Fig. 1 is a structural schematic diagram of a road generation device according to an example embodiment of the present application. The road generation device can be realized by software, hardware or a combination of both to become all or part of the device. The device 1 comprises an acquisition module 10, a similarity confirmation module 20, a clustering module 30 and a new road generation module 40.

[0119] The acquisition module 10 is configured to acquire at least one first trajectory segment not existing in a road network map from original trajectory data.

[0120] The similarity confirmation module 20 is configured to calculate the similarity between the at least one first trajectory segment by using a preset distance index.

[0121] The clustering module 30 is configured to perform clustering processing on the at least one first trajectory segment based on the similarity to obtain a segment clustering cluster.

[0122] The new road generation module 40 is configured to acquire a segment vector of a second trajectory segment in the segment clustering cluster and determine a new road in the road network map based on the segment vector.

[0123] Optionally, the acquisition module 10 is specifically configured to acquire a matching road in a preset matching range of a first trajectory point in the original trajectory data from the road network map, wherein the first trajectory point is any trajectory point in a plurality of trajectory points collected in the original trajectory data.

[0124] A trajectory vector is acquired based on a line connecting the first trajectory point and a second trajectory point in the original trajectory data, an included angle between the trajectory vector and a road vector of the matching road is calculated, and the second trajectory point is a trajectory point obtained before the first trajectory point in a chronological order of collection time of the original trajectory data.

[0125] The matching road with the included angle less than an included angle threshold is confirmed as a first initial candidate road corresponding to the first trajectory point.

[0126] An observation probability between the first trajectory point and the first initial candidate road is calculated.

[0127] A transition probability from the first initial candidate road to a second initial candidate road is calculated, and the second initial candidate road is an initial candidate road corresponding to the second trajectory point.

[0128] Whether a target candidate road matching the first trajectory point exists in the first initial candidate road is confirmed based on the observation probability and the transition probability.

[0129] If not, at least one first trajectory segment not existing in the road network map is generated based on the first trajectory point.

[0130] Optionally, the acquisition module 10 is specifically configured to calculate a road surface distance between the first initial candidate road and the first trajectory point;

[0131] The first initial candidate road with the road surface distance greater than a distance threshold is removed.

[0132] Optionally, the acquisition module 10 is specifically configured to calculate a trajectory point distance between the first trajectory point and the second trajectory point if a maximum value of the observation probability is less than a probability threshold;

[0133] If the trajectory point distance is greater than a complement road threshold, a non-penetration trajectory segment is generated based on a line segment between the first trajectory point and the second trajectory point, and the non-penetration trajectory segment is taken as at least one first trajectory segment that does not exist in the road network map.

[0134] Optionally, the similarity degree confirmation module 20 is specifically configured to acquire a midpoint longitude and latitude coordinate of a target trajectory segment in the first trajectory segment, and confirm a grid where the target trajectory segment is located based on the midpoint longitude and latitude coordinate;

[0135] Based on the distance index, a similarity degree between the target trajectory segment and all first trajectory segments in the grid where the target trajectory segment is located and a neighborhood grid of the grid where the target trajectory segment is located is calculated.

[0136] Optionally, the new road generation module 40 is specifically configured to confirm a minimum circumscribed rectangle of the segment cluster;

[0137] An average vector direction of a segment vector of the second trajectory segment is acquired, and a center line of the minimum circumscribed rectangle consistent with the average vector direction is confirmed, and the center line is taken as a new road corresponding to the segment cluster in the road network map.

[0138] Optionally, the acquisition module 10 is further configured to acquire position information of a trajectory point, and sequentially save the position information into a preset queue according to a time sequence of acquisition;

[0139] A distance proximity degree between a third trajectory point and a fourth trajectory point in the preset queue is calculated, the third trajectory point is any acquired trajectory point in the preset queue, and the fourth trajectory point is two trajectory points adjacent to the third trajectory point in the preset queue;

[0140] Based on the distance proximity degree, whether the third trajectory point is a drift trajectory point is confirmed;

[0141] If yes, the third trajectory point is filtered from the preset queue, and original trajectory data is generated.

[0142] Optionally, the acquisition module 10 is further configured to obtain a drift trajectory point based on the distance closeness, filter the drift trajectory point from the preset queue, and generate an initial trajectory point sequence.

[0143] identify a stay point in the initial trajectory point sequence, filter the stay point from the initial trajectory point sequence, and generate original trajectory data.

[0144] It should be noted that the road generation device provided in the above embodiments is used to execute the road generation method, and the above-mentioned division of each functional module is only used as an example. In actual application, the above-mentioned functions can be completed by different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the above-described functions. In addition, the road generation device and the road generation method provided in the above embodiments belong to the same concept, and the implementation process is described in detail in the method embodiments, which will not be repeated here.

[0145] The above-mentioned serial numbers of the embodiments of the present application are only for description, and do not represent the advantages or disadvantages of the embodiments. In some cases, the actions or steps recited in the claims can be executed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0146] The embodiments of the present application also provide a storage medium, which stores a computer program. The computer program is executed by a processor to implement the road generation method of the above-mentioned embodiments. The specific implementation process can be referred to the specific description of the above-mentioned embodiments, which will not be repeated here. Figures 2-6 The embodiments of the present application also provide a storage medium, which stores a computer program. The computer program is executed by a processor to implement the road generation method of the above-mentioned embodiments. The specific implementation process can be referred to the specific description of the above-mentioned embodiments, which will not be repeated here. Figures 2-6

[0147] Please refer to Figure 8 , which shows the structural schematic diagram of the road generation device provided by an exemplary embodiment of the present application. The road generation device in the present application can include one or more of the following components: a processor 110, a memory 120, an input device 130, an output device 140 and a bus 150. The processor 110, the memory 120, the input device 130 and the output device 140 can be connected through the bus 150.

[0148] ​The processor 110 can include one or more processing cores. The processor 110 connects various parts within the entire road generation device by various interfaces and lines, performs various functions of the terminal 100 and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 120, and calling data stored in the memory 120. Alternatively, the processor 110 can be implemented in at least one of a hardware form of a digital signal processing (DSP), a field-programmable gate array (FPGA), a programmable logic array (PLA). The processor 110 can integrate one or a combination of a central processing unit (CPU), a graphics processor (GPU), and a modem. Among them, the CPU mainly processes an operating system, a user page, and an application program, etc.; the GPU is responsible for rendering and drawing display content; and the modem is used for processing wireless communication. It can be understood that the above-mentioned modem can also not be integrated into the processor 110, but can be implemented by a separate communication chip.

[0149] The memory 120 can include a random access memory (RAM) and can also include a read-only memory (ROM). Alternatively, the memory 120 includes a non-transitory computer-readable storage medium. The memory 120 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 120 can include a program storage area and a data storage area, wherein the program storage area can store instructions for implementing an operating system, instructions for implementing at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-mentioned various method embodiments, etc., and the operating system can be an Android system, an IOS system developed by Apple Inc., a system developed based on the Android system or the IOS system, or other systems.

[0150] The memory 120 can be divided into operating system space and user space. The operating system runs in the operating system space, while native and third-party applications run in user space. To ensure that different third-party applications can achieve good running performance, the operating system allocates corresponding system resources for each application. However, different application scenarios within the same third-party application have different requirements for system resources. For example, in local resource loading scenarios, third-party applications have high requirements for disk read speed; in animation rendering scenarios, third-party applications have high requirements for GPU performance. Since the operating system and third-party applications are independent of each other, the operating system often cannot promptly perceive the current application scenario of a third-party application, resulting in the operating system's inability to adapt system resources accordingly.

[0151] In order for the operating system to distinguish the specific application scenarios of third-party applications, it is necessary to establish data communication between the third-party applications and the operating system. This would allow the operating system to obtain the current scenario information of the third-party applications at any time, and then perform targeted system resource adaptation based on the current scenario.

[0152] The input device 130 is used to receive input instructions or data, and includes, but is not limited to, a keyboard, mouse, camera, microphone, or touch device. The output device 140 is used to output instructions or data, and includes, but is not limited to, a display device and a speaker. In one example, the input device 130 and the output device 140 can be combined, and the input device 130 and the output device 140 can be a touch display screen.

[0153] The touch display screen can be designed as a full-screen, curved screen, or irregularly shaped screen. It can also be designed as a combination of a full-screen and a curved screen, or a combination of an irregularly shaped screen and a curved screen; however, this application does not limit the specific design in this regard.

[0154] In addition, those skilled in the art will understand that the structure of the road generation device shown in the above figures does not constitute a limitation on the road generation device. The road generation device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the road generation device may also include radio frequency circuits, input units, sensors, audio circuits, WiFi modules, power supplies, Bluetooth modules, etc., which will not be described in detail here.

[0155] exist Figure 8 In the road generation device shown, the processor 110 can be used to call the road generation application stored in the memory 120 and specifically perform the following operations:

[0156] acquire at least one first trajectory segment not existing in the road network map from the original trajectory data;

[0157] adopt a preset distance index to calculate a similarity degree between the at least one first trajectory segment;

[0158] perform clustering processing on the at least one first trajectory segment based on the similarity degree to obtain a segment clustering cluster;

[0159] acquire a segment vector of a second trajectory segment in the segment clustering cluster, and determine a newly added road in the road network map based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment clustering cluster.

[0160] In one embodiment, the processor 110, when performing the operation of acquiring at least one first trajectory segment not existing in the road network map from the original trajectory data, specifically performs the following operations:

[0161] acquire a matching road within a preset matching range of a first trajectory point in the original trajectory data from the road network map; the first trajectory point being any trajectory point in a plurality of trajectory points collected in the original trajectory data;

[0162] acquire a trajectory vector based on a connection line of the first trajectory point and a second trajectory point in the original trajectory data, calculate an included angle between the trajectory vector and a road vector of the matching road, the second trajectory point being a trajectory point obtained in a chronological order before the first trajectory point;

[0163] confirm the matching road with the included angle smaller than an included angle threshold as a first initial candidate road corresponding to the first trajectory point;

[0164] calculate an observation probability between the first trajectory point and the first initial candidate road;

[0165] calculate a transition probability from the first initial candidate road to a second initial candidate road, the second initial candidate road being an initial candidate road corresponding to the second trajectory point;

[0166] confirm whether there is a target candidate road matching the first trajectory point in the first initial candidate road based on the observation probability and the transition probability;

[0167] if not, generate at least one first trajectory segment not existing in the road network map based on the first trajectory point.

[0168] In one embodiment, the processor 110, after performing the operation of confirming the matching road with the included angle smaller than the included angle threshold as the first initial candidate road corresponding to the first trajectory point, further performs the following operations:

[0169] calculating a road surface distance between the first initial candidate road and the first trajectory point;

[0170] eliminating the first initial candidate road whose road surface distance is greater than a distance threshold.

[0171] In one embodiment, the processor 110, after performing the calculation of the observation probability between the first trajectory point and the first initial candidate road, further performs the following operations:

[0172] if the maximum value of the observation probability is less than a probability threshold, calculating a trajectory point distance between the first trajectory point and the second trajectory point;

[0173] if the trajectory point distance is greater than a complement road threshold, generating a non-passable trajectory segment based on a line connecting the first trajectory point and the second trajectory point, and taking the non-passable trajectory segment as at least one first trajectory segment that does not exist in the road network map.

[0174] In one embodiment, the processor 110, when performing the calculation of the similarity between the at least one first trajectory segment by using a preset distance indicator, specifically performs the following operations:

[0175] obtaining the midpoint longitude and latitude coordinates of a target trajectory segment in the first trajectory segment, and confirming the grid where the target trajectory segment is located based on the midpoint longitude and latitude coordinates;

[0176] calculating the similarity between the target trajectory segment and the grid where the target trajectory segment is located and all first trajectory segments in the neighbor grid of the grid based on the distance indicator.

[0177] In one embodiment, the processor 110, when obtaining the segment vector of the second trajectory segment in the segment cluster, and determining the newly added road in the road network map based on the segment vector, specifically performs the following operations:

[0178] confirming the minimum circumscribed rectangle of the segment cluster;

[0179] obtaining the average vector direction of the segment vector of the second trajectory segment, and confirming the center line of the minimum circumscribed rectangle consistent with the average vector direction, taking the center line as the corresponding newly added road of the segment cluster in the road network map.

[0180] In one embodiment, the processor 110, before performing the obtaining of the at least one first trajectory segment that does not exist in the road network map in the original trajectory data, further performs the following operations:

[0181] Collect position information of the trajectory points, and save the position information in a preset queue in sequence according to the time sequence of collection;

[0182] Calculate a distance proximity of a third trajectory point and a fourth trajectory point in the preset queue, the third trajectory point being any collected trajectory point in the preset queue, and the fourth trajectory point being two adjacent trajectory points before and after the third trajectory point in the preset queue;

[0183] Confirm whether the third trajectory point is a drift trajectory point based on the distance proximity;

[0184] If yes, filter the third trajectory point from the preset queue to generate original trajectory data.

[0185] In one embodiment, when the processor 110 performs the operations of obtaining a drift trajectory point based on the distance proximity, filtering the drift trajectory point from the preset queue, and generating original trajectory data, the processor 110 specifically performs the following operations:

[0186] Obtain a drift trajectory point based on the distance proximity, filter the drift trajectory point from the preset queue, and generate an initial trajectory point sequence;

[0187] Identify a stay point in the initial trajectory point sequence, filter the stay point from the initial trajectory point sequence, and generate original trajectory data.

[0188] In the embodiments of the present application, at least one first trajectory segment not present in the road network map is obtained from the original trajectory data, a preset distance index is used to calculate the similarity between the at least one first trajectory segment, the at least one first trajectory segment is clustered based on the similarity, to obtain a segment cluster, a segment vector of a second trajectory segment in the segment cluster is obtained, and a new road is determined in the road network map based on the segment vector. The first trajectory segment that cannot be matched with the road network map is obtained from the collected original trajectory data, the similarity between the first trajectory segments is measured based on the unknown road mining scene by using the preset distance index, a better clustering effect can be obtained, and then the segment vector of the second trajectory segment in the segment cluster obtained after clustering is used to determine the new road, so that a more accurate road direction can be obtained, and the accuracy of the new road is improved. Moreover, the first trajectory segment that cannot be matched is obtained by using the hidden Markov algorithm for map matching. Compared with directly superimposing the trajectory data on the road network topology and determining whether it is a new road according to whether it coincides with the road network topology, a more accurate new road can be obtained. Moreover, the performance of the map matching algorithm based on the hidden Markov model is improved by using the azimuth angle pruning and candidate road filtering method, and compared with the original algorithm, the efficiency and effect are obviously improved. At the same time, the situation of crossing is considered, and the accuracy of trajectory point matching is further improved. In addition, the position information of the trajectory points is collected, and the collected trajectory points in the preset queue are preprocessed, the redundant parts and parts not concerned in the data are removed through drift point filtering and stay point filtering, and the data quality of the original trajectory data is improved. Moreover, the latitude and longitude point coordinates of the second trajectory segment are used as indexes to obtain similar segments, the calculation amount of the clustering process is reduced, and the performance of the clustering algorithm is improved. The average vector direction of the segment vector in the segment cluster is obtained to confirm the center line of the minimum bounding rectangle of the segment cluster, the center line represents the unmatched trajectory in the segment cluster, and the new road is generated based on the center line, so that the accuracy of the position of the new road can be improved.

[0189] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when the program is executed, the processes of the above-mentioned embodiments can be included. The storage medium can be a magnetic disc, an optical disc, a read-only memory (ROM) or a random access memory (RAM).

[0190] The above only describes the preferred embodiments of the present application, and cannot limit the scope of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope of the present application.

Claims

1. A road generation method characterized by comprising: The method comprises the following steps: acquiring at least one first trajectory segment not existing in a road network map from original trajectory data; calculating the similarity between the at least one first trajectory segment by using a preset distance index; performing clustering processing on the at least one first trajectory segment based on the similarity to obtain a segment clustering cluster; acquiring a segment vector of a second trajectory segment in the segment clustering cluster, and determining a newly added road in the road network map based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment clustering cluster; the step of acquiring at least one first trajectory segment not existing in a road network map from original trajectory data comprises: acquiring a matching road within a preset matching range of a first trajectory point in the original trajectory data from the road network map, the first trajectory point being any trajectory point in a plurality of trajectory points collected in the original trajectory data; acquiring a trajectory vector based on a line connecting the first trajectory point and a second trajectory point in the original trajectory data, calculating an included angle between the trajectory vector and a road vector of the matching road, the second trajectory point being a trajectory point collected before the first trajectory point in a time sequence of the original trajectory data; confirming the matching road with the included angle less than an included angle threshold as a first initial candidate road corresponding to the first trajectory point; calculating an observation probability between the first trajectory point and the first initial candidate road; if the maximum value of the observation probability is less than a probability threshold, calculating a trajectory point distance between the first trajectory point and the second trajectory point; if the trajectory point distance is greater than a road threshold, generating a non-penetrable trajectory segment based on a line connecting the first trajectory point and the second trajectory point, and taking the non-penetrable trajectory segment as at least one first trajectory segment not existing in the road network map; calculating a transition probability from the first initial candidate road to a second initial candidate road, the second initial candidate road being an initial candidate road corresponding to the second trajectory point; confirming whether there is a target candidate road matching the first trajectory point in the first initial candidate road based on the observation probability and the transition probability; if not, generating at least one first trajectory segment not existing in the road network map based on the first trajectory point.

2. The method of claim 1, wherein, after the step of confirming the matching road with the included angle less than the included angle threshold as the first initial candidate road corresponding to the first trajectory point, the method further comprises the following steps: calculating a road surface distance between the first initial candidate road and the first trajectory point; eliminating the first initial candidate road with the road surface distance greater than a distance threshold.

3. The method of claim 1, wherein, the step of calculating the similarity between the at least one first trajectory segment by using a preset distance index comprises the following steps: acquiring a midpoint longitude and latitude coordinate of a target trajectory segment in the first trajectory segment, and confirming a grid where the target trajectory segment is located based on the midpoint longitude and latitude coordinate; calculating the similarity between the target trajectory segment and all first trajectory segments in the grid where the target trajectory segment is located and neighbor grids of the grid based on the distance index.

4. The method of claim 1, wherein, The segment vector of the second trajectory segment in the segment cluster is obtained, and a new road in the road network map is determined based on the segment vector, the second trajectory segment being a first trajectory segment belonging to the same segment cluster, comprising: confirming a minimum bounding rectangle of the segment cluster; obtaining an average vector direction of the segment vector of the second trajectory segment, and confirming a center line of the minimum bounding rectangle consistent with the average vector direction, and taking the center line as a corresponding new road of the segment cluster in the road network map.

5. The method according to any one of claims 1 to 4, characterized in that, Before the step of obtaining at least one first trajectory segment not existing in the road network map from the original trajectory data, comprising: collecting position information of the trajectory points, and sequentially saving the position information into a preset queue according to the time sequence of collection; calculating the distance proximity between a third trajectory point and a fourth trajectory point in the preset queue, the third trajectory point being any collected trajectory point in the preset queue, and the fourth trajectory point being two adjacent trajectory points before and after the third trajectory point in the preset queue; based on the distance proximity, confirming whether the third trajectory point is a drift trajectory point; if yes, filtering the third trajectory point from the preset queue to generate the original trajectory data.

6. The method of claim 5, wherein, The step of obtaining the drift trajectory point based on the distance proximity, filtering the drift trajectory point from the preset queue, and generating the original trajectory data, comprising: obtaining the drift trajectory point based on the distance proximity, filtering the drift trajectory point from the preset queue, and generating an initial trajectory point sequence; identifying a stay point in the initial trajectory point sequence, filtering the stay point from the initial trajectory point sequence, and generating the original trajectory data.

7. A road generating apparatus characterized by comprising: The device comprises: The acquisition module is configured to acquire at least one first trajectory segment not present in the road network map from the original trajectory data; the acquisition of the at least one first trajectory segment not present in the road network map from the original trajectory data comprises: acquiring a matching road within a preset matching range of a first trajectory point in the original trajectory data from the road network map; the first trajectory point is any trajectory point in a plurality of trajectory points collected in the original trajectory data; acquiring a trajectory vector based on a connection line of the first trajectory point and a second trajectory point in the original trajectory data, calculating an included angle between the trajectory vector and a road vector of the matching road, and confirming the matching road with the included angle less than an included angle threshold as a first initial candidate road corresponding to the first trajectory point; calculating an observation probability between the first trajectory point and the first initial candidate road; if a maximum value of the observation probability is less than a probability threshold, calculating a trajectory point distance between the first trajectory point and the second trajectory point; if the trajectory point distance is greater than a road threshold, generating a non-penetrable trajectory segment based on the connection line between the first trajectory point and the second trajectory point, and taking the non-penetrable trajectory segment as the at least one first trajectory segment not present in the road network map; calculating a transition probability from the first initial candidate road to a second initial candidate road corresponding to the second trajectory point; and confirming whether there is a target candidate road matching the first trajectory point in the first initial candidate road based on the observation probability and the transition probability, and generating the at least one first trajectory segment not present in the road network map based on the first trajectory point if there is no target candidate road. The similarity confirmation module is configured to calculate a similarity between the at least one first trajectory segment by using a preset distance index. The clustering module is configured to perform clustering processing on the at least one first trajectory segment based on the similarity to obtain a segment clustering cluster. The new road generation module is configured to acquire a segment vector of a second trajectory segment in the segment clustering cluster, and determine a new road in the road network map based on the segment vector.

8. An electronic device, comprising: The processor and the memory are included. The memory stores a computer program, and the computer program is adapted to be loaded and executed by the processor to implement the steps of the method in any one of claims 1 to 6. The storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the method in any one of claims 1 to 6.

9. A storage medium, characterized by ​

Images