Method, device and equipment for determining set membership of POI (Point of Interest) and storage medium

Abstract

The invention provides a method for determining a set membership of a point of interest (POI), and relates to the technical field of artificial intelligence, in particular to the technical fields of map navigation, automatic driving, intelligent traffic and the like. The method comprises the steps that scene information corresponding to a target parent point POI is determined according to attribute information of the target parent point POI and attribute information of a candidate parent point POI, and the scene information is used for representing whether the target parent point POI and the candidate parent point POI are partitioned and / or overlapped or not; the tracks passing through the candidate child point POIs are clustered, clustering track clusters are obtained, and the candidate child point POIs are child point POIs contained in the geographic space corresponding to the target parent point POIs; and according to a corresponding relationship between the target parent point POI and the clustering track cluster and scene information corresponding to the target parent point POI, determining a target child point POI having a parent-child relationship with the target parent point POI from the candidate child point POIs. According to the method, the efficiency and accuracy of determining the set membership of the POI are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of artificial intelligence technology, specifically to technical fields such as map navigation, autonomous driving, and intelligent transportation, and more particularly to a method, apparatus, device, and storage medium for determining parent-child relationships of points of interest (POIs). Background Art

[0002] In geographic information systems, a POI (Point of Interest) can be a house, a shop, a mailbox, a bus stop, and so on. The parent-child relationship between POIs (i.e., the subordinate relationship between multiple POIs) is widely used in map navigation product strategies. For example, in destination guidance scenarios, the navigation destination of a child point inherits the guidance point of the parent point. In another example, in indoor navigation scenarios, the floor navigation results of a parent point (such as a shopping mall) are displayed as a collection of child points. Summary of the Invention

[0003] The present disclosure provides a method, apparatus, device, and storage medium for determining a parent-child relationship between points of interest (POIs).

[0004] According to a first aspect of the present disclosure, a method for determining a parent-child relationship of a point of interest (POI) is provided, comprising: determining scene information corresponding to a target parent point POI based on attribute information of the target parent point POI and attribute information of candidate parent point POIs, wherein the scene information is used to characterize whether there is partitioning and / or overlap between the target parent point POI and the candidate parent point POI; clustering trajectories passing through the candidate child point POIs to obtain clustered trajectory clusters, wherein the candidate child point POIs are child point POIs contained in the geographic space corresponding to the target parent point POI; and determining, from the candidate child point POIs, a target child point POI having a parent-child relationship with the target parent point POI based on the correspondence between the target parent point POI and the clustered trajectory clusters and the scene information corresponding to the target parent point POI.

[0005] According to a second aspect of the present disclosure, a device for determining a parent-child relationship of a point of interest (POI) is provided, comprising: an information determination module configured to determine scene information corresponding to a target parent point POI based on attribute information of the target parent point POI and attribute information of a candidate parent point POI, wherein the scene information is used to characterize whether there is partitioning and / or overlap between the target parent point POI and the candidate parent point POI; a clustering module configured to cluster trajectories passing through candidate child point POIs to obtain clustered trajectory clusters, wherein the candidate child point POIs are child point POIs contained in the geographic space corresponding to the target parent point POI; and a child point determination module configured to determine, from the candidate child point POIs, a target child point POI having a parent-child relationship with the target parent point POI based on the correspondence between the target parent point POI and the clustered trajectory clusters and the scene information corresponding to the target parent point POI.

[0006] According to a third aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described in any implementation manner in the first aspect.

[0007] According to a fourth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method described in any implementation manner of the first aspect.

[0008] According to a fifth aspect of the present disclosure, a computer program product is provided, comprising a computer program, which implements the method described in any implementation manner of the first aspect when executed by a processor.

[0009] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are provided to facilitate a better understanding of the present invention and do not constitute a limitation of the present disclosure. Figure 1 is an exemplary system architecture diagram in which the present disclosure may be applied; Figure 2 is a flowchart of an embodiment of a method for determining a parent-child relationship of a point of interest (POI) according to the present disclosure; Figure 3 is a flowchart of another embodiment of a method for determining a parent-child relationship of a point of interest (POI) according to the present disclosure; Figure 4-1 This is a schematic diagram of a simple scenario; Figure 4-2 It is a schematic diagram of the partitioning scenario; Figure 4-3 It is a schematic diagram of overlapping scenes; Figure 4-4 It is a schematic diagram of the partition overlap scenario; Figure 5 is a flowchart of another embodiment of a method for determining a parent-child relationship of a point of interest (POI) according to the present disclosure; Figure 6 yes Figure 5 A flowchart of the step of determining the initial sub-point POI; Figure 7 yes Figure 5 A flow chart of the steps of determining the target sub-point POI; Figure 8 1 is a schematic structural diagram of an embodiment of an apparatus for determining a parent-child relationship of a point of interest (POI) according to the present disclosure; Figure 9 It is a block diagram of an electronic device for implementing the method for determining the parent-child relationship of points of interest (POI) according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0011] The following description of exemplary embodiments of the present disclosure is made in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding. These details should be considered as merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0012] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments in the present disclosure may be combined with each other. The present disclosure will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0013] Figure 1 An exemplary system architecture 100 is shown to which an embodiment of a method for determining a parent-child relationship of a point of interest (POI) or an apparatus for determining a parent-child relationship of a point of interest (POI) of the present disclosure may be applied.

[0014] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, 103, and 104, a network 105, and a server 106. Network 105 is a medium for providing communication links between terminal devices 101, 102, 103, and 104 and server 106. Network 105 may include various connection types, such as wired or wireless communication links or fiber optic cables.

[0015] Users can use terminal devices 101, 102, 103, 104 to interact with server 106 via network 105 to receive or send information, etc. Various client applications can be installed on terminal devices 101, 102, 103, 104.

[0016] Terminal devices 101, 102, 103, and 104 can be hardware or software. When terminal devices 101, 102, 103, and 104 are hardware, they can be various electronic devices, including but not limited to smartphones, tablet computers, laptop computers, and desktop computers. When terminal devices 101, 102, 103, and 104 are software, they can be installed in the aforementioned electronic devices. They can be implemented as multiple software programs or software modules, or as a single software program or software module. This is not specifically limited here.

[0017] The server 106 can provide various services. For example, the server 106 can analyze and process the attribute information of the target parent point POI and the attribute information of the candidate parent point POI obtained from the terminal devices 101, 102, 103, and 104, and generate a processing result (such as the target child point POI).

[0018] It should be noted that server 106 can be either hardware or software. When server 106 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or as a single server. When server 106 is software, it can be implemented as multiple software programs or software modules (for example, to provide distributed services), or as a single software program or software module. This is not specifically limited here.

[0019] It should be noted that the method for determining the parent-child relationship of points of interest (POI) provided in the embodiments of the present disclosure is generally executed by the server 106 . Accordingly, the apparatus for determining the parent-child relationship of points of interest (POI) is generally provided in the server 106 .

[0020] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is merely illustrative. Any number of terminal devices, networks and servers may be provided as required.

[0021] Continue to refer Figure 2 , which shows a process 200 of an embodiment of a method for determining the parent-child relationship of a point of interest (POI) according to the present disclosure. The method for determining the parent-child relationship of a point of interest (POI) includes the following steps: Step 201: Determine the scene information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI.

[0022] In this embodiment, the execution subject (eg Figure 1 The server 106 shown in the figure determines the scene information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI, wherein the scene information is used to indicate whether there is a partition and / or overlap between the target parent point POI and the candidate parent point POI.

[0023] When drawing a map, a large number of POIs are collected. You can determine whether a POI is a parent POI by its name or category, thereby obtaining a parent POI set containing a large number of parent POIs. It should be noted that a parent POI refers to the top node in the POI hierarchy in a geographic information system, such as a large building, park, scenic spot, etc., which usually contains many child POIs within its spatial range.

[0024] Next, any parent POI from the parent POI set is selected as the target parent POI. The execution entity first determines candidate parent POIs corresponding to the target parent POI. Candidate parent POIs are POIs of the same category as the target parent POI within a preset distance range. For example, all POIs within 1 kilometer of the target parent POI are traversed and selected as candidate parent POIs if they are of the same category as the target parent POI. It should be noted that the preset distance range can be set based on actual circumstances and is not specifically limited in this embodiment.

[0025] Afterwards, the execution entity will obtain the attribute information of the target parent point POI and the attribute information of the candidate parent point POI, and then determine the scene information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI. Here, the attribute information may include name information, location information, etc. The scene information is used to indicate whether the target parent point POI and the candidate parent point POI are partitioned and / or overlapped. That is, the target parent point POI and the candidate parent point POI may have neither partitions nor overlaps, which is called a simple scene; the target parent point POI and the candidate parent point POI may only have partitions, which is called a partition scene; the target parent point POI and the candidate parent point POI may only have overlaps, which is called an overlapping scene; of course, the target parent point POI and the candidate parent point POI may also have both partitions and overlaps, which is called a partition overlapping scene.

[0026] When the scenario information indicates that the target parent POI is partitioned and overlapped, the execution entity may also determine information such as the names of the partitioned POIs of the target parent POI. A partitioned POI is a POI that creates a partition with the target parent POI. The execution entity may also determine information such as the names of the overlapping POIs of the target parent POI. An overlapping POI is a POI that overlaps with the target parent POI. It is understood that the number of partitioned POIs is greater than or equal to two, and the number of overlapping POIs is greater than or equal to one.

[0027] Step 202: Cluster the trajectories passing through the candidate sub-points POI to obtain clustered trajectory clusters.

[0028] In this embodiment, the execution entity clusters the trajectories passing through the candidate sub-points POI to obtain clustered trajectory clusters, wherein the candidate sub-points POI are sub-points POI contained in the geographic space corresponding to the target parent point POI.

[0029] Specifically, the execution entity first identifies all sub-points (POIs) contained in the geographic space corresponding to the target parent POI and selects these sub-points (POIs) as candidate sub-points (POIs). This means that there is at least one candidate sub-point POI. It then obtains the user's POI arrival trajectory that passes through any of the candidate sub-points (POIs), thereby generating a user trajectory set. Finally, the trajectories in the user trajectory set are clustered based on the similarity between any two trajectories in the set, thereby generating clustered trajectory clusters.

[0030] For example, a user's arrival trajectory for any candidate POI along the route within a preset time period (e.g., 30 days) is obtained. Specifically, the last X meters of the user's trajectory before reaching the candidate POI is intercepted. For example, X can be 300, meaning the last 300 meters of the user's trajectory before reaching the candidate POI are captured. Then, trajectory points with dwell times exceeding 10 minutes or abnormal speed are removed. Missing trajectory points are repaired using linear interpolation or Kalman filtering to obtain a smoothed trajectory. Finally, a clustering algorithm such as DBSCAN (Density-Based Spatial Clustering of Applications with Noise) is used to generate multiple clusters of trajectories. Clusters with more than X trajectories are retained (where X is a positive integer that can be set based on actual business needs).

[0031] Step 203 : According to the correspondence between the target parent point POI and the cluster trajectory cluster and the scene information corresponding to the target parent point POI, a target child point POI having a parent-child relationship with the target parent point POI is determined from the candidate child point POIs.

[0032] In this embodiment, the execution entity determines, from the candidate child POIs, a target child POI that has a parent-child relationship with the target parent POI based on the correspondence between the target parent POI and the clustered trajectory clusters and the scene information corresponding to the target parent POI. Specifically, the execution entity may first filter out initial child POIs from the candidate child POIs based on the correspondence between the target parent POI and the clustered trajectory clusters, and then filter out the target child POI from the initial child POIs based on the scene information of the target parent POI.

[0033] Here, the correspondence between the target parent POI and the clustered trajectory clusters can mean that the target parent POI is located in one cluster within the clustered trajectory clusters, or that the target parent POI is located in multiple clusters within the clustered trajectory clusters. That is, based on the correspondence between the target parent POI and the clustered trajectory clusters, the initial child POIs that may have a parent-child relationship with the target parent POI can be preliminarily determined.

[0034] Then, combined with the target parent's scene information, the target child POI is determined from the initial child POI. The target child POI is the child POI that has a parent-child relationship with the target parent POI. For example, if the scene information is a simple scene or a partitioned scene, the initial child POI can be directly used as the target child POI. If the scene information is an overlapping scene or a partitioned overlapping scene, the target child POI needs to be further determined based on the child category of the initial child POI and the parent category of the target parent POI.

[0035] The method for determining the parent-child relationship of a point of interest (POI) provided in the disclosed embodiment first determines the scene information corresponding to the target parent POI based on the attribute information of the target parent POI and the attribute information of the candidate parent POIs; then clusters the trajectories passing through the candidate child POIs to obtain clustered trajectory clusters; finally, based on the correspondence between the target parent POI and the clustered trajectory clusters and the scene information corresponding to the target parent POI, determines the target child POI that has a parent-child relationship with the target parent POI from the candidate child POIs. The method for determining the parent-child relationship of a point of interest in this embodiment determines the target child POI corresponding to the target parent POI by combining the scene information of the target parent POI and the clustered trajectory clusters corresponding to the arrival trajectories of the candidate child POIs. That is, the method first explores the spatial correlation between multiple POIs and comprehensively determines the target child POI that has a parent-child relationship with the target parent POI based on the correlation between the scene information and the trajectory clusters to which the POIs belong. This allows for more accurate determination of the parent-child relationship of POIs, improving the efficiency and accuracy of determining the parent-child relationship of POIs.

[0036] In addition, in the technical solutions involved in this disclosure, the acquisition, storage, use, processing, transportation, provision and disclosure of user personal information involved (such as the attribute information of the target parent point POI and the attribute information of the candidate parent point POI involved in this disclosure) are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0037] Continue to refer Figure 3 , Figure 3 A process 300 of another embodiment of a method for determining a parent-child relationship of a point of interest (POI) according to the present disclosure is shown. The method for determining a parent-child relationship of a point of interest (POI) comprises the following steps: Step 301: Determine the partition information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI.

[0038] In this embodiment, the execution subject (eg Figure 1The server 106 (shown) determines the zone information corresponding to the target parent POI based on the attribute information of the target parent POI and the attribute information of the candidate parent POIs. The process for determining the target parent POI, the attribute information of the target parent POI, the candidate parent POIs, and the attribute information of the candidate parent POIs can be found in the previous embodiment and will not be further described here.

[0039] The partition information here refers to whether the target parent point POI and the candidate parent point POI are in a partition relationship. For example, if the target parent point POI is XX Mall and the candidate parent point POI is Area A of XX Mall, the similarity between the names of the target parent point POI and the candidate parent point POI can be used to determine that Area A of XX Mall is the partition POI of XX Mall. At this time, the partition information of the target parent point POI is determined to be partitioned.

[0040] In some optional implementations of this embodiment, the attribute information includes: name information and location information; and step 301 includes: Step 3011: Based on the similarity between the name information of the target parent point POI and the name information of the candidate parent point POI, it is determined whether there is a first parent point POI that meets a first preset condition among the candidate parent point POIs.

[0041] The above-mentioned execution entity can remove the partition words (such as District A, District 1, East District, etc.) in the names of the target parent point POI and the candidate parent point POI through the regular template, and then calculate the similarity (such as cosine similarity, etc.) between the name of the target parent point POI and the name of the candidate parent point POI. If the similarity meets the preset similarity threshold requirement, the candidate parent point POI meets the first preset condition, where the first preset condition is the name condition, and the candidate parent point POI that meets the name condition is used as the first parent point POI.

[0042] Step 3012: In response to determining that a first parent point POI exists, based on the position information of the bounding box of the surface of interest where the first parent point POI is located and the position information of the bounding box of the surface of interest where the target parent point POI is located, determine whether a second parent point POI that meets a second preset condition exists in the first parent point POI.

[0043] After determining a first parent POI that meets a first preset condition with the target parent POI, the execution entity will further determine whether a second parent POI within the first parent POI meets a second preset condition, based on the position information of the bounding box of the area of ​​interest (AOI) where the first parent POI resides and the position information of the bounding box of the target parent POI. The second preset condition is the location condition. In other words, the execution entity will determine a parent POI from the first parent POI that meets the location condition with the target parent POI and use it as the second parent POI.

[0044] In some optional implementations of this embodiment, step 3012 includes: (1) In response to determining that both the target parent point POI and the first parent point POI have bounding boxes, and the bounding box of the first parent point POI is within the bounding box of the target parent point POI, it is determined that the first parent point POI has a second parent point POI that meets a second preset condition.

[0045] If the AOI where the target parent point POI is located and the AOI where the first parent point POI is located both have bounding boxes, and the positional relationship between the AOI bounding box of the target parent point POI and the AOI bounding box of the first parent point POI is: the AOI bounding box of the first parent point POI is within the AOI bounding box of the target parent point POI, then it can be directly determined that the current first parent point POI meets the second preset condition and is used as the second parent point POI.

[0046] It should be noted that in map data production, a border will be drawn for POIs with clear boundaries or building entities to identify the spatial range of the POI, namely the AOI border.

[0047] (2) In response to determining that the interest surface where the target parent point POI is located has a border, the interest surface where the first parent point POI is located has no border, and the coordinates of the first parent point POI are located within the border of the interest surface where the target parent point POI is located, determining that there is a second parent point POI in the first parent point POI that meets the second preset condition.

[0048] If the AOI where the target parent point POI is located has a border and the AOI where the first parent point POI is located has no border, and the positional relationship between the AOI border of the target parent point POI and the first parent point POI is: the coordinates of the first parent point POI are within the AOI border of the target parent point POI, then it can be directly determined that the current first parent point POI meets the second preset condition and is used as the second parent point POI.

[0049] Therefore, based on the positional relationship between the AOI bounding box of the target parent point POI and the AOI bounding box of the first parent point POI, it is accurately determined from the first parent point POI whether there is a second parent point POI that meets the position requirements, thereby improving the accuracy of the partition information of the target parent point POI.

[0050] Step 3013: In response to determining that the second parent point POI exists, determine that the partition information corresponding to the target parent point POI is an existing partition.

[0051] After determining that there is a second parent point POI in the first parent point POI that meets the location requirement of the target parent point POI, the execution entity will determine that the target parent point POI has a partition, and the partition POI of the target parent point POI is the second parent point POI.

[0052] Therefore, it is judged whether there is a partition POI of the target parent point POI among the candidate parent point POIs through the name condition and the location condition, thereby accurately determining the partition information of the target parent point POI.

[0053] Step 302 : constructing first multi-dimensional feature information and second multi-dimensional feature information according to the attribute information of the target parent point POI and the attribute information of the candidate parent point POI respectively.

[0054] In this embodiment, the execution entity constructs the first multi-dimensional feature information according to the attribute information of the target parent point POI, and constructs the second multi-dimensional feature information according to the attribute information of the candidate parent point POI.

[0055] The multidimensional feature information construction process here is as follows: First, multidimensional features are extracted, including: a sub-point category presence feature (N-dimensional): a Boolean value of 1 for the presence of a shopping sub-point and 0 for the absence of a shopping sub-point; a sub-point category proportion feature (N-dimensional): the proportion of shopping sub-points = the number of shopping sub-points divided by the total number of sub-points, a continuous value ranging from 0 to 1; the top 1 industry for sub-point proportion (1-dimensional): an integer from 1 to N; the top 2 industries for sub-point proportion (1-dimensional): an integer from 1 to N; the top 3 industries for sub-point proportion (1-dimensional): an integer from 1 to N; an industry monopoly degree feature (1-dimensional): (top 1 proportion - top 2 proportion) / top 1 proportion, used to measure the degree of industry concentration within a region; and the total number of sub-points (1-dimensional). These features are then concatenated into a single vector to obtain the multidimensional feature information described above.

[0056] By performing the above process on the target parent point POI and the candidate parent point POI respectively, the first multi-dimensional feature information corresponding to the target parent point POI and the second multi-dimensional feature information corresponding to the candidate parent point POI can be obtained.

[0057] Step 303: Determine overlapping information corresponding to the target parent point POI based on the similarity between the first multidimensional feature information and the second multidimensional feature information.

[0058] In this embodiment, the execution entity calculates the similarity between the first multi-dimensional feature information and the second multi-dimensional feature information, and determines the overlapping information corresponding to the target parent point POI according to the similarity result.

[0059] Here, parent point overlap scenarios are defined as multiple physically overlapping parent points within the same spatial range. For example, if a shopping mall and an office building share the same building, the mall and office building are considered overlapping, and the value is a Boolean value (0 for non-overlapping, 1 for overlapping). In particular, parent-child relationships (such as a convenience store within a shopping mall) are not considered overlapping, nor are multiple instances of the same parent class (such as two supermarkets or two office buildings).

[0060] In some optional implementations of this embodiment, the execution entity may also pre-train an overlapping scene recognition model, and use the overlapping scene recognition model to determine the overlapping information of the target parent point POI.

[0061] Specifically, the POI classification system commonly used in the map industry can be adopted to select N typical sub-point POI categories (which are required to represent the purpose of the building, such as the company category, because sub-points of the company category are widely present in commercial office buildings, but will not appear in shopping malls), including transportation facilities, food, tourist attractions, companies and enterprises, etc.

[0062] Then, a feature vector corresponding to the parent POI is constructed. Specifically, the following features are used: Child category presence feature (N-dimensional): A Boolean value of 1 indicates the presence of a shopping-related child; a Boolean value of 0 indicates the absence of a shopping-related child. Child category proportion feature (N-dimensional): The proportion of shopping-related children is calculated by dividing the number of shopping-related children by the total number of children, a continuous value between 0 and 1. The industry with the highest proportion of children (1-dimensional): A value between 1 and N integers. The industry with the highest proportion of children (1-dimensional): A value between 1 and N integers. The industry with the highest proportion of children (1-dimensional): A value between 1 and N integers. The industry monopoly degree feature (1-dimensional): (top 1 proportion - top 2 proportion) / top 1 proportion, which measures the concentration of industries within a region. The total number of children (1-dimensional). Finally, these features are concatenated into a single vector to obtain the feature vector corresponding to the parent POI.

[0063] Next, we define parent point overlap scenarios: If multiple physically overlapping parent points exist within the same spatial range, this scenario is considered overlapping. For example, if a shopping mall and an office building share the same building, this scenario is considered overlapping. The value is a Boolean value (0 for non-overlapping, 1 for overlapping). In particular, parent-child relationships (such as a convenience store within a shopping mall) are not considered overlapping. Multiple instances of the same parent class (such as two supermarkets or two office buildings) are also not considered overlapping.

[0064] Next, the dataset is screened and manually annotated. Specifically, all AOI bounding boxes of the parent point POI are obtained. Spatial calculation functions are used to automatically filter out pairs of parent points whose AOI boundaries intersect. These pairs are considered overlapping parent points and added to the candidate set of overlapping parent points. If a parent point POI does not have any other parent points whose AOI boundaries intersect, the parent point POI is considered a non-overlapping parent point and added to the candidate set of non-overlapping parent points. The true label values ​​of the candidate parent points (including both overlapping and non-overlapping parent points) are then manually verified to obtain the true values ​​corresponding to the training set samples.

[0065] Finally, the training set is trained using machine learning models such as Xgboost (distributed gradient boosting library) and logistic regression to generate an overlapping scene recognition model. This trained overlapping scene recognition model can then be used to determine the overlapping information of the target parent point POI.

[0066] Step 304: The partition information and the overlapping information are integrated to obtain the scene information corresponding to the target parent point POI.

[0067] In this embodiment, the above-mentioned execution entity will integrate the partition information and overlap information to obtain the scene information of the target parent point POI. The scene information is used to characterize whether the target parent point POI and the candidate parent point POI have partitions and / or overlaps, that is, the target parent point POI and the candidate parent point POI may have neither partitions nor overlaps, the target parent point POI and the candidate parent point POI may only have partitions, the target parent point POI and the candidate parent point POI may only have overlaps, or the target parent point POI and the candidate parent point POI may have both partitions and overlaps.

[0068] In some optional implementations of this embodiment, step 304 includes: in response to determining that the partition information is that there is no partition and the overlap information is that there is no overlap, determining that the scene information corresponding to the target parent point POI is a simple scene; in response to determining that the partition information is that there is a partition and the overlap information is that there is no overlap, determining that the scene information corresponding to the target parent point POI is a partition scene; in response to determining that the partition information is that there is no partition and the overlap information is that there is overlap, determining that the scene information corresponding to the target parent point POI is an overlapping scene; in response to determining that the partition information is that there is a partition and the overlap information is that there is overlap, determining that the scene information corresponding to the target parent point POI is a partition overlapping scene.

[0069] In this implementation, if the target parent point POI and the candidate parent point POI have neither partitions nor overlaps, the scene information of the target parent point POI is determined to be a simple scene. Figure 4-1 A schematic diagram of a simple scenario, such as Figure 4-1 As shown, the target parent point POI is XX shopping mall, which has neither partitions nor overlaps, and its scene information is determined to be a simple scene.

[0070] If the target parent point POI and the candidate parent point POI are partitioned but not overlapped, the scene information of the target parent point POI is determined to be a partitioned scene. Figure 4-2 is a schematic diagram of the partition scenario, such as Figure 4-2 As shown, the target parent point POI is XX shopping mall, which has partitions but no overlap. Its scene information is determined to be a partitioned scene, and the partitioned POIs are area A and area B.

[0071] If the target parent point POI and the candidate parent point POI do not have partitions but overlap, the scene information of the target parent point POI is determined to be an overlapping scene. Figure 4-3 is a schematic diagram of overlapping scenes, such as Figure 4-3 As shown, the target parent point POI is XX shopping mall, which overlaps with the office building but does not have a partition. Its scene information is determined to be an overlapping scene, and the overlapping POI is the office building.

[0072] If the target parent point POI and the candidate parent point POI are both partitioned and overlapped, the scene information of the target parent point POI is determined to be a partition overlapping scene. Figure 4-4 This is a schematic diagram of the partition overlap scenario, such as Figure 4-4 As shown, the target parent point POI is XX shopping mall, which has partitions A and B, and overlaps with office buildings A and B. Its scene information is determined to be a partition overlapping scene.

[0073] Thus, the scene information of the target parent point POI is accurately determined by comprehensively analyzing the partition and overlap between the target parent point POI and the candidate parent point POI.

[0074] Step 305: Cluster the trajectories passing through the candidate sub-points POI to obtain clustered trajectory clusters.

[0075] Step 306 : According to the correspondence between the target parent point POI and the cluster trajectory cluster and the scene information corresponding to the target parent point POI, a target child point POI having a parent-child relationship with the target parent point POI is determined from the candidate child point POIs.

[0076] Steps 305-306 are basically the same as steps 202-203 of the aforementioned embodiment. For specific implementation methods, reference can be made to the aforementioned description of steps 202-203, which will not be repeated here.

[0077] from Figure 3 It can be seen that Figure 2 Compared with the corresponding embodiments, the method for determining the parent-child relationship of the point of interest (POI) in this embodiment highlights the step of determining the scene information corresponding to the target parent point POI. The method first determines the partition information of the target parent point POI based on the name and location information of the target parent point POI and the candidate parent point POI, and then determines the overlapping information of the target parent point POI based on the multi-dimensional feature information of the target parent point POI and the candidate parent point POI. Finally, the scene information is generated by combining the partition information and the overlapping information, thereby accurately determining the scene information of the target parent point POI, thereby improving the accuracy of the subsequent determination of the child point POI corresponding to the target parent point POI based on the scene information.

[0078] Continue to refer Figure 5 , Figure 5A process 500 of another embodiment of a method for determining a parent-child relationship of a point of interest (POI) according to the present disclosure is shown. The method for determining a parent-child relationship of a point of interest (POI) comprises the following steps: Step 501: Determine the partition information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI.

[0079] Step 502 : constructing first multi-dimensional feature information and second multi-dimensional feature information according to the attribute information of the target parent point POI and the attribute information of the candidate parent point POI respectively.

[0080] Step 503: Determine overlapping information corresponding to the target parent point POI based on the similarity between the first multidimensional feature information and the second multidimensional feature information.

[0081] Step 504: The partition information and the overlapping information are integrated to obtain the scene information corresponding to the target parent point POI.

[0082] Steps 501-504 are basically the same as steps 301-304 of the aforementioned embodiment. For specific implementation methods, reference can be made to the aforementioned description of steps 301-304, which will not be repeated here.

[0083] Step 505: Divide the geographic space corresponding to the target parent point POI to obtain at least one grid.

[0084] In this embodiment, the execution subject (eg Figure 1 The server 106 shown in FIG2 will divide the geographic space corresponding to the target parent point POI into at least one grid. For example, the geographic space corresponding to the target parent point POI is divided into grids of fixed size (e.g., 10 km 10km), thereby obtaining at least one grid.

[0085] Step 506 : For each grid, obtain at least one trajectory passing through the candidate sub-point POI in the grid, and cluster the at least one trajectory based on the similarity between any two trajectories in the at least one trajectory to obtain a trajectory cluster corresponding to the grid.

[0086] In this embodiment, for each grid, the execution entity obtains at least one trajectory that passes through a candidate POI in that grid. For example, at least one user arrival trajectory that passes through a candidate POI in that grid within a preset time period (e.g., 30 days) is obtained. Here, the trajectory of the user reaching the candidate POI can be intercepted for the last X meters before the candidate POI, where X can be 300, for example. Then, trajectory points with a dwell time exceeding 10 minutes or abnormal speed (e.g., excessive speed) are removed. Missing trajectory points are repaired using linear interpolation or Kalman filtering to obtain a smooth trajectory sequence.

[0087] Each trajectory point can be expressed as: , where x and y are the latitude and longitude information of the trajectory point, and t is the arrival time information.

[0088] Furthermore, a trajectory sequence Expressed as: ； in, , that is, the trajectory sequence is arranged according to the arrival time.

[0089] Afterwards, the execution entity clusters the at least one trajectory based on the similarity between any two trajectories in the at least one trajectory to obtain a trajectory cluster corresponding to the grid.

[0090] Here, we can use methods such as DTW (Dynamic Time Warping) to calculate the similarity between each two trajectories. Take the DTW algorithm as an example: Distance metric between two trajectory points for: ; Among them, the spatial weight coefficient (0.7) and the temporal weight coefficient (0.3) can be customized.

[0091] Cumulative distance of dynamic time warping of two trajectory sequences The formula is: ; in, It represents the minimum cumulative distance between the first i points of sequence X and the first j points of sequence Y.

[0092] Afterwards, cluster at least one trajectory using a clustering algorithm such as DBSCAN to obtain the trajectory cluster corresponding to the grid, and retain clusters with a trajectory count greater than X (X can be set according to actual business needs).

[0093] Step 507: merge the trajectory clusters corresponding to at least one grid to obtain a clustered trajectory cluster.

[0094] In this embodiment, the execution entity merges the trajectory clusters corresponding to all grids to obtain clustered trajectory clusters.

[0095] Therefore, through grid division, the user's POI arrival trajectory of the candidate child points along the way is obtained. By clustering the POI arrival trajectory, the problem of missing POI parent-child relationships in scenarios without text mentions is solved, which improves the comprehensiveness and accuracy of determining POI parent-child relationships.

[0096] Step 508 : determining an initial child point POI from the candidate child points POI according to the correspondence between the target parent point POI and the clustered trajectory cluster.

[0097] In this embodiment, the execution entity determines an initial child POI from the candidate child POIs based on the correspondence between the target parent POI and the clustered trajectory clusters. Here, the correspondence between the target parent POI and the clustered trajectory clusters can mean that the target parent POI is located in one cluster within the clustered trajectory clusters, or that the target parent POI is located in multiple clusters within the clustered trajectory clusters. In other words, based on the correspondence between the target parent POI and the clustered trajectory clusters, initial child POIs that may have a parent-child relationship with the target parent POI can be preliminarily determined. Typically, there are multiple initial child POIs.

[0098] Step 509 : determining a target child point POI having a parent-child relationship with the target parent point POI from the initial child point POI according to the scene information.

[0099] In this embodiment, the execution entity determines a target child POI that has a parent-child relationship with the target parent POI from the initial child POI based on the scene information. Specifically, the target child POI is determined from the initial child POI based on the scene information of the target parent POI. The target child POI is the child POI that has a parent-child relationship with the target parent POI.

[0100] For example, if the scene information is a simple scene or a partitioned scene, the initial sub-point POI can be directly used as the target sub-point POI; if the scene information is an overlapping scene or a partitioned overlapping scene, it is necessary to further comprehensively determine the target sub-point POI based on information such as the sub-point category of the initial sub-point POI and the parent point category of the target parent point POI.

[0101] from Figure 5 It can be seen that Figure 3Compared with the corresponding embodiments, the method for determining the parent-child relationship of the point of interest (POI) in this embodiment highlights the step of determining the target child point (POI) based on the scene information of the target parent point and the clustered trajectory cluster. By combining the scene information with the correlation between the clustered trajectory clusters corresponding to the POI arrival trajectory, the target child point (POI) is comprehensively determined, thereby more accurately determining the target child point (POI) having a parent-child relationship with the target parent point (POI), thereby improving the accuracy of the target child point (POI).

[0102] Continue to refer Figure 6 , Figure 6 Shown Figure 5 A process 600 of determining an initial sub-point POI in the embodiment of the present invention. This step includes: Step 601 : In response to determining that a target parent point POI exists only in a first trajectory cluster among clustered trajectory clusters, all candidate child point POIs included in the first trajectory cluster are used as initial child point POIs.

[0103] If the execution entity determines that the target parent point POI exists only in one trajectory cluster among the clustered trajectory clusters, that is, the target parent point POI exists only in the first trajectory cluster among the clustered trajectory clusters, then all candidate child point POIs contained in the first trajectory cluster are used as initial child point POIs.

[0104] Step 602 : In response to determining that the target parent point POI exists in multiple trajectory clusters in the clustered trajectory cluster, an initial child point POI is determined from all candidate child point POIs included in the multiple trajectory clusters according to the partition information of the target parent point POI.

[0105] If the execution entity determines that the target parent point POI exists in multiple trajectory clusters in the clustered trajectory cluster, then the initial child point POI will be determined from all candidate child point POIs contained in these multiple trajectory clusters according to the partition information of the target parent point POI.

[0106] In some optional implementations of this embodiment, step 602 includes: Step 6021: In response to determining that the partition information of the target parent point POI indicates that no partition exists, all candidate child point POIs included in the plurality of trajectory clusters are used as initial child point POIs.

[0107] If the partition information of the target parent point POI indicates that the target parent point POI does not have any partition, then all candidate child point POIs contained in the multiple trajectory clusters are directly used as initial child point POIs.

[0108] Step 6022: In response to determining that the partition information of the target parent point POI indicates that a partition exists, the partition child point POI is used as the initial child point POI.

[0109] If the partition information of the target parent point POI indicates that the target parent point POI has a partition, a child point POI having a parent-child relationship with the partition POI of the target parent point POI, namely the partition child point POI, will be determined first, and the partition child point POI will be used as the initial child point POI.

[0110] For example, if the target parent point POI has partitions, and the partitions are area A and area B, then the above execution entity will first determine the child point POIs that have a parent-child relationship with area A, and the child point POIs that have a parent-child relationship with area B, and use the child point POIs of area A and area B as the initial child point POIs.

[0111] Therefore, when the target parent point POI exists in multiple trajectory clusters, the initial child point POI can be accurately determined according to the partition information of the target parent point POI.

[0112] In some optional implementations of this embodiment, the method further includes the step of determining partition sub-points (POIs), specifically, in response to determining that the partition POI corresponding to the target parent point (POI) appears in a second trajectory cluster among the multiple trajectory clusters, taking all candidate sub-points (POIs) contained in the second trajectory cluster as partition sub-points (POIs); in response to determining that the partition POI corresponding to the target parent point (POI) does not appear in the second trajectory cluster among the multiple trajectory clusters, taking all candidate sub-points (POIs) located within the bounding box of the surface of interest where the partition POI is located as partition sub-points (POIs).

[0113] In this implementation, if the partition corresponding to the target parent point POI appears in a trajectory cluster among multiple trajectory clusters, the trajectory cluster where it appears is recorded as the second trajectory cluster, and all candidate sub-point POIs contained in the second trajectory cluster are used as partition sub-point POIs, where the number of second trajectory clusters can be one or more.

[0114] If the partition corresponding to the target parent point POI does not appear in any of the multiple trajectory clusters, the partition sub-point POI is determined according to the positional relationship between the AOI bounding box where the partition POI is located and the candidate sub-point POI, that is, all candidate sub-point POIs located within the AOI bounding box of the partition POI are used as the partition sub-point POI.

[0115] Therefore, according to the correspondence between the target parent point POI and the clustered trajectory cluster, the initial sub-point POI can be accurately and quickly determined from the candidate sub-point POIs.

[0116] Continue to refer Figure 7 , Figure 7 Shown Figure 5 A process 700 of determining a target sub-point POI in the embodiment of the present invention. This step includes: Step 701: In response to determining that the scene information is a simple scene or a partitioned scene, an initial child point POI is determined as a target child point POI having a parent-child relationship with a target parent point POI.

[0117] If the scene information is a simple scene, the initial child point POI is directly determined as the target child point POI having a parent-child relationship with the target parent point POI, that is, the initial child point POI is directly determined as the child point POI of the target parent point POI.

[0118] If the scene information is a partitioned scene, the initial child point POI is directly determined as the target child point POI that has a parent-child relationship with the target parent point POI. Specifically, the output is the partition relationship between the target parent point POI and the partition POI, as well as the child point POI corresponding to the partition POI.

[0119] For example, if the target parent POI is partitioned into Area A and Area B, the initial child POIs include the child POIs in Area A and Area B, and the initial child POI is the target child POI. The final output is the partition relationship between the target parent POI and Area A and Area B, as well as the child POIs in Area A and Area B.

[0120] In step 702, in response to determining that the scene information is an overlapping scene or a partitioned overlapping scene, a target child point POI having a parent-child relationship with the target parent point POI is determined from the initial child point POI based on the category of the initial child point POI, the category of the target parent point POI, and the similarity between the name information of the initial child point POI and the name information of the target parent point POI.

[0121] If the scene information is determined to be an overlapping scene, the text similarity between the name or address of the initial child point POI and the names or addresses of multiple overlapping parent point instances is calculated, and a decision is made based on the initial child point POI category, floor, and parent point category to determine the target child point POI that has a parent-child relationship with the target parent point POI from the initial child point POI.

[0122] It is understandable that if the category of the initial child POI is company and the floor is 25, then the parent point of the initial child POI should be an office building; if the category of the initial child POI is food and the floor is B1, then the parent point of the initial child POI should be a shopping mall.

[0123] For example, if the overlapping parent POIs are an office building and a shopping mall, and an initial child POI is classified as food and located on the first floor, and the name similarity of this initial child POI is calculated to be more closely matched to the office building name, then the parent POI of this initial child POI can be determined to be the office building. This means that the target child POI of the office building parent POI is determined to be this initial child POI.

[0124] If the scene information is determined to be a partitioned overlapping scene, then, as with the overlapping scene, it is necessary to calculate the text similarity between the name or address of the initial child point POI and the names or addresses of multiple overlapping parent point instances, and make a comprehensive decision based on the initial child point POI category, floor, and parent point category to determine the target child point POI that has a parent-child relationship with the target parent point POI from the initial child point POI.

[0125] It is understandable that the overlapping parent point instances in the overlapping scene are shopping malls and office buildings (refer to Figure 4-3 ), while the overlapping parent point instances in the partition overlap scenario are Area A and Office Building A, or Area B and Office Building B (refer to Figure 4-4 ).

[0126] Therefore, by combining the scene information with the correlation between the clustered trajectory clusters corresponding to the POI arrival trajectory, the target sub-point POI is comprehensively determined, thereby improving the accuracy of determining the target sub-point POI that has a parent-child relationship with the target parent point POI.

[0127] Further references Figure 8 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a device for determining the parent-child relationship of a point of interest POI. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0128] like Figure 8 As shown, the apparatus 800 for determining the parent-child relationship of a point of interest (POI) in this embodiment includes: an information determination module 801, a clustering module 802, and a child point determination module 803. The information determination module 801 is configured to determine scene information corresponding to a target parent point POI based on attribute information of the target parent point POI and attribute information of candidate parent point POIs, wherein the scene information is used to indicate whether there is a partition and / or overlap between the target parent point POI and the candidate parent point POIs. The clustering module 802 is configured to cluster trajectories passing through the candidate child point POIs to obtain clustered trajectory clusters, wherein the candidate child point POIs are child points contained in the geographic space corresponding to the target parent point POI. The child point determination module 803 is configured to determine, from the candidate child point POIs, target child points that have a parent-child relationship with the target parent point POI based on the correspondence between the target parent point POI and the clustered trajectory clusters and the scene information corresponding to the target parent point POI.

[0129] In this embodiment, in the apparatus 800 for determining the parent-child relationship of a point of interest POI, the specific processing of the information determination module 801, the clustering module 802 and the child point determination module 803 and the technical effects thereof can be referred to respectively. Figure 2The relevant descriptions of steps 201-203 in the corresponding embodiment are not repeated here.

[0130] In some optional implementations of this embodiment, the information determination module 801 includes: a partition information determination submodule, configured to determine the partition information corresponding to the target parent point POI based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI; a feature construction submodule, configured to construct first multidimensional feature information and second multidimensional feature information based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI, respectively; an overlapping information determination submodule, configured to determine the overlapping information corresponding to the target parent point POI based on the similarity between the first multidimensional feature information and the second multidimensional feature information; and a scene information determination submodule, configured to integrate the partition information and overlapping information to obtain the scene information corresponding to the target parent point POI.

[0131] In some optional implementations of this embodiment, the attribute information includes: name information and location information; and the partition information determination submodule includes: a first determination unit, configured to determine whether there is a first parent point POI that meets the first preset condition in the candidate parent point POI based on the similarity between the name information of the target parent point POI and the name information of the candidate parent point POI; a second determination unit, configured to, in response to determining that there is a first parent point POI, determine whether there is a second parent point POI that meets the second preset condition in the first parent point POI based on the location information of the bounding box of the interest surface where the first parent point POI is located and the location information of the bounding box of the interest surface where the target parent point POI is located; a third determination unit, configured to, in response to determining that there is a second parent point POI, determine that the partition information corresponding to the target parent point POI is an existing partition.

[0132] In some optional implementations of this embodiment, the second determination unit is further configured to: in response to determining that both the interest surface where the target parent point POI is located and the interest surface where the first parent point POI is located have bounding boxes, and the bounding box of the interest surface where the first parent point POI is located is located within the bounding box of the interest surface where the target parent point POI is located, determine that there is a second parent point POI in the first parent point POI that meets the second preset condition; or in response to determining that the interest surface where the target parent point POI is located has a bounding box, the interest surface where the first parent point POI is located has no bounding box, and the coordinates of the first parent point POI are located within the bounding box of the interest surface where the target parent point POI is located, determine that there is a second parent point POI in the first parent point POI that meets the second preset condition.

[0133] In some optional implementations of this embodiment, the scene information determination submodule is further configured to: in response to determining that the partition information is that there is no partition and the overlap information is that there is no overlap, determine that the scene information corresponding to the target parent point POI is a simple scene; in response to determining that the partition information is that there is a partition and the overlap information is that there is no overlap, determine that the scene information corresponding to the target parent point POI is a partition scene; in response to determining that the partition information is that there is no partition and the overlap information is that there is overlap, determine that the scene information corresponding to the target parent point POI is an overlapping scene; in response to determining that the partition information is that there is a partition and the overlap information is that there is overlap, determine that the scene information corresponding to the target parent point POI is a partition overlapping scene.

[0134] In some optional implementations of this embodiment, the sub-point determination module 803 includes: an initial sub-point determination sub-module, configured to determine an initial sub-point POI from candidate sub-point POIs based on the correspondence between the target parent point POI and the cluster trajectory cluster; and a target sub-point determination sub-module, configured to determine a target sub-point POI having a parent-child relationship with the target parent point POI from the initial sub-point POI based on scene information.

[0135] In some optional implementations of this embodiment, the initial child point determination submodule includes: a first child point determination unit configured to, in response to determining that the target parent point POI exists only in the first trajectory cluster among the clustered trajectory clusters, use all candidate child point POIs included in the first trajectory cluster as initial child point POIs; and a second child point determination unit configured to, in response to determining that the target parent point POI exists in multiple trajectory clusters among the clustered trajectory clusters, determine the initial child point POI from all candidate child point POIs included in the multiple trajectory clusters according to partition information of the target parent point POI.

[0136] In some optional implementations of this embodiment, the second child point determination unit is further configured to: in response to determining that the partition information of the target parent point POI is that no partition exists, use all candidate child point POIs contained in the multiple trajectory clusters as initial child point POIs; in response to determining that the partition information of the target parent point POI is that a partition exists, use the partition child point POI as the initial child point POI, wherein the partition child point POI is a child point POI that has a parent-child relationship with the partition POI corresponding to the target parent point POI.

[0137] In some optional implementations of this embodiment, the apparatus 800 for determining the parent-child relationship of a point of interest (POI) further includes: a first partition sub-point determination module configured to, in response to determining that the partition POI corresponding to the target parent POI appears in a second trajectory cluster among the multiple trajectory clusters, select all candidate sub-points (POIs) contained in the second trajectory cluster as partition sub-points (POIs); and a second partition sub-point determination module configured to, in response to determining that the partition POI corresponding to the target parent POI does not appear in the second trajectory cluster among the multiple trajectory clusters, select all candidate sub-points (POIs) located within a bounding box of the interest surface where the partition POI is located as partition sub-points (POIs).

[0138] In some optional implementations of this embodiment, the target sub-point determination sub-module is further configured to: in response to determining that the scene information is a simple scene or a partitioned scene, determine the initial sub-point POI as a target sub-point POI having a parent-child relationship with the target parent point POI; in response to determining that the scene information is an overlapping scene or a partitioned overlapping scene, determine the target sub-point POI having a parent-child relationship with the target parent point POI from the initial sub-point POI based on the category of the initial sub-point POI, the category of the target parent point POI, and the similarity between the name information of the initial sub-point POI and the name information of the target parent point POI.

[0139] In some optional implementations of this embodiment, the clustering module 802 is further configured to: divide the geographic space corresponding to the target parent point POI to obtain at least one grid; for each grid, obtain at least one trajectory passing through the candidate child point POI in the grid; cluster the at least one trajectory based on the similarity between any two trajectories in the at least one trajectory to obtain a trajectory cluster corresponding to the grid; and merge the trajectory clusters corresponding to the at least one grid to obtain a clustered trajectory cluster.

[0140] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0141] Figure 9 A schematic block diagram of an example electronic device 900 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are provided as examples only and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0142] like Figure 9As shown, device 900 includes a computing unit 901, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 902 or a computer program loaded from a storage unit 908 into a random access memory (RAM) 903. RAM 903 may also store various programs and data required for the operation of device 900. Computing unit 901, ROM 902, and RAM 903 are connected to each other via a bus 904. An input / output (I / O) interface 905 is also connected to bus 904.

[0143] Multiple components in device 900 are connected to I / O interface 905, including: an input unit 906, such as a keyboard, mouse, etc.; an output unit 907, such as various types of displays, speakers, etc.; a storage unit 908, such as a magnetic disk, optical disk, etc.; and a communication unit 909, such as a network card, modem, wireless communication transceiver, etc. The communication unit 909 allows device 900 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0144] The computing unit 901 can be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the computing unit 901 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 901 performs the various methods and processes described above, such as the method for determining parent-child relationships of points of interest (POIs). For example, in some embodiments, the method for determining parent-child relationships of points of interest (POIs) can be implemented as a computer software program tangibly embodied in a machine-readable medium, such as the storage unit 908. In some embodiments, part or all of the computer program can be loaded and / or installed onto the device 900 via the ROM 902 and / or the communication unit 909. When the computer program is loaded into the RAM 903 and executed by the computing unit 901, one or more steps of the method for determining parent-child relationships of points of interest (POIs) described above can be performed. Alternatively, in other embodiments, the computing unit 901 may be configured in any other appropriate manner (eg, by means of firmware) to execute the method for determining the parent-child relationship of points of interest (POIs).

[0145] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0146] The program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. Such program code may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the program code is executed by the processor or controller, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The program code may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0147] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0148] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0149] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), and the Internet.

[0150] A computer system may include a client and a server. The client and server are generally remote from each other and typically interact through a communication network. The client-server relationship arises through computer programs running on the respective computers and having a client-server relationship with each other. The server may be a cloud server, a server in a distributed system, or a server integrated with a blockchain.

[0151] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in this disclosure can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved. This is not limited herein.

[0152] The above specific embodiments do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure shall be included within the scope of protection of this disclosure.

Claims

1. A method for determining a parent-child relationship of a point of interest (POI), comprising: Determining scene information corresponding to the target parent point POI based on attribute information of the target parent point POI and attribute information of the candidate parent point POI, wherein the scene information is used to indicate whether there is a partition and / or overlap between the target parent point POI and the candidate parent point POI; Clustering the trajectories passing through the candidate sub-points POI to obtain clustered trajectory clusters, wherein the candidate sub-points POI are sub-points POI contained in the geographic space corresponding to the target parent point POI; According to the corresponding relationship between the target parent point POI and the cluster trajectory cluster and the scene information corresponding to the target parent point POI, a target child point POI having a parent-child relationship with the target parent point POI is determined from the candidate child points POI.

2. The method according to claim 1, wherein The determining, based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI, the scene information corresponding to the target parent point POI includes: Determining the partition information corresponding to the target parent point POI according to the attribute information of the target parent point POI and the attribute information of the candidate parent point POI; Constructing first multidimensional feature information and second multidimensional feature information according to the attribute information of the target parent point POI and the attribute information of the candidate parent point POI respectively; Determining overlapping information corresponding to the target parent point POI based on the similarity between the first multidimensional feature information and the second multidimensional feature information; The partition information and the overlapping information are combined to obtain the scene information corresponding to the target parent point POI.

3. The method according to claim 2, wherein: The attribute information includes: name information and location information; and The determining, based on the attribute information of the target parent point POI and the attribute information of the candidate parent point POI, the partition information corresponding to the target parent point POI includes: Determining, based on the similarity between the name information of the target parent point POI and the name information of the candidate parent point POI, whether there is a first parent point POI that meets a first preset condition among the candidate parent point POIs; In response to determining that the first parent point POI exists, determining whether there is a second parent point POI that meets a second preset condition in the first parent point POI based on the position information of the bounding box of the surface of interest where the first parent point POI is located and the position information of the bounding box of the surface of interest where the target parent point POI is located; In response to determining that the second parent point POI exists, the partition information corresponding to the target parent point POI is determined to be an existing partition.

4. The method according to claim 3, wherein: The determining, based on the position information of the frame of the surface of interest where the first parent point POI is located and the position information of the frame of the surface of interest where the target parent point POI is located, whether there is a second parent point POI in the first parent point POI that meets a second preset condition includes: In response to determining that both the surface of interest where the target parent point POI is located and the surface of interest where the first parent point POI is located have bounding boxes, and the bounding box of the surface of interest where the first parent point POI is located is within the bounding box of the surface of interest where the target parent point POI is located, determining that the first parent point POI has a second parent point POI that meets a second preset condition; or In response to determining that the surface of interest where the target parent point POI is located has a border, the surface of interest where the first parent point POI is located has no border, and the coordinates of the first parent point POI are located within the border of the surface of interest where the target parent point POI is located, it is determined that a second parent point POI that meets a second preset condition exists in the first parent point POI.

5. The method according to claim 2, wherein: The step of synthesizing the partition information and the overlapping information to obtain scene information corresponding to the target parent point POI includes: In response to determining that the partition information indicates that no partition exists and the overlap information indicates that no overlap exists, determining that the scene information corresponding to the target parent point POI is a simple scene; In response to determining that the partition information indicates that a partition exists and the overlap information indicates that there is no overlap, determining that the scene information corresponding to the target parent point POI is a partition scene; In response to determining that the partition information indicates that no partition exists and the overlap information indicates that overlap exists, determining that the scene information corresponding to the target parent point POI is an overlapping scene; In response to determining that the partition information indicates that a partition exists and the overlap information indicates that an overlap exists, it is determined that the scene information corresponding to the target parent point POI is a partition overlap scene.

6. The method according to claim 5, wherein: The determining, from the candidate child points POIs, a target child point POI having a parent-child relationship with the target parent point POI according to the correspondence between the target parent point POI and the cluster trajectory cluster and the scene information corresponding to the target parent point POI, includes: Determine an initial sub-point POI from the candidate sub-points POI according to the correspondence between the target parent point POI and the cluster trajectory cluster; A target child point POI having a parent-child relationship with the target parent point POI is determined from the initial child point POI according to the scene information.

7. The method according to claim 6, wherein: The determining of the initial sub-point POI from the candidate sub-point POIs according to the correspondence between the target parent point POI and the cluster trajectory cluster includes: In response to determining that the target parent point POI exists only in a first trajectory cluster in the clustered trajectory clusters, all candidate child point POIs included in the first trajectory cluster are used as the initial child point POI; In response to determining that the target parent point POI exists in multiple trajectory clusters in the clustered trajectory cluster, the initial child point POI is determined from all candidate child point POIs included in the multiple trajectory clusters according to the partition information of the target parent point POI.

8. The method according to claim 7, wherein: The determining the initial child point POI from all candidate child point POIs included in the multiple trajectory clusters according to the partition information of the target parent point POI includes: In response to determining that the partition information of the target parent point POI is the non-existent partition, all candidate child point POIs included in the plurality of trajectory clusters are used as the initial child point POI; In response to determining that the partition information of the target parent point POI is the existing partition, a partition child point POI is used as the initial child point POI, wherein the partition child point POI is a child point POI having a parent-child relationship with the partition POI corresponding to the target parent point POI.

9. The method according to claim 8, further comprising: In response to determining that the partition POI corresponding to the target parent point POI appears in a second trajectory cluster among the multiple trajectory clusters, taking all candidate sub-point POIs included in the second trajectory cluster as the partition sub-point POI; In response to determining that the sub-region POI corresponding to the target parent point POI does not appear in the second trajectory cluster among the multiple trajectory clusters, all candidate sub-points POI located within the bounds of the interest surface where the sub-region POI is located are used as the sub-region sub-points POI.

10. The method according to claim 6, wherein: The determining, from the initial child point POI according to the scene information, a target child point POI having a parent-child relationship with the target parent point POI includes: In response to determining that the scene information is the simple scene or the partitioned scene, determining the initial child point POI as a target child point POI having a parent-child relationship with the target parent point POI; In response to determining that the scene information is the overlapping scene or the partitioned overlapping scene, a target child point POI having a parent-child relationship with the target parent point POI is determined from the initial child point POI based on the category of the initial child point POI, the category of the target parent point POI, and the similarity between the name information of the initial child point POI and the name information of the target parent point POI.

11. The method according to claim 1, wherein The clustering of the trajectories passing through the candidate sub-points POI to obtain clustered trajectory clusters includes: Divide the geographic space corresponding to the target parent point POI to obtain at least one grid; For each grid, obtain at least one trajectory that passes through the candidate sub-point POI in the grid, and cluster the at least one trajectory based on the similarity between any two trajectories in the at least one trajectory to obtain a trajectory cluster corresponding to the grid; The trajectory clusters corresponding to the at least one grid are merged to obtain the clustered trajectory cluster.

12. A device for determining a parent-child relationship of a point of interest (POI), comprising: an information determination module configured to determine scene information corresponding to the target parent point POI based on attribute information of the target parent point POI and attribute information of the candidate parent point POI, wherein the scene information is used to indicate whether there is a partition and / or overlap between the target parent point POI and the candidate parent point POI; a clustering module configured to cluster trajectories passing through candidate sub-points (POIs) to obtain clustered trajectory clusters, wherein the candidate sub-points (POIs) are sub-points (POIs) contained in the geographic space corresponding to the target parent point (POI); The child point determination module is configured to determine a target child point POI having a parent-child relationship with the target parent point POI from the candidate child points POI according to the correspondence between the target parent point POI and the cluster trajectory cluster and the scene information corresponding to the target parent point POI.

13. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 11.

14. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are configured to cause the computer to execute the method according to any one of claims 1 to 11.

15. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 11.

Images