A method, device and storage medium for generating space-time domain

Abstract

The present application discloses a method, device and storage medium for generating a spatiotemporal domain, the method comprising: respectively determining the first points of interest corresponding to each point within a first preset range, wherein the point is a location point provided with a collection device; based on the attributes of the first points of interest, initially clustering each of the points to obtain a number of clusters; and according to at least one preset activity mode of the target, performing secondary clustering on the several clusters to obtain a number of spatiotemporal domains corresponding to the first preset range. Through the above methods, the present application can improve the accuracy and practicality of the spatiotemporal domain generation, thereby improving the effectiveness of target clustering based on the spatiotemporal domain.

Description

Technical Field

[0001] The present application relates to the field of big data, and specifically to a spatiotemporal domain generation method, device and storage medium. Background Art

[0002] With the rapid development of information technology, especially the continuous advancement of geographic information systems (GIS), big data analysis, the Internet of Things (IoT) and mobile computing technology, the demand for processing and analyzing spatiotemporal data is growing. The spatiotemporal domain, as a concept that integrates the dimensions of time and space, is crucial for understanding and predicting dynamic events and behavior patterns in the real world.

[0003] However, events and behavior patterns in the real world are changing dynamically, and existing methods find it difficult to capture and adapt to such changes in real time, resulting in reduced accuracy and practicality of generating spatiotemporal domains, and further leading to problems such as deviations in target clustering results based on the spatiotemporal domain. Summary of the invention

[0004] In order to solve the above technical problems, the technical solution adopted in this application is: to provide a spatiotemporal domain generation method, device and storage medium, so as to at least solve the problem that the accuracy and practicality of generating spatiotemporal domains in related technologies are reduced, which further leads to deviations in target clustering results based on the spatiotemporal domain.

[0005] According to an embodiment of the present invention, a spatiotemporal domain generation method is provided, comprising:

[0006] Determine first points of interest corresponding to respective points within a first preset range, wherein the points are locations where collection devices are provided;

[0007] Based on the attributes of the first interest points, initially clustering the points to obtain a number of clusters;

[0008] According to at least one preset activity mode of the target, the plurality of clusters are secondary clustered to obtain a plurality of spatiotemporal domains corresponding to the first preset range, wherein each of the spatiotemporal domains includes at least one of the points.

[0009] In order to solve the above technical problems, a technical solution adopted in the present application is: to provide an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and when the computer program is executed by the processor, it is used to implement the spatiotemporal domain generation method and / or the target clustering method in the above technical solution.

[0010] In order to solve the above technical problems, a technical solution adopted in the present application is: to provide a computer-readable storage medium, which is used to store a computer program. When the computer program is executed by a processor, it is used to implement the spatiotemporal domain generation method and / or the target clustering method in the above technical solution.

[0011] Through the above scheme, the beneficial effect of the present application is: the spatiotemporal domain generation method provided by the present application determines the first point of interest corresponding to each point within the first preset range respectively, and based on the attributes of the first point of interest, initially clusters each point to obtain a number of clusters, and performs secondary clustering on the several clusters according to at least one preset activity mode of the target to obtain a number of spatiotemporal domains corresponding to the first preset range; in this way, through more accurate matching of interest points corresponding to points and secondary clustering of attribute-based clustering methods and activity modes, it is possible to effectively process dynamically changing spatiotemporal data, fuse multi-source information, and consider complex activity patterns to achieve more accurate and practical spatiotemporal domain generation and analysis, thereby improving the accuracy and effectiveness of target clustering. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:

[0013] Figure 1 It is a flowchart of an embodiment of a spatiotemporal domain generation method provided by the present application;

[0014] Figure 2 It is a flowchart of another embodiment of the spatiotemporal domain generation method provided by the present application;

[0015] Figure 3 It is a flowchart of an initial spatiotemporal domain generation method provided by the present application;

[0016] Figure 4 It is a schematic diagram of a time-space domain division provided by this application;

[0017] Figure 5 It is a schematic diagram of a process of adaptively adjusting the spatiotemporal domain provided by the present application;

[0018] Figure 6 It is a structural schematic diagram of an embodiment of an electronic device provided by the present application;

[0019] Figure 7 It is a structural schematic diagram of an embodiment of a computer-readable storage medium provided by the present application. DETAILED DESCRIPTION

[0020] The present application is further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present application, but are not intended to limit the scope of the present application. Similarly, the following examples are only some embodiments of the present application rather than all embodiments, and all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0021] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0022] It should be noted that the terms "first", "second", etc. in this application are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Thus, the features defined as "first", "second", etc. can explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "multiple" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally also includes steps or units that are not listed, or optionally also includes other steps or units inherent to these processes, methods, products or devices.

[0023] The spatiotemporal domain is a concept that combines time and space dimensions to describe and analyze the specific events and geographic locations where events or phenomena occur in the real world. In some dimensions, it can also be understood as the spatial range that can be reached within a certain period of time. Through the analysis of the spatiotemporal domain, the laws of how events and phenomena change over time and space can be revealed. In the actual application of target clustering, the division of the spatiotemporal domain helps to improve the accuracy of the target clustering results.

[0024] Traditional clustering schemes extract features from massive amounts of captured data, and then use the features to cluster similar targets. This type of method has high requirements for feature engineering and clustering algorithms, requires a large amount of comparison calculations, and requires the correct identification and clustering of interference targets in a large range. In order to solve the above problems, a new clustering scheme introduces the concept of spatiotemporal domain, pre-clusters the target data within a certain time and space, increases the effectiveness of calculations, introduces fewer interference targets, and increases the accuracy of clustering.

[0025] However, the current target clustering based on the spatiotemporal domain strongly relies on information in specific dimensions, without comprehensive evaluation and adaptive adjustment, resulting in low efficiency of the clustering system in matching targets and low accuracy of target clustering results.

[0026] In response to the above-mentioned problems, the present application provides a method for generating a spatiotemporal domain, by respectively determining the first point of interest corresponding to each point within a first preset range, initially clustering each point based on the attributes of the first point of interest to obtain a number of clusters, and performing secondary clustering on the number of clusters according to at least one preset activity mode of the target to obtain a number of spatiotemporal domains corresponding to the first preset range; in this way, through more accurate matching of the points of interest corresponding to the points and secondary clustering of the attribute-based clustering method and the activity mode, it is possible to effectively process dynamically changing spatiotemporal data, fuse multi-source information, and consider complex activity patterns, so as to achieve more accurate and real-time spatiotemporal domain generation and analysis, thereby improving the accuracy and effectiveness of target clustering.

[0027] See also Figure 1 , Figure 1 1 is a flow chart of an embodiment of the time-space domain generation method provided by the present application. It should be noted that if there are substantially the same results, this embodiment does not take Figure 1 The process sequence shown is limited. Figure 1 As shown, this embodiment includes:

[0028] S110: Determine first points of interest corresponding to respective points within a first preset range.

[0029] A first preset range where a spatiotemporal domain needs to be generated is predefined, and all or part of the points within the first preset range are acquired, wherein the points are location points where a collection device is provided, such as a camera device or a sensor.

[0030] For each point, determine the first point of interest corresponding to it. A point of interest (POI) is usually a specific location on a map, such as a restaurant, store, park or museum, etc. The first point of interest can be the POI that is geographically closest or functionally most relevant to the point.

[0031] In one embodiment, the location information of each point within a first preset range is obtained, and the first point of interest corresponding to each point is selected from at least one second point of interest within a second preset range of each point. Exemplarily, using the obtained GPS information of each point, a query request is sent to the API of the map service provider to retrieve the point of interest data within a certain range near each point, that is, to obtain at least one second point of interest within the second preset range of each point, and filter out the POI information related to the location of each point from the query results, such as based on multiple factors such as distance, category or preference; perform a correlation evaluation on the filtered POI, wherein the correlation may include the distance to the corresponding point, the frequency of visits, the rating and historical data, etc.; based on the correlation evaluation results of each point of interest, the second point of interest with the highest correlation among the at least one second point of interest corresponding to each point is selected as the first point of interest.

[0032] S120: Based on the attributes of the first interest point, each point is initially clustered to obtain a plurality of clusters.

[0033] The attribute information of each first point of interest is obtained, which may include naming information, category, service type, score, etc. Based on the attributes of the first points of interest, each point corresponding to each first point of interest is initially clustered to obtain a plurality of clusters.

[0034] In one embodiment, based on the attribute information of the first interest point corresponding to each point, the attribute similarity between different first interest points is determined, and the points corresponding to the first interest points whose attribute similarities meet the similarity requirements are divided into the same cluster.

[0035] In one example, the attribute information includes naming information, and semantic analysis is performed on the attribute information of each first point of interest to obtain the attribute similarity between different first points of interest. Exemplarily, semantic analysis in natural language processing (NLP) is performed on the attributes such as the naming information of the first point of interest to understand the deep meaning of the attribute information; by comparing the attribute information of different first points of interest, the similarity between different first points of interest is calculated, and semantically similar first points of interest are found by establishing a bag-of-words model, a TF-IDF model, and an LSI model, and then the first points of interest that meet the similarity requirements or the points corresponding to each first point of interest are initially clustered to obtain a number of clusters.

[0036] S130: Perform secondary clustering on a plurality of clusters according to at least one preset activity mode of the target to obtain a plurality of spatiotemporal domains corresponding to the first preset range.

[0037] According to the needs of the actual application scenario, at least one preset activity mode of the target is preset, such as walking, cycling, driving, etc. These activity modes will affect the spatial range that the target can cover within a specific time. Based on the preset at least one activity mode, the range of influence of at least one activity mode on a certain space within a certain time is determined, and secondary clustering is performed in combination with several clusters obtained by initial clustering to obtain several spatiotemporal domains corresponding to the first preset range.

[0038] In one embodiment, at least one preset activity mode includes a first activity mode, and the maximum activity distance corresponding to the first activity mode is obtained as the first activity distance. For each cluster, at least two points within the first activity distance under the cluster are divided into the same first spatiotemporal domain to obtain at least one first spatiotemporal domain corresponding to the cluster. It should be noted that the distance between at least two points is less than the first activity distance, wherein the distance between at least two points can be calculated based on the position information of the points or based on the position information of the first point of interest corresponding to the two points. The difference between the two is calculated.

[0039] In one embodiment, at least one preset activity mode further includes at least one second activity mode, each second activity mode corresponds to a different maximum activity distance, and each second activity mode corresponds to a maximum activity distance greater than the first activity distance. According to at least two activity modes of the target, a plurality of clusters are secondary clustered to obtain the spatiotemporal domains of each target corresponding to the first preset range.

[0040] Exemplarily, for each second activity mode, the maximum activity distance corresponding to the second activity mode is obtained as the second activity distance, wherein the second activity distance can be used as a criterion for evaluating whether two spatiotemporal domains can be merged. For each target spatiotemporal domain, the first spatiotemporal domain whose distance from the target spatiotemporal domain is within the second activity distance is used as the spatiotemporal domain to be merged corresponding to the second activity mode of the target spatiotemporal domain. The target spatiotemporal domain and the spatiotemporal domain to be merged corresponding to the second activity mode of the target spatiotemporal domain are merged to obtain the second spatiotemporal domain corresponding to the second activity mode. The merging process is based on the maximum activity distance of the second activity mode to ensure that the merged second spatiotemporal domain can reflect the area reachable under the activity mode. Among them, the target spatiotemporal domain is part or all of the first spatiotemporal domain, and the distance between the two spatiotemporal domains is the distance between the domain representative position points of the two spatiotemporal domains, and the domain representative position point can be the center point or representative point of each spatiotemporal domain.

[0041] In one embodiment, the first activity mode is walking, and the second activity mode includes at least one of using a non-motorized vehicle and using a motorized vehicle. The moving speed corresponding to the preset activity mode and the distance that the target moves in a preset time according to the moving speed corresponding to the preset activity mode are obtained as the maximum activity distance corresponding to the preset activity mode. For each target spatiotemporal domain, a plurality of first spatiotemporal domains are obtained based on the maximum activity distance calculated based on different activity modes, and a second spatiotemporal domain is obtained based on the combination of the first spatiotemporal domains, and finally a plurality of spatiotemporal domains within the first preset range are determined.

[0042] See also Figure 2 , Figure 2 FIG. 1 is a flow chart of another embodiment of the time-space domain generation method provided by the present application. It should be noted that if there are substantially the same results, this embodiment is not used. Figure 2 The process sequence shown is limited. Figure 2 As shown, this embodiment includes:

[0043] S201: Perform initial clustering based on first interest points corresponding to each point within a first preset range to obtain a plurality of clusters.

[0044] All or part of the target points within the first preset range are obtained, first interest points corresponding to each point are determined, and each point corresponding to each first interest point is initially clustered based on the first interest point corresponding to each point to obtain a plurality of clusters.

[0045] The above method for confirming the first point of interest and the method for initial clustering may refer to the above steps S110 - S120 , which will not be described in detail herein.

[0046] S202: performing secondary clustering on a plurality of clusters according to at least one preset activity mode of the target to obtain an initial spatiotemporal domain.

[0047] According to the needs of the actual application scenario, at least one preset activity mode of the target is set, and the influence range of each activity mode on the space within a certain period of time is calculated according to the properties and characteristics of the preset activity mode. Based on the determined influence range, several clusters are secondary clustered to obtain the initial spatiotemporal domain, so as to facilitate the subsequent real-time adjustment of the initial spatiotemporal domain according to the real target activity information. The initial spatiotemporal domain includes several first spatiotemporal domains and / or at least one second spatiotemporal domain.

[0048] The method of performing secondary clustering to obtain the initial spatiotemporal domain may refer to the above step S130 and will not be described in detail here.

[0049] S203: Adjusting the initial spatiotemporal domain based on target trajectory information of each point and / or the amount of collected data of each point.

[0050] In order to further improve the accuracy and practicality of the spatiotemporal domain, by continuously collecting data from point locations and analyzing the activity patterns of targets, the spatiotemporal domain is automatically updated and optimized to adapt to changes in the real world, reduce the need for manual intervention, and reduce operation and maintenance costs.

[0051] In one embodiment, based on the target trajectory information obtained by analyzing the collected data of each point, the points in the first spatiotemporal domain are merged or split to better reflect the actual activity mode and trajectory path. Based on the target trajectory information of each point, it is detected whether there is a preset association relationship between at least two points distributed in different first spatiotemporal domains, and the preset association relationship indicates that when the target passes through one point, there is a certain probability that it will also pass through another point.

[0052] If a preset association relationship is detected between at least two points distributed in different first spatiotemporal domains, a new first spatiotemporal domain containing at least two points is regenerated, wherein the points in the new first spatiotemporal domain are newly merged points. Other points in the first spatiotemporal domain where each newly merged point originally resides are used to re-form other new first spatiotemporal domains.

[0053] In this embodiment, by analyzing the target trajectory information, the relationship between the points can be more accurately identified and adjusted, thereby improving the accuracy of the spatiotemporal domain division. At the same time, since the target trajectory information is real-time, the spatiotemporal domain generation method provided in this application can dynamically adjust the spatiotemporal domain according to real-time data to better reflect and adapt to actual activity patterns and needs.

[0054] In one embodiment, the method for adaptively adjusting the spatiotemporal domain further includes: adjusting the domain representative position point of the first spatiotemporal domain based on the amount of collected data of each point, and re-determining the points included in the first spatiotemporal domain based on the new domain representative position point. The domain representative position point is the center or representative position of each spatiotemporal domain, and the point with the largest amount of collected data can be selected as the new domain representative position point.

[0055] Exemplarily, a point within a preset distance range of a domain representative position point of the first spatiotemporal domain is determined as a candidate point, and a candidate point with the largest amount of collected data is selected from a number of candidate points as a new domain representative position point of the first spatiotemporal domain, and the points included in the first spatiotemporal domain are re-determined based on the new domain representative position point. In the process of re-determining the first spatiotemporal domain, it may involve merging or splitting the points of the original first spatiotemporal domain.

[0056] After merging or splitting the points in the first spatiotemporal domain based on the target trajectory information of each point, or adjusting the domain representative position point of the first spatiotemporal domain based on the amount of collected data of each point, it also includes: after adjusting the first spatiotemporal domain, re-determining the second spatiotemporal domain corresponding to each second activity mode based on the adjusted first spatiotemporal domain. The second activity mode refers to other activity modes other than the main activity mode (such as the first activity mode), such as riding or driving other than walking.

[0057] It should be noted that the spatiotemporal domain adaptive adjustment process described in the above embodiment is a dynamic update process. With the continuous collection and analysis of new data, the definition of the spatiotemporal domain will also be continuously adjusted and optimized.

[0058] To better illustrate the spatiotemporal domain generation method of the present application, a spatiotemporal domain generation method in a target clustering application scenario is provided as an exemplary illustration. The spatiotemporal domain generation method in this embodiment can be divided into two steps, namely, generating an initialization spatiotemporal domain based on the physical information of the initial point and the preset target activity speed, and adaptively adjusting the spatiotemporal domain according to the target's trajectory information after the clustering system has run for a period of time.

[0059] First see Figure 3 , Figure 3 This is a flow chart of an initial spatiotemporal domain generation method provided by this application. The physical information of the initial points is used to search for points of interest, semantic processing is performed based on the naming information of the points of interest to pre-aggregate highly similar points, and finally the pre-clustered joint clustering is performed based on the preset target activity speed to generate the initialization spatiotemporal domain. The specific process is as follows:

[0060] (1) Load the GPS information of existing points in the target clustering system.

[0061] (2) Obtain the corresponding representative points of interest based on the GPS information of the point.

[0062] According to the GPS information of the point, the POI points of interest within a certain range of each point are searched, and the POI points of interest with the highest correlation with each point are selected according to the relevance as the representative points of interest of each point.

[0063] (3) Semantic analysis of POI representative points of interest is used to pre-cluster similar points.

[0064] Perform NLP semantic processing and analysis on each POI representative point of interest, establish a bag-of-words model, a TF-IDF model, and an LSI model, and pre-cluster highly similar POI representative points of interest or the points corresponding to each POI representative point of interest.

[0065] (4) Cluster and merge different cluster areas according to the speed of the target's different preset activity modes.

[0066] In this embodiment, the preset activity modes of the target include walking, non-motorized vehicles and motorized vehicles, and different preset activity modes correspond to different speeds. The above pre-clustering results are divided based on the reachable range of the target calculated based on the walking speed to obtain several cluster areas.

[0067] Determine the center point of each cluster area as the representative point of each cluster area, calculate the reachable range of the representative point of each cluster area based on the speed of different target activity modes (such as at least one of non-motor vehicles and / or motor vehicles) within a preset period of time, and perform secondary clustering of other cluster areas that meet the reachability with each cluster area as the center to obtain the initial spatiotemporal domain. For example, refer to Figure 4 , Figure 4 This is a schematic diagram of the time-space domain division provided by the present application. The black dots in the figure represent the positions of various points. Secondary clustering is performed according to the different activity modes of the targets to obtain mutually reachable activity areas.

[0068] See further Figure 5 , Figure 5 This is a schematic diagram of a process of adaptively adjusting the spatiotemporal domain provided by the present application. In order to adjust the point information and more scientifically fit the activity law of the real target in the scene, the spatiotemporal domain generation program of this embodiment also has an adaptive adjustment function.

[0069] After the target clustering system has been running for a period of time, the regular information of the correlation relationship between the points is mined based on the daily data collection volume of each point and the trajectory information after clustering, so as to dynamically adjust the time and space domain. In this way, by regularly reading the data information of the clustering system and intelligently responding to changes in point information, when the user updates the point information, the division of the time and space domain can be automatically adjusted and optimized without human intervention. This feature not only reduces the additional operation and maintenance work caused by changes in point information, but also ensures the accuracy and timeliness of the time and space domain division, reduces maintenance costs, and improves the flexibility and efficiency of the system. The specific process of adaptively updating the time and space domain process is as follows:

[0070] (1) Regularly load the latest point information of the target clustering system.

[0071] (2) Obtain daily data on points within the clustering system and target trajectory information within the system.

[0072] (3) Mining the correlation between points based on the correlation rules of target trajectory information within the system.

[0073] Exemplarily, the FP-Growth algorithm is used to mine the path relationship between different points of the target. By treating points as items and paths as transactions, the FP-Growth algorithm can efficiently identify frequent point combinations and reveal the pattern of target movement. In this way, points that often appear together in the target path can be found, so as to determine the correlation between the points. For example, if the target passes through point A and then passes through point B with a high probability, then we can determine that there is a significant correlation between point A and point B. This process provides data support for the dynamic adjustment of the spatiotemporal domain, allowing us to more accurately reflect the activity patterns of the target.

[0074] (4) Optimize and adjust the time and space domain based on the correlation between points.

[0075] By analyzing the correlation between points, we dynamically optimize the time and space domain. For points with strong correlation, we merge them to form a more representative time and space domain; for points with poor correlation, we split them to improve the accuracy of the time and space domain.

[0076] For points with strong correlation, that is, those points that often appear together in the target path, re-merging them into a space-time domain helps to more accurately reflect the actual activity range and flow pattern of the target.

[0077] On the contrary, for those points with weak correlation, that is, points that rarely appear together in the target path, splitting processing is performed to divide them into different spatiotemporal domains. This helps to reduce false associations and improve the homogeneity of points in the spatiotemporal domain.

[0078] (5) Reselect representative points to update the reachability calculation and adjust the current space-time domain.

[0079] In each optimized spatiotemporal domain, the points with the largest amount of collected data are selected as new representative points in each cluster area, and the accessibility between spatiotemporal domains is re-evaluated based on these points. This process ensures that the division of spatiotemporal domains can keep up with the actual pattern of target activities and provide a solid data foundation for further analysis and decision-making.

[0080] For example, in each cluster area after clustering, the point information with the largest amount of collected data is determined within a certain preset range of the center point of the current cluster area, and the point is used as the new representative point. These points are usually the most active or important locations in the spatiotemporal domain, which can best represent the characteristics of the entire spatiotemporal domain. The newly selected representative points are used to recalculate the accessibility between the spatiotemporal domains, and then adjust the current spatiotemporal domain.

[0081] Based on the several spatiotemporal domains determined within a preset range by the above-mentioned spatiotemporal domain generation method, the present application also provides a target clustering method, clustering the target information corresponding to each spatiotemporal domain to obtain a target clustering result, wherein the target information corresponding to the spatiotemporal domain is obtained based on the collected data of the points in the spatiotemporal domain. The spatiotemporal domain obtained by the above-mentioned spatiotemporal domain generation and adjustment and update method can effectively assist in the efficient comparison in the target clustering process and improve the accuracy of clustering.

[0082] See also Figure 6 , Figure 6 It is a structural diagram of an embodiment of an electronic device provided in the present application. The electronic device 60 includes a memory 61 and a processor 62 that are connected to each other. The memory 61 is used to store a computer program. When the computer program is executed by the processor 62, it is used to implement the spatiotemporal domain generation method and / or the target clustering method in the above-mentioned embodiment.

[0083] The method of the above embodiment may exist in the form of a computer program, so the present application proposes a computer readable storage medium, see Figure 7 , Figure 7 It is a structural diagram of an embodiment of a computer-readable storage medium provided in the present application. The computer-readable storage medium 80 is used to store a computer program 81, which can be executed to implement the spatiotemporal domain generation method and / or the target clustering method in the above-mentioned embodiments.

[0084] The computer-readable storage medium 80 may be a server, a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., which may store program codes.

[0085] The above description is only an embodiment of the present application and does not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for generating space-time domain, characterized in that: The method comprises: Determine first points of interest corresponding to respective points within a first preset range, wherein the points are locations where collection devices are provided; Based on the attributes of the first interest points, initially clustering the points to obtain a number of clusters; According to at least two preset activity modes of the target, the maximum activity distance corresponding to each of the preset activity modes is obtained, and the plurality of clusters are secondary clustered based on the maximum activity distance corresponding to each of the preset activity modes to obtain a plurality of space-time domains corresponding to the first preset range, wherein the at least two preset activity modes include a first activity mode and at least one second activity mode, and each of the space-time domains includes at least one of the points.

2. The method according to claim 1, characterized in that The step of respectively determining the first points of interest corresponding to the points within the first preset range includes: For each of the points, based on the position information of the point, a first point of interest corresponding to the point is selected from at least one second point of interest within a second preset range of the point.

3. The method according to claim 1, characterized in that Based on the attributes of the first interest points, each of the points is initially clustered to obtain a number of clusters, including: Determining attribute similarities between different first interest points based on the attribute information of each of the first interest points; The points corresponding to the first interest points whose attribute similarity meets the similarity requirement are divided into the same cluster.

4. The method according to claim 3, characterized in that: The attribute information includes naming information; And / or, determining the attribute similarity between different first interest points based on the attribute information of each first interest point includes: Semantic analysis is performed on the attribute information of each of the first interest points to obtain attribute similarities between different first interest points.

5. The method according to claim 1, characterized in that The at least two preset activity modes include a first activity mode; The method of obtaining the maximum activity distance corresponding to each of the preset activity modes according to at least two preset activity modes of the target, and performing secondary clustering on the plurality of clusters based on the maximum activity distance corresponding to each of the preset activity modes to obtain a plurality of spatiotemporal domains corresponding to the first preset range includes: Obtaining a maximum activity distance corresponding to the first activity mode as a first activity distance; For each of the clusters, at least two of the points within the first activity distance under the cluster are divided into the same first spatiotemporal domain to obtain at least one first spatiotemporal domain corresponding to the cluster.

6. The method according to claim 5, characterized in that The at least two preset activity modes further include at least one second activity mode, the maximum activity distances corresponding to the second activity modes are different, and the maximum activity distances corresponding to the second activity modes are greater than the first activity distance; The method further comprises: obtaining a maximum activity distance corresponding to each of the preset activity modes according to at least two preset activity modes of the target, performing secondary clustering on the plurality of clusters based on the maximum activity distance corresponding to each of the preset activity modes to obtain a plurality of spatiotemporal domains corresponding to the first preset range; and For each of the second activity modes; Obtaining a maximum activity distance corresponding to the second activity mode as a second activity distance; For each target spatiotemporal domain, the first spatiotemporal domain whose distance to the target spatiotemporal domain is within the second activity distance is used as the spatiotemporal domain to be merged with the target spatiotemporal domain corresponding to the second activity mode, wherein the target spatiotemporal domain is part or all of the first spatiotemporal domain, and the distance between the two spatiotemporal domains is the distance between the domain representative position points of the two spatiotemporal domains; The target spatiotemporal domain and the spatiotemporal domain to be merged corresponding to the target spatiotemporal domain and the second activity mode are merged to obtain a second spatiotemporal domain corresponding to the second activity mode.

7. The method according to claim 6, characterized in that The first activity mode is walking, and the second activity mode includes at least one of using a non-motorized vehicle and using a motorized vehicle; And / or, obtaining the maximum activity distance corresponding to the preset activity mode, including: Obtaining a moving speed corresponding to the preset activity mode; The distance that the target moves in a preset time at a moving speed corresponding to the preset activity mode is obtained as the maximum activity distance corresponding to the preset activity mode.

8. The method according to claim 5, characterized in that After dividing the at least two points within the first activity distance under the cluster into the same first spatiotemporal domain to obtain at least one first spatiotemporal domain corresponding to the cluster, the method further includes at least one of the following steps: Based on the target trajectory information of each of the points, merging or splitting the points in the first time-space domain, wherein the target trajectory information is obtained based on the analysis of the collected data of the points; Based on the amount of collected data of each of the points, the domain representative position point of the first spatiotemporal domain is adjusted, and based on the new domain representative position point, the points included in the first spatiotemporal domain are re-determined.

9. The method according to claim 8, characterized in that The merging or splitting the points in the first spatiotemporal domain based on the target trajectory information of each of the points includes: Based on the target trajectory information, it is detected that there is a preset association relationship between at least two points distributed in different first time-space domains, wherein the preset association relationship indicates that when the target passes through one point, there is a certain probability that the target also passes through another point; Regenerate a new first spatiotemporal domain that at least includes the at least two points, wherein the points in the new first spatiotemporal domain are newly merged points; Using other points in the first spatiotemporal domain where each of the newly merged points originally existed, re-forming other new first spatiotemporal domains; The adjusting the domain representative position point of the first spatiotemporal domain based on the amount of collected data at each of the points includes: Selecting the candidate point with the largest amount of collected data from a number of candidate points as a new domain representative position point of the first spatiotemporal domain, wherein the candidate point is a point within a preset distance range of the domain representative position point of the first spatiotemporal domain; The at least one preset activity mode also includes at least one second activity mode, after merging or splitting the points in the first spatiotemporal domain based on the target trajectory information of each of the points, or adjusting the domain representative position point of the first spatiotemporal domain based on the amount of collected data of each of the points, further including: Based on the adjusted first spatiotemporal domain, the second spatiotemporal domain corresponding to each of the second activity modes is re-determined.

10. A target clustering method, characterized in that: include: Determine a plurality of spatiotemporal domains within a preset range, wherein the plurality of spatiotemporal domains are obtained by using the method according to any one of claims 1 to 9; The target information corresponding to each of the time and space domains is clustered to obtain a target clustering result, wherein the target information corresponding to the time and space domain is obtained based on the collected data of the points in the time and space domain.

11. An electronic device, characterized in that: The electronic device includes a processor and a memory, wherein the processor is coupled to the memory, and the processor is configured to execute one or more steps of the spatiotemporal domain generation method described in any one of claims 1 to 9, and / or one or more steps of the target clustering method described in claim 10 based on instructions stored in the memory.

12. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the spatiotemporal domain generation method according to any one of claims 1 to 9, and / or one or more steps of the target clustering method according to claim 10.

Images