Population flow monitoring method and system based on spatiotemporal gridded agent graph

Abstract

The application provides a population flow monitoring method and system based on a space-time gridding intelligent agent graph, first constructs a space-time grid composed of units with unique space-time codes based on preset rules, and accesses multi-source positioning data to map individual positions to corresponding units in real time, and then dynamically updates the intelligent agent graph according to individual movement; subsequently, the real-time index flow is calculated based on the graph, and the index threshold group is obtained in combination with the historical index baseline library and the typical event graph library; then, if the unit residence degree exceeds the threshold, primary attention is determined; in response, if the flow flux exceeds the threshold, intermediate early warning is determined; in response, if the graph connection degree deviates from or is similar to the typical graph by more than the threshold, high-level abnormality is determined; finally, a graded early warning signal is output according to the determination result; the application provides intelligent decision support for urban fine management and emergency response through the deep integration of space-time grids and intelligent agent graphs.

Description

Technical Field

[0001] This invention relates to the field of population movement monitoring, and more specifically, to a method and system for population movement monitoring based on a spatiotemporal gridded intelligent agent map. Background Technology

[0002] With the acceleration of urbanization and the popularization of IoT sensing technology, massive amounts of multi-source location data provide a rich data foundation for population flow monitoring. However, existing population flow monitoring technologies have significant shortcomings. First, traditional methods often use fixed grid divisions, focusing only on the quantitative changes of individuals within a single spatial unit, failing to deeply couple spatial location with the temporal dimension, and making it difficult to capture the dynamic evolutionary characteristics of individuals over time. Second, existing technologies typically treat each moving individual as an isolated point data point, analyzing only their physical location trajectory, ignoring the social relationships between individuals and the structural information contained in group aggregation. Third, conventional anomaly detection methods often use single or static thresholds for judgment, which easily generate a large number of false alarms when facing periodic fluctuations such as morning and evening rush hours and holidays. Fourth, the judgment logic of current methods usually only makes independent judgments on a single indicator, lacking a multi-level progressive analysis mechanism from individual aggregation and spatial flow to changes in group structure. Finally, due to the lack of a refined hierarchical early warning mechanism, the output results are often rather general.

[0003] Therefore, there is an urgent need for a population flow monitoring technology that can deeply integrate spatiotemporal information with individual social relationships and has intelligent discrimination capabilities. Summary of the Invention

[0004] In view of the above problems, the purpose of this invention is to provide a population flow monitoring method and system based on a spatiotemporal gridded intelligent agent graph. By constructing a spatiotemporal grid and generating an intelligent agent graph, and introducing a chain threshold comparison mechanism, a multi-level progressive anomaly judgment from individual flow to group structure is achieved. Specifically, firstly, by performing multi-level recursive segmentation of geographic space and binding it with time slices, standardized spatiotemporal units with unique codes are generated. Then, by constructing dynamic attribute profiles for each individual and mining social relationship edges between individuals, an intelligent agent graph containing rich semantic information is generated. Secondly, by aggregating and calculating unit dwell time, flow flux, and graph connectivity, a three-dimensional indicator flow covering aggregation, movement, and convergence is constructed. Then, by introducing a historical indicator baseline library and a typical event graph library, adaptive threshold generation matching the spatiotemporal context is achieved. Finally, three-level early warning signals are output through different results of chain comparisons, providing accurate hierarchical decision-making basis for refined urban management and emergency response.

[0005] The first aspect of this invention provides a population flow monitoring method based on a spatiotemporal rasterized intelligent agent map, the method comprising: Based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; Acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; The agent graph corresponding to an individual is dynamically updated based on the changes in the spatiotemporal units of that individual. Based on the agent graph, an indicator flow is calculated, which includes cell dwell degree, flow flux and graph connectivity. Based on the historical indicator baseline library and the typical event graph library, the indicator threshold group is obtained; If the cell dwell time exceeds the dwell time threshold, it is judged as primary concern; In response to initial concerns, if the flow flux exceeds the flux threshold, it is judged as an intermediate warning; In response to the intermediate warning, based on the connectivity of the graph and the typical event graph library, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly.

[0006] Based on the judgment results, a graded early warning signal is output.

[0007] In this scheme, the spatiotemporal grid composed of multiple spatiotemporal units, obtained based on preset geospatial division and time slicing rules, specifically includes: According to the preset multi-level geospatial division rules, the target monitoring area is recursively divided into multiple spatial grid units with unique codes and no overlap. According to the preset time discretization rules, the continuous time axis is divided into multiple continuous and equal-length time slices; Each of the spatial grid cells is combined and bound with each of the time slices to generate a spatiotemporal cell with a unique spatiotemporal code; The spatiotemporal grid is composed of all the aforementioned spatiotemporal units.

[0008] In this scheme, the step of dynamically updating the agent graph corresponding to an individual based on changes in the spatiotemporal units of that individual specifically includes: Based on each individual's real-time location, their identity identifier is mapped to the corresponding spatiotemporal unit, establishing a relationship between the individual and the spatiotemporal unit; Acquire multi-source attribute data streams and construct dynamic attribute profiles for each individual. The attribute profiles include at least basic attributes, behavioral pattern attributes, and social relationship attributes. Based on the co-occurrence frequency and duration of multiple individuals in the same spatiotemporal unit, the social relationship edges between individuals are dynamically constructed and updated; Using individuals as nodes, social relationships as edges, and attribute profiles as node attributes, a dynamically evolving intelligent agent graph is generated.

[0009] In this scheme, the step of calculating the indicator flow based on the agent graph specifically includes: Aggregate the residence information of all individuals within a single spatiotemporal unit to calculate the unit residence degree, which reflects the degree of aggregation of individuals within the spatiotemporal unit; For spatial grid cells on adjacent time slices, the transition matrix of an individual between different cells is analyzed, and the flow flux reflecting the intensity of individual movement between spatiotemporal cells is calculated. For a monitoring area consisting of multiple spatiotemporal units, the agent graph of all individuals in the area is extracted as a subgraph. The network structure features of the subgraph are calculated to obtain the graph connectivity, which reflects the degree of social aggregation of the group.

[0010] In this scheme, the indicator threshold group obtained based on the historical indicator baseline library and the typical event graph library includes: Obtain a historical baseline database, which stores the residence and flux characteristics of each spatiotemporal unit in the same historical period; The dwell time threshold is calculated based on the dwell time characteristics of adjacent time slices; The flux threshold is calculated based on the flux characteristics of adjacent spatial grid cells; Obtain a typical event graph library, which stores the agent sub-graph structure features of the corresponding region when a known event occurs; Based on the preset feature tolerance, the deviation threshold and similarity threshold of the map are obtained.

[0011] In this scheme, based on the judgment result, a graded early warning signal is output, specifically including: If it is determined to be only of primary concern, then a first-level observation warning signal is generated and output. If the warning is determined to be of medium level, a second-level prompt warning signal will be generated and output. If it is determined to be a high-level anomaly, a third-level alarm warning signal will be generated and output. The warning signal includes at least the area where the event occurred, the scope of impact, and the event level information.

[0012] A second aspect of the present invention provides a population flow monitoring system based on a spatiotemporal rasterized intelligent agent map, including a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map. When the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map is executed by the processor, it implements the following steps: Based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; Acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; The agent graph corresponding to an individual is dynamically updated based on the changes in the spatiotemporal units of that individual. Based on the agent graph, an indicator flow is calculated, which includes cell dwell degree, flow flux and graph connectivity. Based on the historical indicator baseline library and the typical event graph library, the indicator threshold group is obtained; If the cell dwell time exceeds the dwell time threshold, it is judged as primary concern; In response to initial concerns, if the flow flux exceeds the flux threshold, it is judged as an intermediate warning; In response to the intermediate warning, based on the connectivity of the graph and the typical event graph library, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly.

[0013] Based on the judgment results, a graded early warning signal is output.

[0014] A third aspect of the present invention provides a computer-readable storage medium comprising a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map, wherein when the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map is executed by a processor, the steps of the population flow monitoring method based on the spatiotemporal rasterized intelligent agent map as described in any of the preceding claims are implemented.

[0015] This invention provides a method and system for monitoring population flow based on a spatiotemporal gridded intelligent agent map. First, a spatiotemporal grid composed of uniquely spatiotemporally coded units is constructed based on preset rules. Multi-source positioning data is then integrated to map individual locations to corresponding units in real time, and the intelligent agent map is dynamically updated based on individual movement. Subsequently, a real-time indicator flow is calculated based on this map, and indicator threshold groups are obtained by combining a historical indicator baseline library and a typical event map library. Then, if the unit dwell rate exceeds the threshold, a primary concern is identified. In response, if the flow throughput exceeds the threshold, a medium-level warning is issued. In response, if the map connectivity deviates from or is similar to a typical map beyond the threshold, a high-level anomaly is identified. Finally, a tiered warning signal is output based on the judgment results. This invention, through the deep integration of spatiotemporal grids and intelligent agent maps, provides intelligent decision support for refined urban management and emergency response. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope.

[0017] Figure 1 A flowchart of a population flow monitoring method based on a spatiotemporal rasterized intelligent agent map according to the present invention is shown; Figure 2 A flowchart illustrating the partitioning process of a spatiotemporal grid according to an embodiment of the present invention is shown; Figure 3 A flowchart illustrating the drawing process of an intelligent agent map according to an embodiment of the present invention is shown; Figure 4 A block diagram of a population flow monitoring system based on a spatiotemporal rasterized intelligent agent map according to the present invention is shown. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Unless otherwise defined, all terms (including technical and scientific terms) used in embodiments of this invention shall have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in a common dictionary shall be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as being interpreted in an idealized or highly formalized sense, unless expressly defined in this embodiment of the invention.

[0020] The terms "first," "second," and similar words used in the embodiments of this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "an," "a," or "the" do not indicate a quantity limitation, but rather indicate the presence of at least one. Similarly, terms such as "including" or "comprising" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The steps preceding or following the steps in the method of the embodiments of this invention are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0021] In addition, the functional modules in the various embodiments of the present invention can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0022] Figure 1 The flowchart of a population flow monitoring method based on a spatiotemporal rasterized intelligent agent map according to the present invention is shown.

[0023] like Figure 1 As shown, the first aspect of this invention discloses a population flow monitoring method based on a spatiotemporal rasterized intelligent agent map, the method comprising: S102, based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; S104, acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; S106, dynamically update the agent graph corresponding to the individual based on the changes in the spatiotemporal unit of the individual; S108, Calculate the index flow based on the agent graph, the index flow including cell dwell degree, flow flux and graph connectivity; S110, based on the historical indicator baseline library and the typical event graph library, obtain the indicator threshold group; S112, If the cell dwell time exceeds the dwell time threshold, it is determined to be a primary concern; S114, in response to primary concern, if the flow flux exceeds the flux threshold, it is judged as a medium warning; S116, in response to intermediate warning, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly based on the graph connectivity and typical event graph library.

[0024] S118, based on the judgment result, outputs a graded early warning signal.

[0025] It should be noted that, in this embodiment, the monitoring system first discretizes the target monitoring area into a spatiotemporal grid composed of multiple spatiotemporal units with unique codes, based on preset geospatial division rules and time slicing rules. Subsequently, the monitoring system receives multi-source positioning data streams from different terminals in real time, identifies each mobile individual, and maps it to the corresponding spatiotemporal unit according to its current coordinates. As individuals move between different spatiotemporal units, the monitoring system continuously tracks their positional change trajectory and dynamically updates the agent graph corresponding to each individual. This graph not only records the individual's location history but also contains their behavioral patterns and the relationships between individuals. Based on this, the monitoring system periodically calculates real-time indicator streams to characterize population flow status using the rich information contained in the agent graphs. These indicator streams include at least unit dwell time (reflecting the degree of individual aggregation), flow flux (reflecting the intensity of spatial movement), and graph connectivity (reflecting the degree of social aggregation of the group). Simultaneously, the monitoring system has a pre-set historical indicator baseline library and a typical event map library. By analyzing the deviation of current indicators from historical baselines and typical maps, a set of indicator thresholds for anomaly judgment is obtained. The monitoring system performs a chain comparison judgment: when the cell dwell degree of a certain spatiotemporal unit exceeds the corresponding dwell degree threshold, it is judged as a primary concern; in response to the primary concern, if the flow flux in the area further exceeds the flux threshold, it is judged as upgraded to an intermediate warning; in response to the intermediate warning, the monitoring system further analyzes the map connectivity of the area. If its deviation from the typical event map exceeds the deviation threshold or its similarity exceeds the similarity threshold, it is finally judged as a high-level anomaly. Finally, based on the above judgment results, the monitoring system outputs the corresponding level of graded warning signal to the management terminal.

[0026] Figure 2 A flowchart illustrating the partitioning process of a spatiotemporal grid provided by an embodiment of the present invention is shown.

[0027] According to embodiments of the present invention, such as Figure 2 As shown, the spatiotemporal grid, composed of multiple spatiotemporal units, is obtained based on preset geospatial division and time slicing rules, specifically including: S202, according to the preset multi-level geospatial division rules, recursively divides the target monitoring area into multiple spatial grid units with unique codes and no overlap. S204, according to the preset time discretization rules, divide the continuous time axis into multiple continuous and equal-length time slices; S206, combine and bind each of the spatial grid units with each of the time slices to generate a spatiotemporal unit with a unique spatiotemporal code; S208, the spatiotemporal grid is composed of all the aforementioned spatiotemporal units.

[0028] It should be noted that, in this embodiment, the monitoring system first recursively segments the target monitoring area according to preset multi-level geospatial division rules. For example, it generates a series of spatial raster units with unique codes and non-overlapping boundaries according to administrative regions, key functional areas, or standard geographic grids, forming a spatially separated structure. Simultaneously, the monitoring system uniformly divides the continuous time axis into multiple continuous and equal-length time slices according to preset time discretization rules, for example, slicing at the minute or second level. Subsequently, the monitoring system binds each spatial raster unit to each time slice through combination operations, generating a spatiotemporal unit with a unique spatiotemporal code. This spatiotemporal unit is both the basic container for data storage and the smallest unit for all subsequent analysis and calculations. Finally, the three-dimensional spatiotemporal raster composed of all spatiotemporal units provides a unified spatiotemporal coordinate system for subsequent population flow monitoring.

[0029] Figure 3 The diagram illustrates a flowchart of an intelligent agent map drawing method provided by an embodiment of the present invention.

[0030] According to embodiments of the present invention, such as Figure 3 As shown, the step of dynamically updating the agent graph corresponding to an individual based on changes in the individual's spatiotemporal units specifically includes: S302, based on the real-time location of each individual, map their identity identifier to the corresponding spatiotemporal unit, and establish the association between the individual and the spatiotemporal unit; S304, acquire multi-source attribute data streams, and construct a dynamic attribute profile for each individual. The attribute profile includes at least basic attributes, behavioral pattern attributes, and social relationship attributes. S306: Based on the co-occurrence frequency and duration of multiple individuals in the same spatiotemporal unit, the social relationship edges between individuals are dynamically constructed and updated. S308 uses individuals as nodes, social relationships as edges, and attribute profiles as node attributes to generate a dynamically evolving intelligent agent graph.

[0031] It should be noted that, in this embodiment, the monitoring system first accurately maps the anonymized identity of each individual to the corresponding spatiotemporal unit based on the received real-time location information, thereby establishing a dynamic association between the individual and the specific spatiotemporal location. Simultaneously, with the individual's permission, the monitoring system also accesses multi-source attribute data streams, such as the individual's basic registration information, travel preferences, and consumption records, to construct a dynamic attribute profile for each individual that evolves over time. This profile covers at least basic attributes, behavioral pattern attributes, and social relationship attributes. Furthermore, by analyzing the historical co-occurrence frequency and duration of a large number of individuals within the same spatiotemporal unit, the monitoring system mines and infers potential social connections between individuals based on a pre-trained social relationship model, such as colleague, neighbor, or hobby relationships, and dynamically constructs and updates social relationship edges between individuals. Finally, the monitoring system uses each individual as a graph node, social relationships as edges connecting nodes, and multi-dimensional dynamic attribute profiles as attributes of nodes to generate an intelligent agent graph that reflects the social network structure of the population and evolves dynamically over time.

[0032] According to an embodiment of the present invention, the step of calculating the index flow based on the agent graph specifically includes: Aggregate the residence information of all individuals within a single spatiotemporal unit to calculate the unit residence degree, which reflects the degree of aggregation of individuals within the spatiotemporal unit; For spatial grid cells on adjacent time slices, the transition matrix of an individual between different cells is analyzed, and the flow flux reflecting the intensity of individual movement between spatiotemporal cells is calculated. For a monitoring area consisting of multiple spatiotemporal units, the agent graph of all individuals in the area is extracted as a subgraph. The network structure features of the subgraph are calculated to obtain the graph connectivity, which reflects the degree of social aggregation of the group.

[0033] It should be noted that, in this embodiment, the monitoring system first aggregates information on all resident individuals within each independent spatiotemporal unit, including the number of resident individuals and the average dwell time of each individual. By comprehensively analyzing this information, the system calculates the unit dwell degree, which quantifies the intensity of individual aggregation within a specific spatiotemporal unit. Secondly, for spatial grid cell pairs on adjacent time slices, the monitoring system analyzes the transfer matrix of the same group of individuals migrating from one spatial cell to another between these two time slices. By statistically analyzing the number and direction of inflow and outflow of individuals, the system calculates the flow flux, which characterizes the spatial mobility of the population. Finally, for a specific monitoring area composed of multiple continuous spatiotemporal units, the monitoring system extracts all individuals within that area and their social relationships, forming a complete intelligent agent subgraph. By calculating the network density, connected components, and other network structure features of this subgraph, the system obtains the graph connectivity, which measures the degree of social aggregation of the group within that area.

[0034] According to an embodiment of the present invention, obtaining the indicator threshold group based on the historical indicator baseline library and the typical event graph library includes: Obtain a historical baseline database, which stores the residence and flux characteristics of each spatiotemporal unit in the same historical period; The dwell time threshold is calculated based on the dwell time characteristics of adjacent time slices; The flux threshold is calculated based on the flux characteristics of adjacent spatial grid cells; Obtain a typical event graph library, which stores the agent sub-graph structure features of the corresponding region when a known event occurs; Based on the preset feature tolerance, the deviation threshold and similarity threshold of the map are obtained.

[0035] It should be noted that, in this embodiment, the monitoring system first acquires a pre-built historical indicator baseline library. This library stores the unit dwell time characteristics and flow flux characteristics of each spatiotemporal unit under multiple historical contemporaneities, categorized by different cycle types such as weekdays, weekends, or holidays. For the spatiotemporal unit to be monitored, the monitoring system retrieves the dwell time sequence corresponding to the historical period from the baseline library based on its temporal position in the time slice. By analyzing the statistical distribution of this sequence, the system dynamically calculates the dwell time threshold applicable to the current monitoring period. Similarly, for the current spatiotemporal unit and its spatial neighboring units, the monitoring system calculates the flux threshold used to determine whether the flow is abnormal based on the flow flux characteristics between adjacent spatial grid units in the historical contemporaneity. Simultaneously, the monitoring system acquires a typical event map library, which includes the intelligent agent sub-map structure features of the affected areas when various confirmed events occurred in history. Based on a preset feature tolerance, the monitoring system compares the real-time extracted sub-map structure with the typical maps in the library. By calculating structural similarity, it obtains the map deviation threshold and similarity threshold used to determine anomalies. The feature tolerance is a feature deviation range dynamically set based on historical experimental data.

[0036] According to an embodiment of the present invention, a graded early warning signal is output based on the determination result, specifically including: If it is determined to be only of primary concern, then a first-level observation warning signal is generated and output. If the warning is determined to be of medium level, a second-level prompt warning signal will be generated and output. If it is determined to be a high-level anomaly, a third-level alarm warning signal will be generated and output. The warning signal includes at least the area where the event occurred, the scope of impact, and the event level information.

[0037] It should be noted that in this embodiment, when only a primary concern judgment is triggered, the monitoring system generates and outputs a first-level observation-level early warning signal. This signal is mainly used to indicate the initial abnormality of individual clustering in the relevant area and suggests continued attention. When an intermediate-level early warning judgment is triggered, the monitoring system generates and outputs a second-level prompt-level early warning signal, indicating that the anomaly has expanded from individual clustering to the spatial flow level and has a tendency to form a local event, requiring the management department to strengthen monitoring. When a high-level anomaly judgment is triggered, the monitoring system generates and outputs a third-level alarm-level early warning signal, confirming the occurrence of a major population flow event with structural characteristics. All output graded early warning signals are encapsulated as structured information packets, which at least include the precise geographical extent of the event, the estimated impact coverage, and the current event level information, so that decision-makers at different levels can take appropriate response measures according to the signal level.

[0038] It is worth mentioning that, before acquiring and analyzing the multi-source positioning data stream, the process also includes: It can access multiple heterogeneous data sources, including at least mobile communication signaling data, location trajectory data, IoT device sensing data, and social media location data; Perform consistency calibration on timestamp accuracy and coordinate system of different data sources; Based on spatiotemporal consistency constraints, outlier detection and removal are performed on the location data; For multi-source localization records of the same individual, a spatiotemporal weighted fusion algorithm is used to determine the optimal estimated location within each time slice; Based on the optimal estimated location, it is mapped to the corresponding spatiotemporal unit.

[0039] It should be noted that in this embodiment, the monitoring system first accesses multiple heterogeneous data sources, including but not limited to signaling data from mobile communication operators, trajectory data from global positioning monitoring systems of various terminal devices, sensing data from IoT sensors deployed in urban infrastructure, and location data from social media with geotags. Since different data sources may differ in time recording accuracy and spatial coordinate reference systems, the monitoring system first normalizes and calibrates the timestamps of all data sources and uniformly transforms the spatial coordinates to a preset standard coordinate system. Subsequently, based on the individual's historical movement patterns and spatiotemporal continuity constraints, the monitoring system performs real-time anomaly detection on the positioning data stream, eliminating outliers caused by signal drift or equipment failure. For the possible multiple duplicate positioning records of the same individual within the same time period, the monitoring system employs a spatiotemporal weighted fusion algorithm, comprehensively considering the signal-to-noise ratio, accuracy level, and temporal proximity of each data source, to calculate an optimal estimated position for each time slice. Finally, the monitoring system uses this optimal estimated position as the individual's true position within the corresponding time slice and maps it to the corresponding spatiotemporal unit.

[0040] It is worth mentioning that the dynamic updating of the agent graph corresponding to an individual based on changes in the individual's spatiotemporal units also includes: Acquire regional functional zoning data and annotate each spatiotemporal unit with semantic labels; When an individual moves between multiple spatiotemporal units, the individual's behavioral semantic trajectory is extracted based on the semantic label sequence of the spatiotemporal unit where it resides. The similarity of behavioral semantic trajectories is analyzed, and individuals are clustered according to their behavioral patterns. The group behavioral labels obtained from the clustering are then fused into the corresponding agent nodes of the individuals as new attributes. Analyze the co-occurrence and collaboration patterns of semantic trajectories to update the weights and types of social relationship edges between individuals.

[0041] It should be noted that in this embodiment, the monitoring system first acquires urban functional zoning data and geographic points of interest data, and labels each spatiotemporal unit with functional semantic tags, such as residential area, commercial area, office area, transportation hub, etc. When an individual resides in multiple spatiotemporal units sequentially within a series of consecutive time slices, the monitoring system extracts the individual's behavioral semantic trajectory based on the sequence of functional semantic tags of these units, thereby transforming simple physical movement into semantic movement with behavioral intent. Subsequently, the monitoring system performs similarity analysis on the behavioral semantic trajectories of all individuals, and uses a clustering algorithm to group individuals with similar behavioral patterns into one category. The group behavioral tags obtained from the clustering are then integrated as new attributes into each agent node, enriching the individual's social identity cognition. At the same time, the monitoring system deeply analyzes the co-occurrence and collaboration patterns of semantic trajectories between individuals. For example, if it finds that two individuals frequently move synchronously from the residential area to the office area at the same time, it updates the weight of the social relationship edge between them accordingly, and may even assign them the relationship type of "colleague" or "peer," so that the agent graph not only reflects physical co-occurrence but also reflects social semantic connections, greatly improving the graph's expressive power and analytical depth.

[0042] Figure 4 A block diagram of a population flow monitoring system based on a spatiotemporal rasterized intelligent agent map according to the present invention is shown.

[0043] like Figure 4 As shown, the second aspect of the present invention discloses a population flow monitoring system 4 based on a spatiotemporal rasterized intelligent agent map, including a memory 41 and a processor 42. The memory includes a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map. When the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map is executed by the processor, it performs the following steps: Based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; Acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; The agent graph corresponding to an individual is dynamically updated based on the changes in the spatiotemporal units of that individual. Based on the agent graph, an indicator flow is calculated, which includes cell dwell degree, flow flux and graph connectivity. Based on the historical indicator baseline library and the typical event graph library, the indicator threshold group is obtained; If the cell dwell time exceeds the dwell time threshold, it is judged as primary concern; In response to initial concerns, if the flow flux exceeds the flux threshold, it is judged as an intermediate warning; In response to the intermediate warning, based on the connectivity of the graph and the typical event graph library, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly.

[0044] Based on the judgment results, a graded early warning signal is output.

[0045] It should be noted that, in this embodiment, the monitoring system first discretizes the target monitoring area into a spatiotemporal grid composed of multiple spatiotemporal units with unique codes, based on preset geospatial division rules and time slicing rules. Subsequently, the monitoring system receives multi-source positioning data streams from different terminals in real time, identifies each mobile individual, and maps it to the corresponding spatiotemporal unit according to its current coordinates. As individuals move between different spatiotemporal units, the monitoring system continuously tracks their positional change trajectory and dynamically updates the agent graph corresponding to each individual. This graph not only records the individual's location history but also contains their behavioral patterns and the relationships between individuals. Based on this, the monitoring system periodically calculates real-time indicator streams to characterize population flow status using the rich information contained in the agent graphs. These indicator streams include at least unit dwell time (reflecting the degree of individual aggregation), flow flux (reflecting the intensity of spatial movement), and graph connectivity (reflecting the degree of social aggregation of the group). Simultaneously, the monitoring system has a pre-set historical indicator baseline library and a typical event map library. By analyzing the deviation of current indicators from historical baselines and typical maps, a set of indicator thresholds for anomaly judgment is obtained. The monitoring system performs a chain comparison judgment: when the cell dwell degree of a certain spatiotemporal unit exceeds the corresponding dwell degree threshold, it is judged as a primary concern; in response to the primary concern, if the flow flux in the area further exceeds the flux threshold, it is judged as upgraded to an intermediate warning; in response to the intermediate warning, the monitoring system further analyzes the map connectivity of the area. If its deviation from the typical event map exceeds the deviation threshold or its similarity exceeds the similarity threshold, it is finally judged as a high-level anomaly. Finally, based on the above judgment results, the monitoring system outputs the corresponding level of graded warning signal to the management terminal.

[0046] According to an embodiment of the present invention, the process of obtaining a spatiotemporal grid composed of multiple spatiotemporal units based on preset geospatial division and time slicing rules specifically includes: According to the preset multi-level geospatial division rules, the target monitoring area is recursively divided into multiple spatial grid units with unique codes and no overlap. According to the preset time discretization rules, the continuous time axis is divided into multiple continuous and equal-length time slices; Each of the spatial grid cells is combined and bound with each of the time slices to generate a spatiotemporal cell with a unique spatiotemporal code; The spatiotemporal grid is composed of all the aforementioned spatiotemporal units.

[0047] It should be noted that, in this embodiment, the monitoring system first recursively segments the target monitoring area according to preset multi-level geospatial division rules. For example, it generates a series of spatial raster units with unique codes and non-overlapping boundaries according to administrative regions, key functional areas, or standard geographic grids, forming a spatially separated structure. Simultaneously, the monitoring system uniformly divides the continuous time axis into multiple continuous and equal-length time slices according to preset time discretization rules, for example, slicing at the minute or second level. Subsequently, the monitoring system binds each spatial raster unit to each time slice through combination operations, generating a spatiotemporal unit with a unique spatiotemporal code. This spatiotemporal unit is both the basic container for data storage and the smallest unit for all subsequent analysis and calculations. Finally, the three-dimensional spatiotemporal raster composed of all spatiotemporal units provides a unified spatiotemporal coordinate system for subsequent population flow monitoring.

[0048] According to an embodiment of the present invention, the step of dynamically updating the agent graph corresponding to an individual based on changes in the spatiotemporal units of that individual specifically includes: Based on each individual's real-time location, their identity identifier is mapped to the corresponding spatiotemporal unit, establishing a relationship between the individual and the spatiotemporal unit; Acquire multi-source attribute data streams and construct dynamic attribute profiles for each individual. The attribute profiles include at least basic attributes, behavioral pattern attributes, and social relationship attributes. Based on the co-occurrence frequency and duration of multiple individuals in the same spatiotemporal unit, the social relationship edges between individuals are dynamically constructed and updated; Using individuals as nodes, social relationships as edges, and attribute profiles as node attributes, a dynamically evolving intelligent agent graph is generated.

[0049] It should be noted that, in this embodiment, the monitoring system first accurately maps the anonymized identity of each individual to the corresponding spatiotemporal unit based on the received real-time location information, thereby establishing a dynamic association between the individual and the specific spatiotemporal location. Simultaneously, with the individual's permission, the monitoring system also accesses multi-source attribute data streams, such as the individual's basic registration information, travel preferences, and consumption records, to construct a dynamic attribute profile for each individual that evolves over time. This profile covers at least basic attributes, behavioral pattern attributes, and social relationship attributes. Furthermore, by analyzing the historical co-occurrence frequency and duration of a large number of individuals within the same spatiotemporal unit, the monitoring system mines and infers potential social connections between individuals based on a pre-trained social relationship model, such as colleague, neighbor, or hobby relationships, and dynamically constructs and updates social relationship edges between individuals. Finally, the monitoring system uses each individual as a graph node, social relationships as edges connecting nodes, and multi-dimensional dynamic attribute profiles as attributes of nodes to generate an intelligent agent graph that reflects the social network structure of the population and evolves dynamically over time.

[0050] According to an embodiment of the present invention, the step of calculating the index flow based on the agent graph specifically includes: Aggregate the residence information of all individuals within a single spatiotemporal unit to calculate the unit residence degree, which reflects the degree of aggregation of individuals within the spatiotemporal unit; For spatial grid cells on adjacent time slices, the transition matrix of an individual between different cells is analyzed, and the flow flux reflecting the intensity of individual movement between spatiotemporal cells is calculated. For a monitoring area consisting of multiple spatiotemporal units, the agent graph of all individuals in the area is extracted as a subgraph. The network structure features of the subgraph are calculated to obtain the graph connectivity, which reflects the degree of social aggregation of the group.

[0051] It should be noted that, in this embodiment, the monitoring system first aggregates information on all resident individuals within each independent spatiotemporal unit, including the number of resident individuals and the average dwell time of each individual. By comprehensively analyzing this information, the system calculates the unit dwelling degree, which quantifies the aggregation intensity of individuals within a specific spatiotemporal unit. Secondly, for spatial grid cell pairs on adjacent time slices, the monitoring system analyzes the transfer matrix of the same group of individuals migrating from one spatial cell to another between these two time slices. By statistically analyzing the number and direction of inflow and outflow of individuals, the system calculates the flow flux, which characterizes the spatial mobility activity of the population. Finally, for a specific monitoring area composed of multiple continuous spatiotemporal units, the monitoring system extracts all individuals within that area and their social relationships, forming a complete intelligent agent subgraph. By calculating the network density, connected components, and other network structure features of this subgraph, the system obtains the graph connectivity, which measures the degree of social aggregation of the group within that area.

[0052] According to an embodiment of the present invention, obtaining the indicator threshold group based on the historical indicator baseline library and the typical event graph library includes: Obtain a historical baseline database, which stores the residence and flux characteristics of each spatiotemporal unit in the same historical period; The dwell time threshold is calculated based on the dwell time characteristics of adjacent time slices; The flux threshold is calculated based on the flux characteristics of adjacent spatial grid cells; Obtain a typical event graph library, which stores the agent sub-graph structure features of the corresponding region when a known event occurs; Based on the preset feature tolerance, the deviation threshold and similarity threshold of the map are obtained.

[0053] It should be noted that, in this embodiment, the monitoring system first acquires a pre-built historical indicator baseline library. This library stores the unit dwell time characteristics and flow flux characteristics of each spatiotemporal unit under multiple historical contemporaneities, categorized by different cycle types such as weekdays, weekends, or holidays. For the spatiotemporal unit to be monitored, the monitoring system retrieves the dwell time sequence corresponding to the historical period from the baseline library based on its temporal position in the time slice. By analyzing the statistical distribution of this sequence, the system dynamically calculates the dwell time threshold applicable to the current monitoring period. Similarly, for the current spatiotemporal unit and its spatial neighboring units, the monitoring system calculates the flux threshold used to determine whether the flow is abnormal based on the flow flux characteristics between adjacent spatial grid units in the historical contemporaneity. Simultaneously, the monitoring system acquires a typical event map library, which includes the intelligent agent sub-map structure features of the affected areas when various confirmed events occurred in history. Based on a preset feature tolerance, the monitoring system compares the real-time extracted sub-map structure with the typical maps in the library. By calculating structural similarity, it obtains the map deviation threshold and similarity threshold used to determine anomalies. The feature tolerance is a feature deviation range dynamically set based on historical experimental data.

[0054] According to an embodiment of the present invention, a graded early warning signal is output based on the determination result, specifically including: If it is determined to be only of primary concern, then a first-level observation warning signal is generated and output. If the warning is determined to be of medium level, a second-level prompt warning signal will be generated and output. If it is determined to be a high-level anomaly, a third-level alarm warning signal will be generated and output. The warning signal includes at least the area where the event occurred, the scope of impact, and the event level information.

[0055] It should be noted that in this embodiment, when only a primary concern judgment is triggered, the monitoring system generates and outputs a first-level observation-level early warning signal. This signal is mainly used to indicate the initial abnormality of individual clustering in the relevant area and suggests continued attention. When an intermediate-level early warning judgment is triggered, the monitoring system generates and outputs a second-level prompt-level early warning signal, indicating that the anomaly has expanded from individual clustering to the spatial flow level and has a tendency to form a local event, requiring the management department to strengthen monitoring. When a high-level anomaly judgment is triggered, the monitoring system generates and outputs a third-level alarm-level early warning signal, confirming the occurrence of a major population flow event with structural characteristics. All output graded early warning signals are encapsulated as structured information packets, which at least include the precise geographical extent of the event, the estimated impact coverage, and the current event level information, so that decision-makers at different levels can take appropriate response measures according to the signal level.

[0056] It is worth mentioning that, before acquiring and analyzing the multi-source positioning data stream, the process also includes: It can access multiple heterogeneous data sources, including at least mobile communication signaling data, location trajectory data, IoT device sensing data, and social media location data; Perform consistency calibration on timestamp accuracy and coordinate system of different data sources; Based on spatiotemporal consistency constraints, outlier detection and removal are performed on the location data; For multi-source localization records of the same individual, a spatiotemporal weighted fusion algorithm is used to determine the optimal estimated location within each time slice; Based on the optimal estimated location, it is mapped to the corresponding spatiotemporal unit.

[0057] It should be noted that in this embodiment, the monitoring system first accesses multiple heterogeneous data sources, including but not limited to signaling data from mobile communication operators, trajectory data from global positioning monitoring systems of various terminal devices, sensing data from IoT sensors deployed in urban infrastructure, and location data from social media with geotags. Since different data sources may differ in time recording accuracy and spatial coordinate reference systems, the monitoring system first normalizes and calibrates the timestamps of all data sources and uniformly transforms the spatial coordinates to a preset standard coordinate system. Subsequently, based on the individual's historical movement patterns and spatiotemporal continuity constraints, the monitoring system performs real-time anomaly detection on the positioning data stream, eliminating outliers caused by signal drift or equipment failure. For the possible multiple duplicate positioning records of the same individual within the same time period, the monitoring system employs a spatiotemporal weighted fusion algorithm, comprehensively considering the signal-to-noise ratio, accuracy level, and temporal proximity of each data source, to calculate an optimal estimated position for each time slice. Finally, the monitoring system uses this optimal estimated position as the individual's true position within the corresponding time slice and maps it to the corresponding spatiotemporal unit.

[0058] It is worth mentioning that the dynamic updating of the agent graph corresponding to an individual based on changes in the individual's spatiotemporal units also includes: Acquire regional functional zoning data and annotate each spatiotemporal unit with semantic labels; When an individual moves between multiple spatiotemporal units, the individual's behavioral semantic trajectory is extracted based on the semantic label sequence of the spatiotemporal unit where it resides. The similarity of behavioral semantic trajectories is analyzed, and individuals are clustered according to their behavioral patterns. The group behavioral labels obtained from the clustering are then fused into the corresponding agent nodes of the individuals as new attributes. Analyze the co-occurrence and collaboration patterns of semantic trajectories to update the weights and types of social relationship edges between individuals.

[0059] It should be noted that in this embodiment, the monitoring system first acquires urban functional zoning data and geographic points of interest data, and labels each spatiotemporal unit with functional semantic tags, such as residential area, commercial area, office area, transportation hub, etc. When an individual resides in multiple spatiotemporal units sequentially within a series of consecutive time slices, the monitoring system extracts the individual's behavioral semantic trajectory based on the sequence of functional semantic tags of these units, thereby transforming simple physical movement into semantic movement with behavioral intent. Subsequently, the monitoring system performs similarity analysis on the behavioral semantic trajectories of all individuals, and uses a clustering algorithm to group individuals with similar behavioral patterns into one category. The group behavioral tags obtained from the clustering are then integrated as new attributes into each agent node, enriching the individual's social identity cognition. At the same time, the monitoring system deeply analyzes the co-occurrence and collaboration patterns of semantic trajectories between individuals. For example, if it finds that two individuals frequently move synchronously from the residential area to the office area at the same time, it updates the weight of the social relationship edge between them accordingly, and may even assign them the relationship type of "colleague" or "peer," so that the agent graph not only reflects physical co-occurrence, but also reflects social semantic connections, greatly improving the graph's expressive power and analytical depth.

[0060] A third aspect of the present invention provides a computer-readable storage medium comprising a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map, wherein when the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map is executed by a processor, the steps of the population flow monitoring method based on the spatiotemporal rasterized intelligent agent map as described in any of the preceding claims are implemented.

[0061] In summary, this invention provides a method and system for monitoring population flow based on a spatiotemporal gridded intelligent agent map. First, a spatiotemporal grid composed of uniquely spatiotemporally coded units is constructed based on preset rules. Multi-source positioning data is then integrated to map individual locations to corresponding units in real time, and the intelligent agent map is dynamically updated based on individual movement. Subsequently, a real-time indicator flow is calculated based on this map, and indicator threshold groups are obtained by combining a historical indicator baseline library and a typical event map library. Then, if the unit dwell rate exceeds the threshold, a primary concern is identified. In response, if the flow throughput exceeds the threshold, a medium-level warning is issued. In response, if the map connectivity deviates from or is similar to the typical map beyond the threshold, a high-level anomaly is identified. Finally, a tiered warning signal is output based on the judgment results. This invention, through the deep integration of spatiotemporal grids and intelligent agent maps, provides intelligent decision support for refined urban management and emergency response.

[0062] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0063] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A population flow monitoring method based on spatiotemporal rasterized intelligent agent maps, characterized in that, The method includes: Based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; Acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; The agent graph corresponding to an individual is dynamically updated based on the changes in the spatiotemporal units of that individual. Based on the agent graph, an indicator flow is calculated, which includes cell dwell degree, flow flux and graph connectivity. Based on the historical indicator baseline library and the typical event graph library, the indicator threshold group is obtained; If the cell dwell time exceeds the dwell time threshold, it is judged as primary concern; In response to initial concerns, if the flow flux exceeds the flux threshold, it is judged as an intermediate warning; In response to the intermediate warning, based on the connectivity of the graph and the typical event graph library, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly. Based on the judgment results, a graded early warning signal is output.

2. The population flow monitoring method based on spatiotemporal rasterized intelligent agent map according to claim 1, characterized in that, The spatiotemporal grid, based on preset geospatial division and time slicing rules, is obtained by generating a spatiotemporal grid composed of multiple spatiotemporal units, specifically including: According to the preset multi-level geospatial division rules, the target monitoring area is recursively divided into multiple spatial grid units with unique codes and no overlap. According to the preset time discretization rules, the continuous time axis is divided into multiple continuous and equal-length time slices; Each of the spatial grid cells is combined and bound with each of the time slices to generate a spatiotemporal cell with a unique spatiotemporal code; The spatiotemporal grid is composed of all the aforementioned spatiotemporal units.

3. The population flow monitoring method based on spatiotemporal rasterized intelligent agent map according to claim 1, characterized in that, The process of dynamically updating the agent graph corresponding to an individual based on changes in the individual's spatiotemporal units specifically includes: Based on each individual's real-time location, their identity identifier is mapped to the corresponding spatiotemporal unit, establishing a relationship between the individual and the spatiotemporal unit; Acquire multi-source attribute data streams and construct dynamic attribute profiles for each individual. The attribute profiles include at least basic attributes, behavioral pattern attributes, and social relationship attributes. Based on the co-occurrence frequency and duration of multiple individuals in the same spatiotemporal unit, the social relationship edges between individuals are dynamically constructed and updated; Using individuals as nodes, social relationships as edges, and attribute profiles as node attributes, a dynamically evolving intelligent agent graph is generated.

4. The population flow monitoring method based on spatiotemporal rasterized intelligent agent map according to claim 1, characterized in that, The step of calculating the indicator flow based on the agent graph specifically includes: Aggregate the residence information of all individuals within a single spatiotemporal unit to calculate the unit residence degree, which reflects the degree of aggregation of individuals within the spatiotemporal unit; For spatial grid cells on adjacent time slices, the transition matrix of an individual between different cells is analyzed, and the flow flux reflecting the intensity of individual movement between spatiotemporal cells is calculated. For a monitoring area consisting of multiple spatiotemporal units, the agent graph of all individuals in the area is extracted as a subgraph. The network structure features of the subgraph are calculated to obtain the graph connectivity, which reflects the degree of social aggregation of the group.

5. The population flow monitoring method based on spatiotemporal rasterized intelligent agent map according to claim 1, characterized in that, The indicator threshold group, obtained based on the historical indicator baseline library and the typical event graph library, includes: Obtain a historical baseline database, which stores the residence and flux characteristics of each spatiotemporal unit in the same historical period; The dwell time threshold is calculated based on the dwell time characteristics of adjacent time slices; The flux threshold is calculated based on the flux characteristics of adjacent spatial grid cells; Obtain a typical event graph library, which stores the agent sub-graph structure features of the corresponding region when a known event occurs; Based on the preset feature tolerance, the deviation threshold and similarity threshold of the map are obtained.

6. The population flow monitoring method based on spatiotemporal rasterized intelligent agent map according to claim 1, characterized in that, Based on the judgment results, a graded early warning signal is output, specifically including: If it is determined to be only of primary concern, then a first-level observation warning signal is generated and output. If the warning is determined to be of medium level, a second-level prompt warning signal will be generated and output. If it is determined to be a high-level anomaly, a third-level alarm warning signal will be generated and output. The warning signal includes at least the area where the event occurred, the scope of impact, and the event level information.

7. A population flow monitoring system based on spatiotemporal rasterized intelligent agent maps, characterized in that, The system includes a memory and a processor. The memory includes a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map. When the processor executes the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map, it performs the following steps: Based on the preset geospatial division and time slicing rules, a spatiotemporal grid composed of multiple spatiotemporal units is obtained; Acquire and analyze multi-source positioning data streams, and map each individual's real-time location to the corresponding spatiotemporal unit; The agent graph corresponding to an individual is dynamically updated based on the changes in the spatiotemporal units of that individual. Based on the agent graph, an indicator flow is calculated, which includes cell dwell degree, flow flux and graph connectivity. Based on the historical indicator baseline library and the typical event graph library, the indicator threshold group is obtained; If the cell dwell time exceeds the dwell time threshold, it is judged as primary concern; In response to initial concerns, if the flow flux exceeds the flux threshold, it is judged as an intermediate warning; In response to the intermediate warning, based on the connectivity of the graph and the typical event graph library, if the graph deviation exceeds the deviation threshold or the graph similarity exceeds the similarity threshold, it is judged as a high-level anomaly. Based on the judgment results, a graded early warning signal is output.

8. A population flow monitoring system based on spatiotemporal rasterized intelligent agent map according to claim 7, characterized in that, The spatiotemporal grid, based on preset geospatial division and time slicing rules, is obtained by generating a spatiotemporal grid composed of multiple spatiotemporal units, specifically including: According to the preset multi-level geospatial division rules, the target monitoring area is recursively divided into multiple spatial grid units with unique codes and no overlap. According to the preset time discretization rules, the continuous time axis is divided into multiple continuous and equal-length time slices; Each of the spatial grid cells is combined and bound with each of the time slices to generate a spatiotemporal cell with a unique spatiotemporal code; The spatiotemporal grid is composed of all the aforementioned spatiotemporal units.

9. A population flow monitoring system based on a spatiotemporal rasterized intelligent agent map according to claim 7, characterized in that, The process of dynamically updating the agent graph corresponding to an individual based on changes in the individual's spatiotemporal units specifically includes: Based on each individual's real-time location, their identity identifier is mapped to the corresponding spatiotemporal unit, establishing a relationship between the individual and the spatiotemporal unit; Acquire multi-source attribute data streams and construct dynamic attribute profiles for each individual. The attribute profiles include at least basic attributes, behavioral pattern attributes, and social relationship attributes. Based on the co-occurrence frequency and duration of multiple individuals in the same spatiotemporal unit, the social relationship edges between individuals are dynamically constructed and updated; Using individuals as nodes, social relationships as edges, and attribute profiles as node attributes, a dynamically evolving intelligent agent graph is generated.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer-readable storage medium includes a population flow monitoring method program based on a spatiotemporal rasterized intelligent agent map. When the population flow monitoring method program based on the spatiotemporal rasterized intelligent agent map is executed by a processor, it implements the steps of the population flow monitoring method based on the spatiotemporal rasterized intelligent agent map as described in any one of claims 1 to 6.

Images