Intelligent community intelligent security scheduling management method based on deep learning

Abstract

The invention relates to the technical field of smart communities, and discloses a smart community intelligent security scheduling management method based on deep learning, which comprises the following steps: acquiring video monitoring data and access control record data of a plurality of communities, cleaning the data, unifying the format and content of the data by adopting a data standardization technology, and storing the unified data into a database; the method comprises the steps of obtaining standardized community security and protection data, mining a personnel flow mode and a vehicle track mode among different communities by adopting a graph algorithm analysis technology on the basis of a community relation network to obtain a cross-community personnel flow mode and a vehicle track mode, and obtaining a cross-community security and protection result according to analysis results of the cross-community personnel flow mode and the vehicle track mode. And an anomaly detection algorithm is adopted to identify potential safety hazards, and if an abnormal condition is detected, early warning information is generated. According to the invention, deep correlation analysis of cross-community security data is realized, and the intelligent level of community security management is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of smart communities, and in particular to a smart community intelligent security dispatching management method based on deep learning. Background Art

[0002] In the intelligent security dispatching management of smart communities, it is necessary to conduct cross-domain correlation mining on massive heterogeneous security data from different communities in order to timely discover potential cross-regional security risks. However, due to the large differences in data formats, data quality and data granularity among various communities, direct data correlation analysis faces many technical challenges. First, it is necessary to clean and standardize the raw data from different sources, but the lack of unified data specifications and standards makes the data preprocessing process complex and time-consuming. Secondly, the storage and management of massive heterogeneous data requires the construction of a suitable data model. Traditional relational databases are difficult to effectively process highly correlated complex data. In addition, cross-community data correlation analysis involves time and space dimensions, and it is necessary to establish an efficient spatiotemporal index to support fast retrieval and calculation, but the indexing technology of high-dimensional spatiotemporal data still needs to be further optimized. Finally, cross-domain data correlation mining requires the use of graph databases and graph algorithms to build a relationship network between people and vehicles for analysis. However, the computational complexity of graph data is high, and it has high requirements for real-time and concurrency.

[0003] Therefore, there is an urgent need for an intelligent security dispatching management method for smart communities based on deep learning. By introducing deep learning technology, the correlation features of heterogeneous data can be automatically learned, and the spatiotemporal indexing and graph computing efficiency can be optimized, thereby achieving efficient, accurate and real-time cross-domain security data association mining. Summary of the invention

[0004] The purpose of the present invention is to overcome one or more of the above-mentioned existing technical problems and provide a smart community intelligent security dispatching management method based on deep learning.

[0005] To achieve the above objectives, the present invention provides a smart community intelligent security dispatch management method based on deep learning, comprising: Obtain video surveillance data and access control record data from multiple communities, clean the acquired data, remove noise and outliers, unify the data format and content, and obtain standardized community security data; Based on standardized community security data, a relationship network model of people and vehicles between communities is constructed. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to associate and represent cross-community data. Based on the relational network model, graph algorithm analysis technology is used to mine the personnel flow patterns and vehicle trajectory patterns between different communities, and the cross-community personnel flow patterns and vehicle trajectory patterns are obtained; Use spatiotemporal indexing technology to organize and manage community security data, thereby performing cross-community data retrieval in spatiotemporal dimensions, associating each piece of data with time and space locations, building a multidimensional index structure, and retrieving personnel and vehicle information within a specific spatiotemporal range; Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards and generate warning information if anomalies are detected; For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

[0006] According to one aspect of the present invention, video surveillance data and access control record data of multiple communities are obtained, and the data are preprocessed to remove noise and outliers in the data; Use data standardization technology to unify the format and content of pre-processed data, eliminate the differences in data format and content in different communities, and obtain standardized community security data; Perform feature extraction on standardized community security data to obtain features such as key frames and target trajectories in video surveillance data, as well as features such as entry and exit time and personnel identity in access control record data; Construct feature vectors of community security data based on the extracted features, and perform numerical conversion on video surveillance data and access control record data; Data fusion technology is used to fuse video surveillance data and access control record data, and the characteristics of the two types of data are comprehensively utilized to obtain community security information; Through data association analysis technology, the implicit relationship in community security data is mined to discover valuable information such as suspicious persons and abnormal events; If a suspicious person or abnormal event is found, the early warning mechanism will be triggered, and the early warning information will be automatically generated and notified to the relevant management personnel so that timely response measures can be taken; According to the needs of community security management, the integrated community security data is visualized to generate intuitive and easy-to-understand charts and maps, which makes it easier for managers to understand the community security situation; The processed community security data is stored in the database to form community security big data resources.

[0007] According to one aspect of the present invention, a basic information table of community personnel and vehicles is constructed based on standardized community security data; Using graph database technology, we create a network model of relationships between people and vehicles in the community, where nodes represent people or vehicles and edges represent the relationships between people or vehicles. By analyzing community security data, we can obtain the correlation information between people and vehicles, and between people and vehicles; According to the acquired association information, corresponding edges are added to the relationship network model to represent the association between people or vehicles, and corresponding attributes are given to the edges; Query and analyze the relationship network model to obtain the association information of people and vehicles across communities; According to the query results, the degree of association between people and vehicles in different communities is determined, and a network diagram of the relationship between people and vehicles across communities is obtained; Through the relationship network diagram, key personnel and vehicles with frequent cross-community activities are identified for focus and analysis; Based on the analysis results, determine the abnormal behavior patterns of people and vehicles across communities and establish an abnormal behavior pattern library; The real-time update mechanism of the graph database is used to update the relationship between people and vehicles in the relationship network model, ensuring that the relationship network model is synchronized with the actual situation, and providing real-time and accurate data support for cross-community personnel and vehicle relationship analysis.

[0008] According to one aspect of the present invention, based on community relationship network data, a personnel flow graph and a vehicle trajectory graph between communities are constructed, wherein nodes represent communities and edges represent personnel flow or vehicle trajectory; The graph clustering algorithm is used to cluster the flow graph of people between communities to obtain the flow pattern of people, where each cluster represents a flow pattern of people; Perform graph embedding on the inter-community vehicle trajectory graph and represent the vehicle trajectory as a low-dimensional vector; By performing cluster analysis on the vehicle trajectory vectors, the vehicle trajectory pattern is obtained, and each cluster corresponds to a vehicle trajectory pattern; According to the personnel flow pattern and vehicle trajectory pattern, a correlation graph of personnel and vehicles across communities is constructed, with the nodes being the personnel flow pattern and vehicle trajectory pattern. If a personnel flow pattern is highly correlated with a vehicle trajectory pattern, an edge is established between the two. The random walk algorithm is used to sample the association graph of people and vehicles across communities to obtain the trajectory association sequence of people flow and vehicles; By mining the association sequence, we can obtain the frequently occurring trajectory patterns of people and vehicles as the association patterns of people and vehicles across communities. Determine the significance of the association pattern of people and vehicles across communities. If the support and confidence of the association pattern exceed the threshold, it will be regarded as a significant association pattern of people and vehicles across communities. Based on the correlation patterns of people and vehicles across communities, the correlation patterns between the flow of people and vehicle trajectories between different communities are determined, revealing the intrinsic connection between the flow of people and vehicle movement across communities.

[0009] According to one aspect of the present invention, spatiotemporal indexing technology is used to organize and manage community security data, associate each piece of data with a time and spatial location, and construct a multidimensional index structure; According to the characteristics of community security data, the data structure of spatiotemporal index is designed, including index information of time dimension and space dimension; Through the time dimension index, the community security data is organized in chronological order to form an index tree of the time dimension. Through the space dimension index, the community security data is divided according to the spatial position to form an index tree of the space dimension. Obtain the time and space dimension information in the spatiotemporal index tree, and construct the query conditions of the spatiotemporal index. If the query conditions include a time range, search for index nodes that meet the conditions in the time dimension index tree; if the query conditions include a space range, search for index nodes that meet the conditions in the space dimension index tree; According to the found time dimension and space dimension index nodes, the corresponding community security data subset is obtained to filter out the personnel and vehicle information that meets the time and space query conditions; The filtered personnel and vehicle information is returned as the result of spatiotemporal cross-community data retrieval, enabling rapid retrieval within a specific spatiotemporal range.

[0010] According to one aspect of the present invention, based on the analysis results of the cross-community personnel flow pattern and vehicle trajectory pattern, an anomaly detection model is constructed, the model including a personnel flow anomaly detection sub-model and a vehicle trajectory anomaly detection sub-model; Clustering algorithm is used to perform cluster analysis on personnel flow data to obtain the normal pattern of personnel flow; By comparing the deviation between real-time personnel flow data and normal patterns, we can determine whether there are abnormalities in personnel flow; If the personnel flow data deviates from the normal mode and exceeds the preset threshold, it is considered that there is abnormal personnel flow, and abnormal personnel flow warning information is generated; Obtain historical data of vehicle trajectories and use sequential pattern mining algorithms to mine frequent patterns of vehicle trajectories; By calculating the similarity between the real-time vehicle trajectory and the frequent pattern, it is determined whether there is an abnormality in the vehicle trajectory; If the similarity between the real-time vehicle trajectory and all frequent patterns is lower than the preset threshold, it is considered that there is an abnormal vehicle trajectory and an abnormal vehicle trajectory warning message is generated; Comprehensive warning information on abnormal personnel flow and abnormal vehicle trajectory, using abnormal information fusion algorithm to determine the final abnormal warning level; According to the abnormal warning level, take corresponding safety precautions to eliminate safety hazards.

[0011] According to one aspect of the present invention, data collection technology is used to obtain real-time location information of people and vehicles in the community, as well as identity information of people and vehicles; Based on the collected location information and identity information, the spatiotemporal trajectory data of people and vehicles flowing between different communities is constructed; Through data analysis technology, the spatiotemporal trajectory data of people and vehicles are analyzed to identify people and vehicles that frequently move between multiple communities; Adopt anomaly detection algorithm to judge whether the current flow behavior is abnormal based on the historical flow patterns of people and vehicles, and obtain the information of abnormal people and vehicles; Based on the information of abnormal persons and vehicles, combined with their identity information, the possible associations between abnormal persons and vehicles are mined through association analysis technology; Using social network analysis technology, we build a relationship map of people and vehicles across communities based on the association between abnormal people and vehicles. Using spatiotemporal data visualization technology, the spatiotemporal trajectory data of abnormal personnel and vehicles are integrated with the relationship map to generate an intuitive spatiotemporal trajectory map; In the time-space trajectory diagram, different colors or icons are used to highlight the identified abnormal persons and vehicles, making it easier for security personnel to locate them quickly; Through visual interactive technology, security personnel are provided with on-picture operation functions to query detailed information of abnormal personnel and vehicles, so as to achieve accurate verification and rapid disposal.

[0012] To achieve the above objectives, the present invention provides a smart community intelligent security dispatching and management system based on deep learning, comprising: Community security data acquisition module: obtains video surveillance data and access control record data from multiple communities, cleans the acquired data, removes noise and outliers, unifies the data format and content, and obtains standardized community security data; Relationship network model building module: Based on standardized community security data, a relationship network model of people and vehicles between communities is built. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to represent the association of cross-community data. Behavior pattern acquisition module: Based on the relationship network model, the graph algorithm analysis technology is used to mine the personnel flow pattern and vehicle trajectory pattern between different communities, and the cross-community personnel flow pattern and vehicle trajectory pattern are obtained; Behavior retrieval module: Uses spatiotemporal indexing technology to organize and manage community security data, thereby performing cross-community data retrieval in spatiotemporal dimensions, associating each piece of data with time and space locations, building a multidimensional index structure, and retrieving personnel and vehicle information within a specific spatiotemporal range; Behavior warning module: Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards. If an abnormal situation is detected, warning information is generated; Early warning and disposal module: For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities, and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

[0013] To achieve the above-mentioned purpose, the present invention provides an electronic device, including a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the above-mentioned intelligent security dispatching management method for a smart community based on deep learning is implemented.

[0014] To achieve the above-mentioned purpose, the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned intelligent security dispatching management method for a smart community based on deep learning is implemented.

[0015] Based on this, the beneficial effects of the present invention are as follows: firstly, the video surveillance and access control record data of multiple communities are obtained, cleaned and standardized, and then the relationship network model of people and vehicles between communities is constructed by using graph database technology to realize the association representation of cross-community data; Use graph algorithm analysis technology to mine the flow of people and vehicle trajectories between different communities, support data retrieval in time and space, use time and space indexing technology to organize and manage security data, and use anomaly detection algorithms to identify potential safety hazards and generate early warning information; Use visualization technology to generate cross-community relationship maps and spatiotemporal trajectory maps to intuitively display abnormal situations; It has achieved in-depth correlation analysis of cross-community security data, effectively improving the level of intelligence in community safety management. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 It is a flow chart of a smart community intelligent security dispatch management method based on deep learning according to an exemplary embodiment; Figure 2 It is a flow chart of a smart community intelligent security dispatching and management system based on deep learning according to an exemplary embodiment. DETAILED DESCRIPTION

[0017] The present invention will now be discussed with reference to exemplary embodiments. It should be understood that the embodiments discussed are only for enabling those skilled in the art to better understand and thereby implement the present invention, rather than implying any limitation on the scope of the present invention.

[0018] As used herein, the term “including” and variations thereof are to be interpreted as open-ended terms meaning “including but not limited to.” The term “based on” is to be interpreted as “based, at least in part, on,” and the terms “one embodiment” and “an embodiment” are to be interpreted as “at least one embodiment.”

[0019] According to one embodiment of the present invention, Figure 1 is a flow chart of a smart community intelligent security dispatch management method based on deep learning according to an exemplary embodiment, such as Figure 1 As shown, to achieve the above purpose, the present invention provides a smart community intelligent security dispatch management method based on deep learning, including: Obtain video surveillance data and access control record data from multiple communities, clean the acquired data, remove noise and outliers, unify the data format and content, and obtain standardized community security data; Based on standardized community security data, a relationship network model of people and vehicles between communities is constructed. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to associate and represent cross-community data. Based on the relational network model, graph algorithm analysis technology is used to mine the personnel flow patterns and vehicle trajectory patterns between different communities, and the cross-community personnel flow patterns and vehicle trajectory patterns are obtained; Use spatiotemporal indexing technology to organize and manage community security data, thereby performing cross-community data retrieval in spatiotemporal dimensions, associating each piece of data with time and space locations, building a multidimensional index structure, and retrieving personnel and vehicle information within a specific spatiotemporal range; Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards and generate warning information if anomalies are detected; For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

[0020] According to one embodiment of the present invention, video surveillance data and access control record data of multiple communities are obtained, and the data are preprocessed to remove noise and outliers in the data; Use data standardization technology to unify the format and content of pre-processed data, eliminate the differences in data format and content in different communities, and obtain standardized community security data; Perform feature extraction on standardized community security data to obtain features such as key frames and target trajectories in video surveillance data, as well as features such as entry and exit time and personnel identity in access control record data; Construct feature vectors of community security data based on the extracted features, and perform numerical conversion on video surveillance data and access control record data; Data fusion technology is used to fuse video surveillance data and access control record data, and the characteristics of the two types of data are comprehensively utilized to obtain community security information; Through data association analysis technology, the implicit relationship in community security data is mined to discover valuable information such as suspicious persons and abnormal events; If a suspicious person or abnormal event is found, the early warning mechanism will be triggered, and the early warning information will be automatically generated and notified to the relevant management personnel so that timely response measures can be taken; According to the needs of community security management, the integrated community security data is visualized to generate intuitive and easy-to-understand charts and maps, which makes it easier for managers to understand the community security situation; The processed community security data is stored in the database to form community security big data resources.

[0021] According to one embodiment of the present invention, a basic information table of community personnel and vehicles is constructed based on standardized community security data; Using graph database technology, we create a network model of relationships between people and vehicles in the community, where nodes represent people or vehicles and edges represent the relationships between people or vehicles. By analyzing community security data, we can obtain the correlation information between people and vehicles, and between people and vehicles; According to the obtained association information, corresponding edges are added to the relationship network model to represent the association relationship between people or vehicles, and corresponding attributes are given to the edges; Query and analyze the relationship network model to obtain the association information of people and vehicles across communities; According to the query results, the degree of association between people and vehicles in different communities is determined, and a network diagram of the relationship between people and vehicles across communities is obtained; Through the relationship network diagram, key personnel and vehicles with frequent cross-community activities are identified for focused attention and analysis; Based on the analysis results, determine the abnormal behavior patterns of people and vehicles across communities and establish an abnormal behavior pattern library; The real-time update mechanism of the graph database is used to update the relationship between people and vehicles in the relationship network model, ensuring that the relationship network model is synchronized with the actual situation, and providing real-time and accurate data support for cross-community personnel and vehicle relationship analysis.

[0022] According to one embodiment of the present invention, based on community relationship network data, a personnel flow graph and a vehicle trajectory graph between communities are constructed, wherein nodes represent communities and edges represent personnel flow or vehicle trajectories; The graph clustering algorithm is used to cluster the flow graph of people between communities to obtain the flow pattern of people, where each cluster represents a flow pattern of people; Perform graph embedding on the inter-community vehicle trajectory graph and represent the vehicle trajectory as a low-dimensional vector; By performing cluster analysis on the vehicle trajectory vectors, the vehicle trajectory pattern is obtained, and each cluster corresponds to a vehicle trajectory pattern; According to the personnel flow pattern and vehicle trajectory pattern, a correlation graph of personnel and vehicles across communities is constructed, with the nodes being the personnel flow pattern and vehicle trajectory pattern. If a personnel flow pattern is highly correlated with a vehicle trajectory pattern, an edge is established between the two. The random walk algorithm is used to sample the association graph of people and vehicles across communities to obtain the trajectory association sequence of people flow and vehicles; By mining the association sequence, we can obtain the frequently occurring trajectory patterns of people and vehicles as the association patterns of people and vehicles across communities. Determine the significance of the association pattern of people and vehicles across communities. If the support and confidence of the association pattern exceed the threshold, it will be regarded as a significant association pattern of people and vehicles across communities. Based on the correlation patterns of people and vehicles across communities, the correlation patterns between the flow of people and vehicle trajectories between different communities are determined, revealing the intrinsic connection between the flow of people and vehicle movement across communities.

[0023] According to one embodiment of the present invention, spatiotemporal indexing technology is used to organize and manage community security data, and each piece of data is associated with a time and spatial location to construct a multidimensional index structure; According to the characteristics of community security data, the data structure of spatiotemporal index is designed, including index information of time dimension and space dimension; Through the time dimension index, the community security data is organized in chronological order to form an index tree of the time dimension. Through the space dimension index, the community security data is divided according to the spatial position to form an index tree of the space dimension. Obtain the time and space dimension information in the spatiotemporal index tree, and construct the query conditions of the spatiotemporal index. If the query conditions include a time range, search for index nodes that meet the conditions in the time dimension index tree; if the query conditions include a space range, search for index nodes that meet the conditions in the space dimension index tree; According to the found time dimension and space dimension index nodes, the corresponding community security data subset is obtained to filter out the personnel and vehicle information that meets the time and space query conditions; The filtered personnel and vehicle information is returned as the result of spatiotemporal cross-community data retrieval, enabling rapid retrieval within a specific spatiotemporal range.

[0024] According to one embodiment of the present invention, based on the analysis results of the cross-community personnel flow pattern and vehicle trajectory pattern, an anomaly detection model is constructed, the model including a personnel flow anomaly detection sub-model and a vehicle trajectory anomaly detection sub-model; Clustering algorithm is used to perform cluster analysis on personnel flow data to obtain the normal pattern of personnel flow; By comparing the deviation between real-time personnel flow data and normal patterns, we can determine whether there are abnormalities in personnel flow; If the personnel flow data deviates from the normal mode and exceeds the preset threshold, it is considered that there is abnormal personnel flow, and abnormal personnel flow warning information is generated; Obtain historical data of vehicle trajectories and use sequential pattern mining algorithms to mine frequent patterns of vehicle trajectories; By calculating the similarity between the real-time vehicle trajectory and the frequent pattern, it is determined whether there is an abnormality in the vehicle trajectory; If the similarity between the real-time vehicle trajectory and all frequent patterns is lower than the preset threshold, it is considered that there is an abnormal vehicle trajectory and an abnormal vehicle trajectory warning message is generated; Comprehensive warning information on abnormal personnel flow and abnormal vehicle trajectory, using abnormal information fusion algorithm to determine the final abnormal warning level; According to the abnormal warning level, take corresponding safety precautions to eliminate safety hazards.

[0025] According to one embodiment of the present invention, data collection technology is used to obtain real-time location information of people and vehicles in the community, as well as identity information of people and vehicles; Based on the collected location information and identity information, the spatiotemporal trajectory data of people and vehicles flowing between different communities is constructed; Through data analysis technology, the spatiotemporal trajectory data of people and vehicles are analyzed to identify people and vehicles that frequently move between multiple communities; Adopt anomaly detection algorithm to judge whether the current flow behavior is abnormal based on the historical flow patterns of people and vehicles, and obtain the information of abnormal people and vehicles; Based on the information of abnormal persons and vehicles, combined with their identity information, the possible associations between abnormal persons and vehicles are mined through association analysis technology; Using social network analysis technology, we build a relationship map of people and vehicles across communities based on the association between abnormal people and vehicles. Using spatiotemporal data visualization technology, the spatiotemporal trajectory data of abnormal personnel and vehicles are integrated with the relationship map to generate an intuitive spatiotemporal trajectory map; In the time-space trajectory diagram, different colors or icons are used to highlight the identified abnormal persons and vehicles, making it easier for security personnel to locate them quickly; Through visual interactive technology, security personnel are provided with on-picture operation functions to query detailed information of abnormal personnel and vehicles, so as to achieve accurate verification and rapid disposal.

[0026] According to one embodiment of the present invention, firstly, raw data is obtained from the monitoring systems and access control systems of multiple communities through an API interface, the data is preprocessed, the median filter is used to remove noise, and the 3σ criterion is used to remove outliers. Then, the data is standardized using the JSON format to unify the format of data from different sources. Key frames are extracted from video data, the YOLOv5 target detection algorithm is used to identify targets such as personnel and vehicles, and the DeepSORT algorithm is used to track the target trajectory. Time, ID and other features are extracted from access control records. The extracted features are converted into feature vectors, such as the personnel feature vector containing attributes such as location coordinates, speed, and body shape. The DS evidence theory is used to fuse the video and access control data, and the confidence after fusion is calculated. The Apriori algorithm is used to mine data association rules, such as discovering frequently occurring suspicious behavior patterns. An early warning threshold is set, and an early warning is automatically triggered when abnormal behavior is found, and the management personnel are notified by SMS. ECharts is used to generate visual charts such as heat maps and trajectory maps. Finally, the processed data is stored in a MongoDB database, and an index is established to optimize query performance. The entire process forms a complete technical chain from data collection, preprocessing, feature extraction to data fusion, correlation analysis, and visualization, realizing the intelligent processing and application of community security data.

[0027] According to one embodiment of the present invention, when constructing a basic information table of community personnel and vehicles, a MySQL database can be used to design a data table containing fields such as personnel ID, name, ID number, vehicle ID, and license plate number. For example, the personnel information table can contain 10,000 records, and the vehicle information table can contain 5,000 records. When creating a network model of personnel and vehicle relationships between communities, a Neo4j graph database can be used, with personnel and vehicles as nodes and association relationships as edges. By analyzing 1,000 community security data within 1 month, using a spatiotemporal association algorithm, if two personnel or vehicles appear at the same time and place, an association edge is added to the graph model, and the attributes such as the association time, place, and frequency are recorded. Using the Cypher query language, the activity trajectory of a person in different communities can be queried, for example, "Zhang San" appears in communities A, B, and C in a week. Visualize the query results, generate a cross-community personnel and vehicle relationship network diagram, and identify key personnel and vehicles with high centrality through the PageRank algorithm. For these key targets, we use association rule mining algorithms to analyze their activity patterns and social networks. For example, we find that "Zhang San" meets "Li Si" in community A every Tuesday morning and frequently travels between communities B and C on weekends. Through cluster analysis, we can find some abnormal behavior patterns, such as 5 cars frequently changing license plates within a month and appearing in crime-prone areas many times. These abnormal patterns are added to the early warning rule library to detect abnormal behaviors in real time. Finally, we use the event-driven mechanism of the graph database to update the relationship network model in real time to ensure the accuracy of the analysis results.

[0028] According to one embodiment of the present invention, firstly, a personnel flow graph and a vehicle trajectory graph between communities are constructed based on community relationship network data. Wherein a node represents a community, and an edge represents personnel flow or a vehicle trajectory. For example, each community can be abstracted as a node, and a directed edge is established between the corresponding community nodes according to the personnel flow record, and the edge weight is the personnel flow amount. Similarly, a vehicle trajectory edge is established between communities according to the vehicle GPS trajectory data. Then, the Louvain algorithm is used to perform community detection on the personnel flow graph, and the community division result with the largest modularity is used as the personnel flow pattern. Then, the Node2Vec algorithm is used to learn the 64-dimensional embedded vector representation of the nodes in the vehicle trajectory graph, and then the vector is clustered using K-Means to obtain 8 vehicle trajectory patterns. On this basis, a cross-community personnel-vehicle association graph is constructed. For example, when the Jaccard similarity of the personnel flow pattern i and the vehicle trajectory pattern j exceeds 6, an edge is added between the corresponding nodes of the two. After that, Node2Vec is used to perform random walk sampling on the association graph, 100 nodes are selected each time, the walk length is 10, and 50 rounds are iterated to obtain a personnel flow-vehicle trajectory association sequence. The PrefixSpan algorithm was used to mine frequent subsequence patterns from the sequence, with the support threshold set to 05 and the confidence threshold set to 8. Finally, 10 significant cross-community personnel-vehicle association patterns were obtained. Through analysis, it was found that the association patterns related to commercial areas and transportation hubs were the most significant, indicating that commercial and transportation activities are the key factors driving the movement of personnel and vehicles across communities. In summary, this study reveals the inherent association between the movement of personnel and vehicles across communities in terms of spatiotemporal patterns, which is of great value for understanding the dynamics of urban people and vehicles and optimizing traffic management.

[0029] According to one embodiment of the present invention, in community security data management, spatiotemporal indexing technology can be used to efficiently organize and retrieve data. First, according to the characteristics of community security data, a multidimensional index structure including time dimension and space dimension is designed. The time dimension index can use a data structure such as a B+ tree to sort the data according to the timestamp to form an index tree of the time dimension. The spatial dimension index can use a spatial index structure such as an R tree and a quadtree to divide the community into different spatial regions, each region corresponds to an index node, and form an index tree of the spatial dimension. When constructing the query condition, if the time range is included, an algorithm such as binary search is used in the time dimension index tree to locate the index node that meets the time condition; if the spatial range is included, an algorithm such as rectangular area intersection is used in the spatial dimension index tree to find the index node that falls within the target area. Through the joint query of the spatiotemporal index tree, a subset of community security data that meets the spatiotemporal conditions can be quickly obtained. Finally, the personnel and vehicle information that meets the specific query conditions is further screened out in the data subset and returned as the result of the spatiotemporal cross-community data retrieval. Through a reasonably designed spatiotemporal index structure and efficient query algorithms, rapid retrieval of massive community security data can be achieved, flexible spatiotemporal query conditions can be supported, and the efficiency and performance of community security data management can be improved.

[0030] According to one embodiment of the present invention, an anomaly detection model is constructed based on the analysis results of the cross-community personnel flow pattern and vehicle trajectory pattern. The model includes a personnel flow anomaly detection submodel and a vehicle trajectory anomaly detection submodel. First, the K-means clustering algorithm is used to perform cluster analysis on the personnel flow data of the past three months, and the personnel flow pattern is divided into five categories, which respectively represent normal flow patterns such as going to work, going home from school, shopping, and seeking medical treatment, so as to obtain the normal pattern of personnel flow. Then, by comparing the degree of deviation between the real-time personnel flow data and the normal pattern, the Euclidean distance between the real-time personnel flow data and each normal pattern is calculated. If the distance from the nearest normal pattern exceeds a preset threshold (such as 5), it is considered that there is an abnormality in personnel flow, and an abnormal warning information of personnel flow is generated. Next, the historical data of vehicle trajectory in the past year is obtained, and the PrefixSpan sequence pattern mining algorithm is used to mine the frequent patterns of vehicle trajectory, and the support threshold is set to 20%. By calculating the similarity between the real-time vehicle trajectory and the frequent patterns (such as using the edit distance algorithm), it is determined whether the vehicle trajectory is abnormal. If the similarity between the real-time vehicle trajectory and all frequent patterns is lower than the preset threshold (such as 8), it is considered that there is an abnormality in the vehicle trajectory, and an abnormal warning information of the vehicle trajectory is generated. Finally, the abnormal warning information of personnel flow and vehicle trajectory is integrated, and the weighted average algorithm is used to assign weights of 6 and 4 respectively to calculate the comprehensive score of abnormal warning and determine the final abnormal warning level (such as 0-20 points for level 1 warning, 20-50 points for level 2 warning, and 50 points or more for level 3 warning). According to the abnormal warning level, corresponding safety precautions are taken, such as strengthening the monitoring frequency of key areas during level 1 warning, checking suspicious personnel and vehicles during level 2 warning, and launching emergency plans during level 3 warning, etc., to eliminate safety hazards.

[0031] In addition, in order to achieve the above-mentioned purpose of the invention, the present invention also provides a smart community intelligent security dispatching management system based on deep learning. Figure 2 is a flow chart of a smart community intelligent security dispatch management system based on deep learning according to an exemplary embodiment, such as Figure 2 As shown, a smart community intelligent security dispatching and management system based on deep learning in the present invention includes: Community security data acquisition module: obtains video surveillance data and access control record data from multiple communities, cleans the acquired data, removes noise and outliers, unifies the data format and content, and obtains standardized community security data; Relationship network model building module: Based on standardized community security data, a relationship network model of people and vehicles between communities is built. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to represent the association of cross-community data. Behavior pattern acquisition module: Based on the relationship network model, the graph algorithm analysis technology is used to mine the personnel flow pattern and vehicle trajectory pattern between different communities, and the cross-community personnel flow pattern and vehicle trajectory pattern are obtained; Behavior retrieval module: Uses spatiotemporal indexing technology to organize and manage community security data, thereby performing cross-community data retrieval in spatiotemporal dimensions, associating each piece of data with time and space locations, building a multidimensional index structure, and retrieving personnel and vehicle information within a specific spatiotemporal range; Behavior warning module: Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards. If an abnormal situation is detected, warning information is generated; Early warning and disposal module: For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities, and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

[0032] In order to achieve the above-mentioned purpose of the invention, the present invention also provides an electronic device, which includes: a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, the above-mentioned smart community intelligent security dispatching management method based on deep learning is implemented.

[0033] In order to achieve the above-mentioned purpose of the invention, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned smart community intelligent security dispatching management method based on deep learning is implemented.

[0034] Based on this, the beneficial effects of the present invention are that, first, video surveillance and access control record data of multiple communities are obtained, cleaned and standardized, and then a relationship network model of people and vehicles between communities is constructed using graph database technology to realize the correlation representation of cross-community data, and graph algorithm analysis technology is used to mine the personnel flow and vehicle trajectory patterns between different communities, support data retrieval in the spatiotemporal dimension, and use spatiotemporal indexing technology to organize and manage security data. Potential safety hazards are identified through anomaly detection algorithms and early warning information is generated. Visualization technology is used to generate cross-community relationship maps and spatiotemporal trajectory maps to intuitively display abnormal situations, thereby realizing in-depth correlation analysis of cross-community security data and effectively improving the level of intelligence in community safety management.

[0035] Those of ordinary skill in the art will appreciate that the modules and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0036] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and equipment can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0037] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or modules, which can be electrical, mechanical or other forms.

[0038] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiments of the present invention.

[0039] In addition, each functional module in the embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0040] If the functions are implemented in the form of software function 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 the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the energy-saving signal sending / receiving method of each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a ROM, a RAM, a magnetic disk, or an optical disk.

[0041] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the inventive concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the present application.

[0042] It should be understood that the size of the serial numbers of each step in the content of the invention and the embodiments of the present invention does not absolutely mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

Claims

1. A smart community intelligent security dispatch management method based on deep learning, characterized in that: include: Obtain video surveillance data and access control record data from multiple communities, clean the acquired data, remove noise and outliers, unify the data format and content, and obtain standardized community security data; Based on standardized community security data, a relationship network model of people and vehicles between communities is constructed. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to associate and represent cross-community data. Based on the relational network model, graph algorithm analysis technology is used to mine the personnel flow patterns and vehicle trajectory patterns between different communities, and the cross-community personnel flow patterns and vehicle trajectory patterns are obtained; Use spatiotemporal indexing technology to organize and manage community security data, so as to perform cross-community data retrieval in spatiotemporal dimensions, associate each piece of data with time and space location, build a multidimensional index structure, and retrieve personnel and vehicle information within a specific spatiotemporal range; Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards and generate warning information if anomalies are detected; For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

2. The method for intelligent security dispatching and management of a smart community based on deep learning as claimed in claim 1, characterized in that: Obtain video surveillance data and access control record data from multiple communities, pre-process the data, and remove noise and outliers from the data; Use data standardization technology to unify the format and content of pre-processed data, eliminate the differences in data format and content in different communities, and obtain standardized community security data; Perform feature extraction on standardized community security data to obtain features such as key frames and target trajectories in video surveillance data, as well as features such as entry and exit time and personnel identity in access control record data; Construct feature vectors of community security data based on the extracted features, and perform numerical conversion on video surveillance data and access control record data; Data fusion technology is used to fuse video surveillance data and access control record data, and the characteristics of the two types of data are comprehensively utilized to obtain community security information; Through data association analysis technology, the implicit relationship in community security data is mined to discover valuable information such as suspicious persons and abnormal events; If a suspicious person or abnormal event is found, the early warning mechanism will be triggered, and the early warning information will be automatically generated and notified to the relevant management personnel so that timely response measures can be taken; According to the needs of community security management, the integrated community security data is visualized to generate intuitive and easy-to-understand charts and maps, which makes it easier for managers to understand the community security situation; The processed community security data is stored in the database to form community security big data resources.

3. A smart community intelligent security dispatch management method based on deep learning as claimed in claim 2, characterized in that: Based on standardized community security data, construct a basic information table of community personnel and vehicles; Using graph database technology, we create a network model of relationships between people and vehicles in the community, where nodes represent people or vehicles and edges represent the relationships between people or vehicles. By analyzing community security data, we can obtain the correlation information between people and vehicles, and between people and vehicles; According to the acquired association information, corresponding edges are added to the relationship network model to represent the association between people or vehicles, and corresponding attributes are given to the edges; Query and analyze the relationship network model to obtain the association information of people and vehicles across communities; According to the query results, the degree of association between people and vehicles in different communities is determined, and a network diagram of the relationship between people and vehicles across communities is obtained; Through the relationship network diagram, key personnel and vehicles with frequent cross-community activities are identified for focus and analysis; Based on the analysis results, determine the abnormal behavior patterns of people and vehicles across communities and establish an abnormal behavior pattern library; The real-time update mechanism of the graph database is used to update the relationship between people and vehicles in the relationship network model, ensuring that the relationship network model is synchronized with the actual situation, and providing real-time and accurate data support for cross-community personnel and vehicle relationship analysis.

4. A smart community intelligent security dispatch management method based on deep learning as claimed in claim 3, characterized in that: According to the community relationship network data, the personnel flow graph and vehicle trajectory graph between communities are constructed, where nodes represent communities and edges represent personnel flow or vehicle trajectory; The graph clustering algorithm is used to cluster the flow graph of people between communities to obtain the flow pattern of people, where each cluster represents a flow pattern of people; Perform graph embedding on the inter-community vehicle trajectory graph and represent the vehicle trajectory as a low-dimensional vector; By performing cluster analysis on the vehicle trajectory vectors, the vehicle trajectory pattern is obtained, and each cluster corresponds to a vehicle trajectory pattern; According to the personnel flow pattern and vehicle trajectory pattern, a correlation graph of personnel and vehicles across communities is constructed, with the nodes being the personnel flow pattern and vehicle trajectory pattern. If a personnel flow pattern is highly correlated with a vehicle trajectory pattern, an edge is established between the two. The random walk algorithm is used to sample the association graph of people and vehicles across communities to obtain the trajectory association sequence of people flow and vehicles; By mining the association sequence, we can obtain the frequently occurring trajectory patterns of people and vehicles as the association patterns of people and vehicles across communities. Determine the significance of the association pattern of people and vehicles across communities. If the support and confidence of the association pattern exceed the threshold, it will be regarded as a significant association pattern of people and vehicles across communities. Based on the correlation patterns of people and vehicles across communities, the correlation patterns between the flow of people and vehicle trajectories between different communities are determined, revealing the intrinsic connection between the flow of people and vehicle movement across communities.

5. The method for intelligent security dispatching and management of a smart community based on deep learning as claimed in claim 4, characterized in that: Use spatiotemporal indexing technology to organize and manage community security data, associate each piece of data with time and space, and build a multidimensional index structure; According to the characteristics of community security data, the data structure of spatiotemporal index is designed, including index information of time dimension and space dimension; Through the time dimension index, the community security data is organized in chronological order to form an index tree of the time dimension. Through the space dimension index, the community security data is divided according to the spatial position to form an index tree of the space dimension. Obtain the time and space dimension information in the spatiotemporal index tree, and construct the query conditions of the spatiotemporal index. If the query conditions include a time range, search for index nodes that meet the conditions in the time dimension index tree; if the query conditions include a space range, search for index nodes that meet the conditions in the space dimension index tree; According to the found time dimension and space dimension index nodes, the corresponding community security data subset is obtained to filter out the personnel and vehicle information that meets the time and space query conditions; The filtered personnel and vehicle information is returned as the result of spatiotemporal cross-community data retrieval, enabling rapid retrieval within a specific spatiotemporal range.

6. A smart community intelligent security dispatch management method based on deep learning as claimed in claim 5, characterized in that: According to the analysis results of the cross-community personnel flow pattern and vehicle trajectory pattern, an anomaly detection model is constructed, which includes a personnel flow anomaly detection sub-model and a vehicle trajectory anomaly detection sub-model; Clustering algorithm is used to perform cluster analysis on personnel flow data to obtain the normal pattern of personnel flow; By comparing the deviation between real-time personnel flow data and normal patterns, we can determine whether there are abnormalities in personnel flow; If the personnel flow data deviates from the normal mode and exceeds the preset threshold, it is considered that there is abnormal personnel flow, and abnormal personnel flow warning information is generated; Obtain historical data of vehicle trajectories and use sequential pattern mining algorithms to mine frequent patterns of vehicle trajectories; By calculating the similarity between the real-time vehicle trajectory and the frequent pattern, it is determined whether there is an abnormality in the vehicle trajectory; If the similarity between the real-time vehicle trajectory and all frequent patterns is lower than the preset threshold, it is considered that there is an abnormal vehicle trajectory and an abnormal vehicle trajectory warning message is generated; Comprehensive warning information on abnormal personnel flow and abnormal vehicle trajectory, using abnormal information fusion algorithm to determine the final abnormal warning level; According to the abnormal warning level, take corresponding safety precautions to eliminate safety hazards.

7. A smart community intelligent security dispatch management method based on deep learning as claimed in claim 6, characterized in that: Use data collection technology to obtain real-time location information of people and vehicles in the community, as well as their identity information; Based on the collected location information and identity information, the spatiotemporal trajectory data of people and vehicles flowing between different communities is constructed; Through data analysis technology, the spatiotemporal trajectory data of people and vehicles are analyzed to identify people and vehicles that frequently move between multiple communities; Adopt anomaly detection algorithm to judge whether the current flow behavior is abnormal based on the historical flow patterns of people and vehicles, and obtain the information of abnormal people and vehicles; Based on the information of abnormal persons and vehicles, combined with their identity information, the possible associations between abnormal persons and vehicles are mined through association analysis technology; Using social network analysis technology, we build a relationship map of people and vehicles across communities based on the association between abnormal people and vehicles. Using spatiotemporal data visualization technology, the spatiotemporal trajectory data of abnormal personnel and vehicles are integrated with the relationship map to generate an intuitive spatiotemporal trajectory map; In the time-space trajectory diagram, different colors or icons are used to highlight the identified abnormal persons and vehicles, making it easier for security personnel to locate them quickly; Through visual interactive technology, security personnel are provided with on-picture operation functions to query detailed information of abnormal personnel and vehicles, so as to achieve accurate verification and rapid disposal.

8. A smart community intelligent security dispatching and management system based on deep learning, characterized in that: include: Community security data acquisition module: obtains video surveillance data and access control record data from multiple communities, cleans the acquired data, removes noise and outliers, unifies the data format and content, and obtains standardized community security data; Relationship network model building module: Based on standardized community security data, a relationship network model of people and vehicles between communities is built. The nodes in the relationship network model represent people or vehicles, and the edges represent the association between people or vehicles, so as to represent the association of cross-community data. Behavior pattern acquisition module: Based on the relationship network model, the graph algorithm analysis technology is used to mine the personnel flow pattern and vehicle trajectory pattern between different communities, and the cross-community personnel flow pattern and vehicle trajectory pattern are obtained; Behavior retrieval module: Uses spatiotemporal indexing technology to organize and manage community security data, thereby performing cross-community data retrieval in spatiotemporal dimensions, associating each piece of data with time and space locations, building a multidimensional index structure, and retrieving personnel and vehicle information within a specific spatiotemporal range; Behavior warning module: Based on the analysis results of cross-community personnel flow patterns and vehicle trajectory patterns, anomaly detection algorithms are used to identify potential safety hazards. If an abnormal situation is detected, warning information is generated; Early warning and disposal module: For the potential safety hazards identified, visualization technology is used to generate cross-community relationship maps and space-time trajectory maps. The relationship maps and trajectory maps show the flow of people and vehicles between different communities, and highlight abnormal people and vehicles, allowing security personnel to verify and deal with them.

9. An electronic device, characterized in that: It includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements a smart community intelligent security dispatching management method based on deep learning as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, it implements a smart community intelligent security dispatching management method based on deep learning as described in any one of claims 1 to 7.

Images