Research On Distributed Storage Technology Of Public Safety Monitoring Data Based On NoSQL

With the widespread deployment of urban public safety monitoring networks in China,it is possible to obtain public safety monitoring data from large-scale multi-source cities.Urban public safety monitoring data provides a data basis for diagnosing,discovering,and warning urban public safety events.Effective management of urban public safety monitoring data is the primary premise and important foundation for supporting urban public safety diagnosis,analysis,and application.The urban public safety monitoring network provides large-scale,unstructured multi-modal monitoring data,such as video,images,remote sensing images,and trajectory data.These data not only contain rich time,space and semantic information,but also have multi-modal characteristics,which provides the possibility of supporting deep cross-modal analysis,mining and application of multi-source spatio-temporal data.However,traditional relational databases are only good at managing structured data.Unstructured spatio-temporal data is usually only stored as multimedia attribute data.It is difficult to support cross-modal query analysis and deep application of unstructured spatio-temporal data.How to consider the time,space and semantic relationship between multi-modal data to achieve efficient storage and distributed management of unstructured multi-source public safety monitoring data has become a key problem to be solved in supporting the application of cross-modal analysis in the field of urban public safety.To this end,this paper is oriented to the application requirements of multi-modal public safety monitoring data,according to the source and characteristics of public safety monitoring data,the data types are analyzed,and the database storage architecture and data storage structure of public safety monitoring data are designed;Taking into account the space,time and semantic characteristics of public safety monitoring data,a unified coding method based on Geohash and Hilbert's multi-level division strategy for spatio-temporal data and spatio-temporal semantic constraints was proposed;a prototype system was built based on MongoDB,and data storage management Based on the function,the prototype system is used to verify the performance of the multi-level partitioning algorithm proposed in this paper.The main research content of the paper includes the following aspects:(1)Storage structure design based on NoSQL database: Aiming at the unstructured and multi-modal characteristics of multi-source public safety monitoring data,taking into account the space-time and semantic correlation between multi-modal data,multi-source public safety monitoring is designed based on NoSQL database MongoDB The unstructured key-value data storage structure of the data provides data structure support for cross-modal analysis applications that support multi-source public safety monitoring data;(2)Spatio-temporal semantic constrained distributed storage strategy: Aiming at the massive characteristics of multi-source public safety monitoring data,from the perspective of time,space and semantics,study the distributed storage strategy of multi-source public safety monitoring data,and establish based on Geohash and Hilbert The multi-level division strategy of spatio-temporal data combined with the curve and the unified coding method of spatio-temporal semantic constraints realize the efficient and distributed storage of massive multi-source public safety monitoring data,and provide data storage for multi-modal public safety monitoring data across modal management,analysis and application Technical Support;(3)Construction and experimental verification of the prototype database system based on MongoDB: designing a multi-source public safety monitoring database prototype system based on the NoSQL database MongoDB,implementing the database storage management function based on the MongoDB prototype experimental system,and verifying the combination of Geohash and Hilbert proposed in this paper The multi-level data partitioning method and the effectiveness of the unified coding method of spatiotemporal semantic constraints,analyze and evaluate the efficiency,data integrity and load balance of the distributed storage strategy of spatiotemporal semantic constraints.