摘要:Our fast-paced, constantly changing world produces massive amounts of data in the 4V categories (volume, velocity, variety and value), which present significant challenges in geo-referenced data management. Virtual geographic environment (VGE), being a human-centred environment, plays a significant role in geographical analysis and simulation. Because of the complexity of geo-processing, time constraints and the large volume of real-time data access, linking calculations and storage has become an effective solution for combining the steps of real-time access and real-time geo-processing in VGE. However, traditional GIS database engines, such as Static GIS or Temporal GIS, that offer an offline interface for managing static geographic features or time-varying geospatial data typically have difficulty in storing streaming data and querying updated information concurrently, due to the ‘store first, and then calculate’ mode. This paper proposed a novel real-time geo-processing database engine that combines a Disk Resident Database (DRDB) and a Main Memory Database (MMDB) to ensure the efficiency of real-time storage, querying, calculation and updating. The proposed engine consists of three modules that respectively satisfy the need of real-time access, complex calculation and ample storage. To illustrate the proposed engine and its implications, a typical real-time geo-processing application, suspect tracking based on video surveillance in a public security setting is studied in this paper.