Video Data Modeling and Querying on Surveillance Videos

To easily access the video content and to manage the huge video archives, powerful video data models are required. Video data model and its supported queries may change the application requirements. In this study, we present a video data model and query types for surveillance videos. Our data model considers video metadata, automatically extracted moving objects, spatial positions of the moving objects, and events of the moving objects. Based on proposed model, we apply semantic, spatial, regional, temporal, size-based and trajectory queries on surveillance videos. Our system allows users to access to query interfaces via Web browsers. With the help of our system, users can reach to surveillance video archives quickly and query the videos from different aspects