Top-Level Design of Intelligent Video Surveillance System for Abandoned Objects

摘要

In the area of public safety and video surveillance, it is particularly important to monitor suspicious abandoned objects and pedestrian abandonment in public places. By analyzing the characteristics of abandonment behavior, this paper put forward the theory of life cycle of abandoned objects, which describes the feature of abandoned objects and abandonment behaviors overtime, and proposed the top-level architecture and workflow of the abandoned objects intelligent monitoring system based on this theory. The system tracks pedestrians entering the video scene, captures offloading behaviors, extracts motion features and outputs suspicious images. It also captures static images of abandoned objects through video surveillance, extract image feature and monitor object carriers. Finally, images, videos and other structured data is collected and as the input of a multi-characteristic risk assessment model, which outputs the suspicious level of the abandonment event. This paper fills the blank in the theoretical research on abandonment behavior and gives solutions to the shortcomings of traditional remnant monitoring systems.