Intelligent Surveillance System Based On Moving Objects Analysis

The moving objects detection is one of the most important parts of the intelligent video surveillance technology. Without accurate moving objects detection, it is very hard to undertake further processing such as tracking and recognition of the moving objects.This paper begins with a brief introduction to object detection and tracking, and a method based on the background subtraction is discussed in more details. Two background update methods are compared:one is based on linear model and the other one is based Gaussian Mixture Model (GMM). A new method for separating moving objects from background is proposed in the paper, and which makes use of the relationship among different connected areas. Simulation result shows that this method can improve the accuracy of moving objects detection. The second part of this paper is focused on the tracking of moving objects. Two different kinds of methods are analyzed, they are Kalman filtering and objects-relevance based methods. The third part is related with the design and implementation of an intelligent video surveillance system. This system has the following functions:moving objects detection, tracking, and counting according to traces. To verify the performance of proposed video surveillance system, simulation has been undertaken and the experimental results have also been analyzed.The software based video surveillance system provides end users with Graphic User Interface to mark the moving objects and show its traces in real time. It can also count the number of moving objects passing through a pre-defined area.