Research And Implementation On Key Technologies Of Intelligent Surveillance System Of Object Safty Detection

Intelligent video surveillance technology is an emerging and hot field of computer vision. The system is significant in practice in military and civil. For the security applications of intelligent video surveillance, the paper focuses on moving target tracking and Stolen/Abandoned objects detection two fields.(1)For tracking moving object, we propose a target tracking algorithm combines mean-shift and particle filter in this paper. The switching condition of both methods is the value of the similarity between the object template and the test template in current frame. Experimental results show that it is robust to noise and improves the performance in real-time.(2)For detecting Stolen and Abandoned Objects, At first, we propose a novel approach to extract the moving object in the dynamic background, which is based on the mixed Gaussian background modeling and background subtraction with temporal different weighted algorithm. Then we infer whether the object is a "stationary object". Afterwards we employ snake active contour model to extract precise contours of the object.This paper discusses two algorithms used to detect stolen and abandoned objects and based on the method of similar shapes and colors. The fusion of these data from these methods can be regarded as the final result in the decision. The experimental results show that, the proposed methods can detect and distinguish correctly the stolen and abandoned objects.