Video surveillance in dynamic environment for detecting human and vehicles, is one of the current challenging research area in computer vision. It is a key technology for efficient management of traffic, public safety, military security. CCTV surveillance systems are more common nowadays because of its easy deployment. But the dependency upon human operator to monitor and classify the objects in the video sequence is very costly and inefficient. Thus an intelligent surveillance system which could detect and classify the moving objects are proposed using this system. Detection of moving objects in video stream is the first relevant step in classification. The proposed method uses Gaussian Mixture Model algorithm (GMM) for object detection and a shadow removal method is also used. Thework mainly focused on object classification. Color, texture and gradient methods along with the kNN classifier used for object classification.

Keywords : Visual surveillance, GMM, Shadow removal, Shadow detection, Object classification, LBP, HOG, kNN.