Object Detection and Tracking Approaches for Video Surveillance Over Camera Network

Abstract

Object detection and tracking are the most challenging part of any computer vision applications. In computer vision, video surveillance is a popular research area in a dynamic environment, particularly for security reasons. The video surveillance technology plays a crucial role to prevent crime, terrorism etc. The video outputs are filtered and processed by human operators and in case of a forensic, the high volume of data made it difficult to track any object. This work has been done with an aim to reduce the effort of human operators with an increase in the response time to forensic events. It involves designing of an efficient object tracking system for simple environments where the camera is static, background is simple and no similar object to the one being tracked is present. The system is provided with network configuration of the cameras and roads of the surveillance area, videodumps and an image of the object to be tracked. It tracks the objects through the videos and dumps the tracked portions of the videos where the object was present.