Moving object tracking within surveillance video sequences based on EDContours

Abstract

We present a technique to track moving object in a video sequences created by still camera. Contours of each frame is detected using EDContours. Background model obtained with three point estimation technique is created from contour maps of video sequence. Differences between contour map of current frame and background model is our target frame. Post-processing operations are applied to the target frame to reduce noises and moving objects are revealed. Last, objects are inclosed with bounding box. Performance of our work is determined using with bounding boxes, considered as trackers.