Moving Object Detection using Segmentation Techniques

Abstract

Moving object detection is the basic step for the analysis of video. It is the monitoring of the behavior, activities or other changing information and aims at background subtraction to locate the object position in video frame. A fast and accurate moving object detection technique is important to detect, recognize and track objects over a sequence of images. This research area has been studied for decades; many techniques have been reported and applied on different video surveillance applications. However, there are still some unsolved problems need to be addressed due to multiple objects present in the scene, whereby we wish to determine the position of the same object across time. These types of tasks require not only good initial object detection but reliable body part segmentation as well. Thresholding has found to be a well-known technique for background subtraction such that pixels labelled corresponds to object and 0 to background. Thresholding is further divide into the global and local thresholding techniques.