Moving Object Segmentation using Enhanced Laplacian Thresholding Method for Video Surveillance

Abstract

Identifying moving objects from a video surveillance is a fundamental and critical task in many computer vision applications.Video segmentation is one such application which has been studied for several decades and still remains a difficult problem for the system to automatically and accurately segment moving objects from video sequence with various backgrounds and global motions in real time.This paper proposes a new approach namely ELT (Enhanced Laplacian Thresholding) for real time video segmentation.The aim of this video enhancement method is to improve the visual appearance of the video and future-automated video processing like analysis, detection, segmentation, recognition, and surveillance for traffic, criminal justice systems and other areas that include analysis on large scale.