期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2012
卷号：1
期号：4
页码：590-599
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Surveillance Cameras has proven to be a key factor in enhancing the public security in many countries around the world . In spite of advancements in image processing and machine vision techniques very less is applied to actual implementation of such surveillance systems. Traditional methods wherein stand alone cameras provide just live feed to the observer are still deployed at most of the places but the need of today's world is to add some intelligence to this systems. In this paper we will be focusing on important features to build an intelligent surveillance system and its architecture using computer vision libraries of opencv. Background Subtraction , segmentation, tracking and human behavior analysis are step by step implemented to build such a system. A detailed analysis of background subtraction methods and their comparison is provided. Lastly we discuss the future work which can lead to human behavior analysis like crowd counting, intrusion detection, loitering , crowd movement , vehicle tracking and all.
