An abandoned and removed object detection algorithm in a reactive smart surveillance system

Abstract

Important issues such as low image quality and human operators' reactivity can seriously reduce the effectiveness of video surveillance systems. Digital video surveillance is used in conjunction with video analytics, which is the semantic analysis of video data through signal and image processing techniques. This paper presents a novel algorithm for real time detection of abandoned and removed objects, in the context of a reactive smart surveillance system, which notifies the occurrence of such events of interest to registered users, within seconds. The proposed algorithm is demonstrated to be completely adaptive and computationally efficient.