Static object detection for video surveillance

Abstract

One essential component of security systems at public locations, such as airports, bus stops, train stations, marketplaces, etc., is the video surveillance. More powerful and efficient automated technical developments are needed for video surveillance. When an unattended object is left in the public places it will be considered as suspicious object as the terrorist assaults have escalated globally in recent years. The people in public areas must be protected from this attack by using safety precautions. Complex surveillance recordings make it difficult to identify abandoned or removed objects due to a number of factors, such as occlusion, abrupt changes in lighting, and so on. A novel approach is proposed in this article for the identification and classification of a static object in a public place. The main aim of this work is the automatic detection of abandoned objects. This method consists of two steps: static item detection using background subtraction and motion estimation; and (ii) abandoned luggage recognition using convolutional neural networks. (CNN). By applying the background subtraction method using fuzzy integral, the suggested method extracts foreground items. Afterwards the static object is detected using hierarchical Finite state machine (FSM). And finally, the object is classified using the CNN algorithm Yolo V5. In terms of accuracy, precision, and recall, the performance of the suggested algorithm is compared with that of traditional approaches.