An Effcient, Automatic Abandoned Luggage Detection using DenseNet for Intelligent Video Surveillance Systems

Abstract

Over the past few years, terrorism has been rising in many parts of the world. Most of the terrorist
attacks employ different kinds of bombs and explosives. Compared to the frequently used bombs like
suicide bomb, car bombs, barrel bombs etc., luggage or baggage bombs can be prevented easily using
a surveillance system. The preclusion of such explodes is made possible by detecting and tracking
different luggage and its owner in public areas like airports, railway stations, malls etc. by using the
surveillance camera installed in that area. The method proposed in the paper automatically detects
and tracks abandoned luggage items in the crowded areas using DenseNet. Haar cascade classifier is
used to distinguish human beings from the luggage. Finally the distance between the luggage and its
owner is determined to identify and track the abandoned luggage when the owner walks away from it.
Comparative study of various methods for abandoned luggage detection reveals that the present work
significantly reduces the false alarm rate and improves the overall performance.