Anomalous Crowd Behavior Detection Using Raspberri PI

摘要

Anomalous crowd behavior detection with proper accuracy is an important aspect all over the world. To detect the abnormal activity, suspicious persons, suspicious objects, any accident, bomb explosion in crowded area like station, fair, bus stand and like others are huge concern for us. The aim of this paper is the fast detection of an anomalous event by the use of TensorFlow algorithm in python software and after detecting the anomalous event there is a raspberry pi hardware interfacing which will create an alarming sound to aware everyone about the unexpected situation. This system can be affordable by all easily. To calculate the accuracy and sensitivity of a given video, PETS2009 datasets have been used in this paper. The proposed work finds out a reasonably good accuracy and high sensitivity.