Movement Classification Technique for Video Forensic Investigation

Movement classification or activity analysis is one of the most important areas in video surveillance. However, manually detecting, classifying and analyzing interesting moving objects by humans does not guarantee absolute correctness. When considering thereal environment and trying to relate the way objects interact in a surveillance covered area, it is not so easy interpreting every activity correctly. These challenges posed by defining and classifying objects’ behaviours as normal or abnormalmovements. These challengescan be tackled using video analytic technologies. The objective of video analytic technologies is to detect the presence of objects that are moving in its field of view and to classify their movements for security, traffic monitoring and safety applications. There are a lot of hurdles faced by video analytic systems that impede their ability to perform accurately. This study presents a review of movement classification techniques and algorithms, which can tackle the challenges of realistic and practical outdoor surveillance scenarios.