Smart Surveillance System

In recent times, it has become very necessary to secure each and every place (public/private), like stations, bus stands, agricultural fields and also other general areas from intruders. For this video surveillance is playing a very important role to secure the areas. In our project, we present an intelligent framework to detect the multiple events in the region of interest with this video surveillance. From the region of interest we get a set of variables, create classes like humans or vehicles, attributes of the classes like speed or locality, and create notions to detect and understand the activities in the circumstances. The objects may be pet, human or any vehicle and are detected on the basis of ratio of height and width. After calculation if objects falls into range then a boundary box is set around that object and a signal is given to the users. If this intruder system is used in the house then an additional algorithm can be set to check if a familiar face is detected, in this case system does not consider the person as intruder. Intruders are found out and marked by a boundary box or colored red in the frames given by video of CCTV or any camera footage.