ABSTRACT
Automated intrusion detection in the warehouse is a more challenging task in real-time era. The algorithms available in literature are suitable for stored videos but not for real-time videos. Very few algorithms are work in real-time but they took more computation power so that is not adapted by cyber-physics hardware. In this chapter, some video analytics techniques are discussed, compared the results, and explained suitable selection of algorithm for an application. In video surveillance, all algorithms are not suite for all application. Most of the algorithms are environment specific which means algorithm design for a specific environment and it acts best in that situation. Some algorithms are work better on most of the environment in the change detection dataset 2014. The dataset consists of 12 categories of videos and each category has more videos for evaluation of the proposed algorithm. The universal algorithm for motion detection is a visual background extractor. It gives the best result in most of the videos in the change 30detection dataset 2014. And also some cyber-physical hardware with their use and how to interface with algorithms has been discussed in this chapter.
