GOAT FLOCK SURVEILLANCE: A VIDEO ANALYTICS FRAMEWORK USING QUADCOPTER

Abstract

In the present study, goat flock surveillance algorithm using video analytics was determined. The surveillance video camera was mounted over a quadcopter camera, which captured the videos of flocks. A video analytics algorithm using Haar features and the Ada Boost classifier was performed. The technique for tracking of flocks is based on the Kanade-Lucas-Tomasi feature (KLT) tracker. The algorithm presented here allows goat flock surveillance without human guidance and was helpful in real-time monitoring and management of goat flocks. The proposed algorithm made monitoring of goat flocks economical and could be applied to any kind of animal flock in different environments. The qualitative and quantitative analysis carried out for successful goat detection demonstrated the efficiency of the proposed algorithm as compared to the state-of-the-art goat flock surveillance and detection algorithms. The proposed algorithm successfully tracked the goat with accuracy of 93%.