Tracking of Person using Monocular Vision by Autonomous Navigation Test Bed (ANT)

This paper has the capability to detect the person and follow that person by the surveillance mobile robot (ANT) in outdoor and indoor environment. Monocular based vision tracking technique used in the paper and following method used for surveillance robot is proposed, as it combines face detection, person’s upper body features extraction and Moment calculation modules proposed in this paper. We are extracting torso based color as a key feature after face detection; it gives the features as input to the module named moment calculation which is used to track the target person by calculating the motion parameters. We introduce speech generation module in order to develop interaction between the person and surveillance robot. We also introduce wireless communication in order to send the signal from base station to the remote location to know where ANT is located. Experiment results validate the robust performance of the proposed approach.