Evaluation of Target Tracking Approaches using Infrared Imagery

Abstract

In this modern world, security is one of the crucial aspects of life either from cross border security or in house security. To secure from the unwanted access, several target detections and tracking algorithms have been proposed and found in the literature for visual imaging. There is number of fields where target tracking or moving object tracking is concerned such as medical, ATM surveillance, border security, quality checking, weather forecasting, defence-security, sea shore security and many more. However, visual imaging is not well capable of dealing with all type of target tracking, thermal imaging or infrared imaging is one of the great options to handle effectively those object where lights are low or near to zero. In this paper, we are focussing on evaluating some of the popular target detection algorithms with the Kalman filter which were mostly used for tracking the object and compared with mean shift filter and blob analysis with Gaussian mixture model on infrared video sequences. We also performed statistical analysis with all five algorithms for real-world benchmark datasets. This analysis will provide great exposure to develop a practical algorithm to achieve tracking the target.