Vigilance and surveillance reinforced using mathematical approaches in object tracking techniques

Abstract

Visual tracking is crucial to the study of object recognition and has been utilized in a variety of realistic settings, such as robotics, traffic monitoring, self-driving automobiles, forensics, and more. This research concentrates on techniques for counting the total number of individuals entering or exiting a space under the watchful eye of a camera. The techniques described here can detect the number of persons in a scene, both for a single individual and for many passers in front of the camera. With the aid of surveillance that use the centroid concept, an effective solution has been devised for monitoring. Secondly, in this study, object tracking methods utilising deep learning are also reviewed and analysed. This study also compares the effectiveness of various algorithms on the LaSOT, VOT2015, VOT2016, VOT2017 and OTB2015 tests.