A Comprehensive Survey on Multi Object Tracking Under Occlusion in Aerial Image Sequences

Abstract

Object detection and tracking is a one of the major research topics in the field of computer vision. It is a major component in applications like video surveillance, people monitoring, traffic monitoring etc. Object tracking algorithms typically estimate the location of the target in each frame in a sequence, given initial location of the target as a bounding box. In spite of extensive research reported on tracking, object tracking under occlusion remains a challenge still. There are many practical challenges in object tracking that have to be addressed. This survey paper aims to compile the various techniques available for object tracking under occlusion. Details about the various performance measures available to evaluate the tracking performance under occlusion is also discussed. This paper also points to a list of publicly available datasets that are used to test the tracking algorithms.