Vision-Based Guidance for Tracking Multiple Dynamic Objects

Kashish Dhal, Pritam Karmokar, Animesh Chakravarthy, William J. Beksi

Abstract: In this paper, we introduce a novel vision-based framework for tracking multiple active objects using guidance laws based on a rendezvous cone method. These guidance laws enable an unmanned aircraft system, equipped with a monocular camera, to continuously observe a set of moving objects within the field of view of its sensor. During the multi-object tracking process, we detect and categorize feature point estimators for controlling the occurrence of occlusions in a comprehensive fashion. Furthermore, we extend our open-source simulation environment and perform a series of simulations to show the efficacy of our proposed approach.