Real-time color-based tracking via a marker interface

Abstract

Presents a general architectural solution for concurrent real-time color-based tracking of multiple objects. The presented architecture, called Cocoa, is suggested for a mobile robot interacting with humans. The main contributions of the architecture are markers and color judges. Markers are a situated representation of the robot's environment. Color judges separate color-based segmentation methods from the process of labeling images. Furthermore, methodology for coping with objects entering and leaving the robot's field of view is suggested. Tracking experiments performed with a system implemented on a Nomad XR4000 robot are presented as well.