Image Based Solution to Occlusion Problem for Multiple Robots Navigation

• Main Authors: Taj Mohammad Khan, Muhamad AhmedChoudhry
• Format: Article
• Language: English
• Published: Mehran University of Engineering and Technology 2012-04-01
• Series: Mehran University Research Journal of Engineering and Technology
• Subjects: Camera Modelling; Fusion of Data from Multiple Vision Sensors; Detection; Occlusion Detection; Mapping; Navigation.
• Online Access: http://publications.muet.edu.pk/research_papers/pdf/pdf624.pdf

In machine vision, occlusions problem is always a challenging issue in image based mapping and navigation tasks. This paper presents a multiple view vision based algorithm for the development of occlusion-free map of the indoor environment. The map is assumed to be utilized by the mobile robots within the workspace. It has wide range of applications, including mobile robot path planning and navigation, access control in restricted areas, and surveillance systems. We used wall mounted fixed camera system. After intensity adjustment and background subtraction of the synchronously captured images, the image registration was performed. We applied our algorithm on the registered images to resolve the occlusion problem. This technique works well even in the existence of total occlusion for a longer period.