A Simultaneous Estimation Method of Camera Pose and Environment Model using Head Mounted Stereo Vision

Abstract

To realize multipurpose recognition system for various robots, it is necessary to design the recognition system which is independent from robot body. When human equipped the system, we can obtain sensor logs with difficult motion for robots, and apply recognized information from logs to robot navigation. In this paper, we make a head mounted stereo vision sensor system and propose a simultaneous estimation method of camera pose and environment model in real environment. We show some environment models as results.