Vision Tracking Algorithm for Augmented Reality System of Teleoperation Mobile Robots

Abstract

Teleoperation mobile robot is an effective solution to the operation in unfamiliar, dangerous or unreachable environment in the deep space. With Augmented Reality(AR), teleoperation mobile robots can reach high accuracy and efficiency. To get perfect AR effect, the feature extraction and matching algorithm are achieved based on reconstruction model and k-d tree. The pose estimation used LM algorithm is discussed. A new vision tracking algorithm is built-up based on 3D reconstruction model for AR teleoperation mobile robot. With experiments, it can proved that the new vision tracking algorithm can meet the requirement of teleoperation mobile robot systems in manned deep space exploration.