An ORB based visual SLAM system by RGB-D camera of LeTV

Abstract

This paper presents a visual simultaneously localization and mapping (SLAM) system based on Oriented FAST and Rotated BRIEF (ORB) and the RGB-D camera accessory of LeTV. Frame matching is performed through ORB feature and random sample and consensus (RANSAC) algorithm. G2o library is used as graph based pose optimization framework. Alternative key frame generation method is proposed to get increased processing capacity. Dense point cloud map which can be used for following intelligent robot related applications is generated from the depth information captured with color image simultaneously by the calibrated RGB-D camera.