Study on the Indoor SLAM Using Kinect

Abstract

The simultaneous localization and mapping (SLAM) using kinect is examined. Building map is very useful to avoid obstacles and probed object. In order to solve the rapidly SLAM problem, we focus on the vision data and depth data(: RGB-Ddata) from Kinect that is a 3D sensor provided by Microsoft. The proposed method detects some landmarks extracted from images. Depth data of landmark points is used for estimation of consecutive frame status (rotation matrix, translation vector) by ICP algorithm. Putting detected landmark points into Smirnov-Grubbs test to remove outliers, the accuracy is improved. In this paper, we report the results obtained by actual environment.