2A2-F08 ステレオ視とSIFT特徴点追跡を用いた移動ロボットのための環境地図生成(移動ロボットのための視覚)

概要

論文の詳細を見る
This paper presents the method of generating the environmental map using stereo vision and SIFT(Scale Invariant Feature Transform) feature. The purpose of this research is to make the map to estimate the robot position when moving with a monocular camera. The feature points are tracked based on the distance of SIFT feature vector, and their 3D position are estimated with the Kalman filter. Each feature point on the map is composed of the 3D position with its covariance, some SIFT feature vectors, the camera position and the glance vector when this feature point was ever observed. By using this environmental map, the 2D robot position and orientation can be estimated without any stereo cameras or range sensors.
一般社団法人日本機械学会の論文
2007-05-11