Fast 3D feature extraction method for moving stereo

This paper presents a method of extracting 3D feature information using a pinhole camera model. We use image pairs which are obtained by a single camera mounted on a robot that is moving through a simulated environment. We compare the computation time-result of singular value decomposition (SVD) and proposed method. We tested the result of a computation time on simulation as the camera moved. This proposed method reduces the computation time with preprocessing, and the reduction of computation time increases as the number of features and motions increase. This proposed method can be used to observe the environment of robot in real time.