Improved simultaneous localization and mapping by stereo camera and SURF

Abstract

Simultaneous Localization and Mapping (SLAM) has been an active area of research lately. MonoSLAM is a SLAM frame working with a single camera without odometry information for executing SLAM The camera motion estimation and incremental map building from new landmarks are computed using the Extended Kalman Filter framework In this paper, we propose a revised approach to improve the speed of executing SLAM. We use a stereo camera to acquire two images of the environment at the same time, and use SURF algorithm to detect features in the images. The distance between the landmark and the camera can be calculated by the pair of images using a revised algorithm. Using this approach, we can reduce the time of distance calculation and increase the SLAM execution speed.