Mapping and Localization in 3D Environments Using a 2D Laser Scanner and a Stereo Camera

Abstract

2D laser scanners have been widely used for accomplishing a number of challenging AI and robotics tasks such as mapping of large environments and localization in highly dynamic environments. However, using only one 2D laser scanner could be insufficient and less reliable for accomplishing tasks in 3D environments. The problem could be solved using multiple 2D laser scanners or a 3D laser scanner for performing 3D perception. Unfortunately, the cost of such 3D sensing systems is still too high for enabling AI and robotics applications. In this paper, we propose to use a 2D laser scanner and a stereo camera for accomplishing simultaneous localization and mapping (SLAM) in 3D indoor environments in which the 2D laser scanner is used for SLAM and the stereo camera is used for 3D mapping. The experimental results demonstrate that the proposed system is lower cost yet effective, and the obstacle detection rate is significant improved compares to using one 2D laser scanner for mapping.