6DOF SLAM with Stereo-in-hand

In this paper we describe a system that carries out SLAM using a stereo pair moving with 6DOF as the only sensor. Textured point features are extracted from the images and stored as 3D points if seen in both images with sufficient disparity, or stored as inverse 3D points otherwise. This allows the system to make use of both near and far features that provide distance and orientation, or orientation information, respectively. Unlike other vision only SLAM systems, stereo does not suffer from ’scale drift’ because of unobservability problems, and thus no other information such as gyroscopes or accelerometers is required. Our SLAM algorithm generates sequences of conditionally independent local maps that can share information related to the camera motion and common features being tracked. The system computes the full map using the Divide and Conquer algorithm adapted for conditionally independent local maps, allowing linear time execution. We show experimental results in outdoor urban environments that demonstrate the robustness and scalability of our system.