SLAM-Based Automatic Extrinsic Calibration of a Multi-Camera Rig

We present a general method for fully automatic extrinsic auto-calibration of a fixed multi camera rig, with no requirement for calibration patterns or other infrastructure, which works even in the case where the cameras have completely non-overlapping views. The robot is placed in a natural environment and makes a set of programmed movements including a full horizontal rotation and captures a synchronized image sequence from each camera. These sequences are processed individually with a monocular visual SLAM algorithm. The resulting maps are matched and fused robustly based on corresponding invariant features, and then all estimates are optimised full joint bundle adjustment, where we constrain the relative poses of the cameras to be fixed. We present results showing accurate performance of the method for various two and four camera configurations. Download PDF




      Lightweight SLAM and Navigation with a Multi-Camera Rig

An interesting recent branch of SLAM algorithms using vision has taken an appealing approach which can be characterised as simple, robust and lightweight when compared to the more established and complex geometrical methods. These lightweight approaches typically comprise mechanical odometry or simple visual odometry for local motion estimation; appearance-based loop closure detection using either whole image statistics or invariant feature matching; and some type of efficient pose graph relaxation. However, such algorithms have so far been proven only as localisation systems, since they have not offered the semantic demarcation of free space and obstacles necessary to guide fully autonomous navigation and exploration. In this paper we investigate how to apply and augment such approaches with other lightweight techniques to permit fully autonomous navigation and exploration around a large and complicated room environment.











.....................................................................................................................................
Gerardo Carrera M