Sequence-based autonomous robot navigation by log-polar image transform and epipolar geometry

Abstract

A sequence-based visual navigation system for autonomous robots is proposed. A sequence of reference images representing a transportation route is taken as the navigation clues. When a robot navigates in a route, two navigation modes, with better reduction in total transportation time required, are switched according to the difference of scales between the robot's current view and the targeted reference image. When the scale is large, log-polar transform is applied to match regions in two images. A less accurate moving direction is inferred from the matching result. After the difference of scale is reduced when the robot approaches the targeted waypoint, SIFT features are matched between two images. An accurate moving direction is decided from the epipolar geometry which is computed from matched SIFT features. The robot's moving velocity is adjusted accordingly, thus reducing the overall transportation time. While other sequence-based navigation methods moves the robot at a maximal velocity of 10 cm/sec, the proposed method moves the robot in an average velocity of 26.5 cm/sec during a 117.3 m long route with 11 reference images. Comparing to a similar sequence-based navigation research, the proposed method does not only reduce transportation time by at least half, but also requires less number of reference images in route transportation applications.