An image-based visual threat cue for autonomous navigation

Abstract

The paper presents a new visual motion cue which we call the visual threat cue (VTC) that provides some measure for a relative change in range as well as clearance between a 3D surface and a fixated observer in motion. This cue is independent of the 3D environment, needs almost no a-priori knowledge about it, is rotation independent, and is measured in [time/sup -1/] units. A practical method to extract the VTC from a sequence of images of a 3D textured surface obtained by a fixated, fixed-focus monocular camera in motion is also presented. This approach of extracting the VTC is independent of the 3D surface texture and needs no optical flow information, 3D reconstruction, segmentation, feature tracking or pre-processing. It can be used in many textured environments and does not depend on the type of texture in the environment. This algorithm to extract the VTC was applied to several real indoor as well as outdoor textures, where we observed a similar behavior for all the textures.