Local visual homing by matched-filter descent in image distances

Abstract In natural images, the distance measure between two images taken at different locations rises smoothly with increasing distance between the locations. This fact can be exploited for local visual homing where the task is to reach a goal location which is characterized by a snapshot image: Descending in the image distance will lead the agent to the goal location. To compute an estimate of the spatial gradient in the distance measure, its value has to be sampled at three non-collinear points. An animal or robot would have to insert exploratory movements into its home trajectory to collect these samples. Here we suggest a method based on the matched-filter concept which allows to estimate the gradient without exploratory movements. Two matched filters -- optical flow fields resulting from translatory movements in the horizontal plane -are used to predict two images in perpendicular directions from the current location. We investigate the relation to differential flow methods appli...

• Type: Journal
• Year: 2006
• Where: BC
• Authors: Ralf Möller, Andrew Vardy
• Keywords: BC 2006, Distance Measure, Image, Matched Filter