A Novel Inverse Perspective Transformation-based Reactive Navigation Strategy

摘要

This paper describes a new vision-based reactivernnavigation strategy addressed to mobile robots, comprising obstaclerndetection and avoidance. While many reactive vision-basedrnsystems are based uniquely on the computation and analysisrnof quantitative information, the proposed algorithm combinesrna quantitative process with a set of qualitative rules to safelyrnexplore unknown environments. The process includes a newrnfeature classifier based on the Inverse Perspective Transformationrnto discriminate points that belong to an object from points thatrnbelong to the ground, and a qualitative method to determine thernorientation and distance of the object with respect to the camerarnin the world coordinate system.