An active-vision system for the recognition of moving objects

Abstract

A novel active-vision system is presented in this paper for moving-object recognition. The proposed system is a natural extension of the static-object recognition system developed earlier in the authors' laboratory. Pre-marking, via circular markers placed on the objects, and active detection, via a camera mounted on a robot, comprise the two principles of the recognition technique. The primary steps of the image-acquisition and recognition procedure are: elliptical shape fitting to a detected marker's boundary, estimation of this marker's 3D-location and orientation, and object standard-view matching. Following the description of the static case, the moving-object recognition technique is introduced with a special focus on the solution of the orientation-duality problem. Utilizing motion information, the duality problem is solved by the acquisition of two consecutive images. An experimental example is given to demonstrate the feasibility of the overall proposed recognition system.