Perception of 3-D Objects by a Robotic Stereo Eye-in-Hand Vision System

Abstract

A 3-D machine perception technique is presented using stereo intensity images and an anthropomorphic robot. Multiple stereo views allow the perception of the third dimension by the solution of a correspondence problem defined in two stereo pairs. The zero crossing technique is used to detect edges on the images and to classify them in order to reduce the number of possible solutions. By the eye-in-hand configuration, several stereo pairs of a scene can be taken moving the robot arm. For a limited number of objects, stationary on the robot table, this technique allows the derivation of a set of three-dimensional sample points that correspond to the physical object edges. The technique mainly consist in a geometrical match, in the 3-D space, of some hypothesized solutions formulated on a couple of stereo pairs, taken from two points of view. Considering objects with straight edges and with no highly textured surfaces, some geometrical properties are statistically verified for the most solution points, so the correspondence problem can be robustly solved.