Fast depth evaluation from pattern projection in conjunction with connected component labeling

This paper presents depth evaluation of object detection for automated assembling robots. Pattern distortion analysis from structured light system figures out an object with the highest depth from its background. An automated assembling robot should priory select and pick an object with the highest depth to reduce physical harm during picking action of the robot arm. Object detection is then combined together with depth evaluation so as to provide contour showing edges of an object with the highest depth. The contour provides shape information to an automated assembling robot, which equips laser based proxy sensor, for picking up and placing an object in an intended place.