Fuzzy-based processing of 3D information for items localization in the automated sorting of recyclable packaging

In order to automate the sorting process that precedes the recycling of packaging material, it is necessary to solve the problem of localizing unknown items running on a moving conveyor belt. No a priori knowledge is available on the position and/or the geometrical or optical properties of the objects to be gripped. Furthermore, they are often in contact and/or overlapped. In this paper, we propose to use the height profile of the objects on the conveyor belt, sampled at fixed space intervals through a simple optical device (a source of structured light and a CCD camera), in conjunction with fuzzy techniques, in order to get a possible geometrical description of the items on the band and to allow the computation of the center point of each object as optimal gripping point. Experimental results confirm the validity of the approach