An eye-hand system for automated paper recycling

This paper presents an robotic system used for detection, sorting and grading of paper objects as a parr of an automated paper recycling system. The system uses stereo vision to estimate in real time the spatial position of moving objects on a conveyor to be picked by the robot arm. It was proved in the literature that stereo vision may be successfully used to estimate position of moving objects, with a priory known shape, and generate commands to a robot arm for picking up the object. A special computing architecture is introduced for performing the task in real time. A vector of geometrical and textural features is used for sorting the paper, according to its grade. The grading process uses color vision and additional contact ultrasonic sensors. A new data fusion paradigm, based on supervised learning, is used. The data fusion algorithm combines stereo vision and ultrasonic sensors to detect and grade paper objects.