Vision based robot for recognizing and grasping fast moving conveyor products

This paper presents a vision guided robot for recognizing and grasping moving objects on the conveyor belt. An industrial camera is utilized to capture the image of moving objects. A robust vision system is applied for recognizing and estimating the moving object in real time. Objects are recognized by using SURF(Speeded-Up Robust Features) key point detector and FLANN(Fast Library for Approximate Nearest Neighbor) based matcher. Then Rotating calipers algorithm is adopted to estimate the centroid position and orientation of the objects. A simple path planning scheme will thus be assigned based on the information obtained from vision system. Experimental result demonstrates that the approach is efficient and reliable.