Institute of Information Science, Academia Sinica

In this talk, I will present a practical vision-based robotic bin-picking system that performs detection and 3D pose estimation of objects in an unstructured bin using a novel camera design, picks up parts from the bin, and performs error detection and pose correction while the part is in the gripper. Two main innovations enable the system to achieve real-time robust and accurate operation. First, a multi-flash camera is used to extract robust depth edges. Second, an efficient shape-matching algorithm called fast directional chamfer matching (FDCM) is introduced to reliably detect objects and estimate their poses. FDCM improves the accuracy of chamfer matching by including edge orientation. It also achieves massive improvements in matching speed using line-segment approximations of edges, a 3D distance transform, and directional integral images. Experiment results show that these speedups, combined with the use of bounds in the spatial and hypothesis domains, give the algorithm sublinear time computational complexity. In addition to significantly improving upon the accuracy of previous chamfer matching methods in all of the evaluated applications, FDCM is up to two orders of magnitude faster than the previous methods.

Ming-Yu Liu received the BE degree from the National Chiao Tung University, Hsinchu, Taiwan in 2003 and the MS and PhD degrees in Electrical Engineering from the University of Maryland College Park, Maryland, USA in 2010 and 2012. He is now a visiting member of research staff at Mitsubishi Electric Research Labs (MERL), Cambridge, Massachusetts, USA. His research interests are in computer vision, machine learning, and robotics. He was on the program committee of several international conferences including CVPR, ICCV, and ECCV. He is a member of the IEEE.