Two-Stage Picking Method for Piled Shiny Objects

In this paper, we propose a novel two-steps algorithm for randomized bin-picking. Since it is difficult to detect the pose of a shiny object randomly piled in a bin, we give up picking objects one by one based on the visual information on the pile. Rather, a robot first roughly picks some of the objects from the pile without using the visual information and roughly place them onto a working table. Then, a robot picks the objects from the working table by detecting their 2D position based on the 2D RGB image. We performed experiments for multiple shiny objects with different shape and weight. Throughout the experimental study, we show that, just by adding one more step of robot motion, we can realize robust bin-picking for shiny objects.