A vison-based grasping strategy for the mineral sorting

In industry application, e.g, mineral sorting, various mineral should be picked from the conveyer. The problem can be regarded as grasping prior unknown objects. In this paper, we aim to present a method to grasp the objects with a three-finger gripper guided by vision. Based on our previous works on attractive region, the way to locate a stable grasp is illustrated in the rotation space of the object. More specifically, the down-hill direction of the attractive region is used to guide the motion of the camera. And then, a grasp planar is proposed to generate hypothetic grasps from the contour of object. The vision and the grasping detection are coupled to reconstruct only few points for contacting, such that, the method can be conducted effectively. At last, the efficiency of the method is illustrated by a simulation.