Robotic grasping based on machine vision and SVM

Abstract

Robotic grasping is one of the hot research topics in the field of robotics, which has widely used in many industrial applications. In this study, we proposed a method for robotic grasping based on machine vision and support vector machine algorithm (SVM). A robotic grasping system with a Baxter robot and a Kinect sensor was taken as the research platform. The calibration of eye-to-hand system was completed based on ArUco marker. The grasping rectangle of object was determined based on SVM algorithm. The proposed method was evaluated through experiments by grasping various shape objects. Experimental results showed that the Baxter robot can grasp various shape objects.