Real-time Object Detection and Grasping Using Background Subtraction in an Industrial Scenario

Abstract

Grasping partially known objects in unstructured environments is one of the most challenging issues in robotics. In this paper, we present a real-time and robust approach for detecting and grasping different objects with a robot manipulator in a partially unstructured scenario. The proposed method is based on two steps: 1) the generation of a background model to localize the objects of interest and 2) the use of depth information to find the grasp pose. Quantitative experiments using a 7 degrees-of-freedom manipulator on different objects demonstrates the effectiveness of the proposed approach.