Unknown object extraction for robot partner using depth sensor

Abstract

This paper describes an object extraction method based on plane detection to extract an unknown object for service robots that use a depth sensor. Recently, depth sensors are used to perceive 3D space in an environment. In robot perception, a depth sensor have used for perceiving unknown environment, such as surface reconstruction, model fitting and so on. Point Cloud Library is famous open source library to deal with 3D point cloud data. However, robot perception for grasping have limitations with high computational costs and low-accuracy for perceiving small objects. Therefore, we proposed the PSO-based plane detection method with RG and the object extraction method based on geometric invariance. To verify accuracy and computational cost for unknown object extraction, we have compared the proposed method with PCL. As an experimental result, we show that the proposed method has higher accuracy and less computational cost drastically for an unknown object extraction.