The Proceedings of JSME annual Conference on Robotics and Mechatronics (Robomec)
Online ISSN : 2424-3124
2021
Session ID : 2P2-H09
Conference information

Recognition of linear deformable objects using point cloud data for various objects mixedly placed
*Daisuke KAJIOKAHiromasa ONAKAZhang CHIYuichiro TODATakayuki MATUNOMamoru MINAMI
Author information
CONFERENCE PROCEEDINGS RESTRICTED ACCESS

Details
Abstract

There are many deformable objects such as strings and paper, clothes surround people grows. It is difficult for a robot to manipulate deformable objects, because it needs to deal with various form. So, the demand that a robot manipulates deformable objects is growing. In our previous study, we have successfully operated a string with a one arm or two arms. However, in the previous research, the object placed in the work area of the manipulator is limited to one string, and it is not assumed to perform the knotting operation of the string under the environment in which some objects are mixed. In this research, A method has been proposed that recognizes an object from the point cloud data acquired by the depth camera and determines whether the point cloud is a string based on the variance of the point cloud on the surface.

Content from these authors
© 2021 The Japan Society of Mechanical Engineers
Previous article Next article
feedback
Top