Contour Tracking of Soft Sheet Materials Using Local Contour Image Data

Technologies related to the handling of cloth products still rely on human hands in the apparel industry and in the fields of medicine, welfare, and nursing care; hence, the automation of these technologies is strongly desired. However, the cloth material in question is characteristically non-linear and anisotropic, so it is not easy to predict its dynamic deformation. Therefore, no guidelines for integrated methods of using robots to handle cloth material have been established. In this paper, we propose a method of tracking the contour of a piece of cloth that has curved edges. The aim of the method is to spread soft, flat objects of different shapes, such as clothes. We show contour extraction performed by a camera mounted on a fingertip of a robot hand and operational plan of a manipulator on the basis of the extracted information. The effectiveness of the method is demonstrated by extracting the contour of a piece of patterned cloth, controlling the hand direction and the depth of grasp, and tracking the contour of a piece of cloth with partly curved edges.