In this paper, an approach for fast and accurate segmentation of Deformable Linear Objects (DLOs) named FASTDLO is presented. A deep convolutional neural network is employed for background segmentation, generating a binary mask that isolates DLOs in the image. Thereafter, the obtained mask is processed with a skeletonization algorithm and the intersections between different DLOs are solved with a similarity-based network. Apart from the usual pixel-wise color-mapped image, FASTDLO also describes each DLO instance with a sequence of 2D coordinates, enabling the possibility of modeling the DLO instances with splines curves, for example. Synthetically generated data are exploited for the training of the data-driven methods, avoiding expensive collection and annotations of real data. FASTDLO is experimentally compared against both a DLO-specific approach and general-purpose deep learning instance segmentation models, achieving better overall performances and a processing rate higher than 20 FPS.
This paper describes the development of a system for a wiring operation, composed by a robotic manipulator and a gripper with tactile sensors. The manipulation of electrical wires, and more in general deformable linear objects, is a challenging task of large interest in many industrial applications. The solution studied, is designed to shape a cable along a given path with fixing points and obstacles exploited to shape the cable itself. The wire position is determined starting from a list of point and then converted to a joint reference pose for the manipulator. Moreover, tension of the wire is controlled on the basis of both the estimation of the robot external wrench and the tactile data by acting on the gripper finger opening. An experimental setup in which a cable must be routed along two linear paths connected by a turn and with four fixing points has been used to validate the proposed solution.
In this paper, the estimation of the 3D shape of a deformable linear object such as an electric cable for the purpose of planning suitable grasping poses is proposed. These poses can be then exploited by a robot manipulator to grasp and manipulate the cable for assembling and manufacturing purposes. The proposed method is based on a previously developed algorithm called ARIADNE providing the segmentation and spline modelling on a single image of electric cables even in cluttered scenarios. By exploiting this result and by collecting measurements from different points of view by means of a 2D camera mounted on the robot end effector, a method to estimate the 3D shape of the cable is here presented. This method is experimentally evaluated in different scenarios showing the capability of providing reliable estimation of the cable shape. The result is then used to drive to robot to grasp the cable in controlled positions.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.