Abstract
In this article, an innovative algorithm for instance segmentation of wires called Ariadne+ is presented. Although vastly present in many manufacturing environments, the perception and manipulation of wires is still an open problem for robotic applications. Wires are deformable linear objects lacking of any specific shape, color, and feature. The proposed approach uses deep learning and standard computer vision techniques aiming at their reliable and time effective instance segmentation. A deep convolutional neural network is employed to generate a binary mask showing where wires are present in the input image, then the graph theory is applied to create the wire paths from the binary mask through an iterative approach that aims to maximize the graph coverage. In addition, the B-Spline model of each instance, useful in manipulation tasks, is provided. The approach has been validated quantitatively and qualitatively using a manually labeled test dataset and by comparing it against the original Ariadne algorithm. The timings performances of the approach have been also analyzed in depth.
Highlights
T HE development of robotic solutions for the manipulation of deformable objects is a topic of interest in several domains of industrial manufacturing, as in the automotive [1], [2], aerospace [3] and textile industries [4] but, for other very diverse fields as robotic surgery [5], [6] and food processing [7]
Deformable Linear Objects (DLOs) are a particular subgroup of deformable objects consisting of wires, cables, strings, ropes and elastic tubes
The main contributions of this work can be summarized as follow: 1) A reliable and time effective complete approach for the instance segmentation of wires in real scenarios; 2) Modelling of the detected wires in terms of B-Splines; 3) An open source implementation of the entire algorithm available online [8]; 4) A comprehensive experimental validation in terms of segmentation capabilities and timings performances of the approach; 5) A comparison against Ariadne [9], the current state-of-the-art segmentation algorithm for DLOs
Summary
T HE development of robotic solutions for the manipulation of deformable objects is a topic of interest in several domains of industrial manufacturing, as in the automotive [1], [2], aerospace [3] and textile industries [4] but, for other very diverse fields as robotic surgery [5], [6] and food processing [7]. Vastly present in every industrial environment, wires and wiring harnesses still represent a problematic task for robotic applications This is the result of few peculiarities embedded in these objects, like not having any specific shape (due to the deformability), nor color, nor any relevant feature that can make them distinguishable with respect to other objects. The main contributions of this work can be summarized as follow: 1) A reliable and time effective complete approach for the instance segmentation of wires in real scenarios; 2) Modelling of the detected wires in terms of B-Splines; 3) An open source implementation of the entire algorithm available online [8]; 4) A comprehensive experimental validation in terms of segmentation capabilities and timings performances of the approach; 5) A comparison against Ariadne [9], the current state-of-the-art segmentation algorithm for DLOs
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
