We propose a bottom-up approach for the instance segmentation of cables (commonly referred in the literature as deformable linear objects). While the state of the art instance segmentation techniques ...propose a bounding box and perform foreground segmentation within each proposed bounding box, we adopt a bottom-up approach as cables can span a considerable part of the image or even the entire image, and therefore, cannot be well localized in a bounding box. In this paper, we show that several operations in the top-down instance segmentation approaches are only applicable for certain classes (i.e., compact objects) such as cars but they are a poor approximation for objects with highly overlapping bounding boxes such as cables. In particular, the non-maximum suppression and RoIPool/RoIAlign operations limit the generalizability of proposal-based instance segmentation methods to such datasets. Furthermore, we introduce a synthetic data generation technique that can also be applied to other popular public datasets such as COCO, Pascal VOC, and Cityscapes.
The manipulation of Deformable Linear Objects (DLOs) is a critical process in which the introduction of automation and autonomous systems is still marginal. In this paper, a novel teleoperation ...framework is proposed in which an intuitive manipulation of DLOs is achieved by means of visual aid. The proposed system could be deployed for manipulating DLOs in hazardous scenarios or for simplifying robot teaching tasks by allowing a faster demonstration time. Experiments are performed involving several subjects and their feedback is collected by means of a survey. The results show that the proposed teleoperation framework simplifies DLOs manipulation and reduces the task completion time by 20% on average.
A planning method for linear object manipulation including knotting/unknotting in the three-dimensional space is proposed. Firstly, topological states of a linear object are represented as sequences ...of crossing points including the crossing type at each crossing point. Secondly, transitions among crossing states are defined. They correspond to operations that change the number of crossing points or permute their sequence. Then, we can generate possible sequences of crossing state transitions, that is, possible manipulation processes from the initial state to a given objective state. Thirdly, a method for determination of actions, that is, grasping points and their moving direction, is proposed in order to realize derived manipulation processes. Furthermore, criteria for evaluation of manipulation processes are introduced in order to reduce the candidates of manipulation plans. Finally, it is demonstrated that our developed system based on the above method can generate and execute manipulation plans for untying an overhand knot.
The analysis of convergence of adaptive regularized generalized identification algorithm under measurement noise is presented. The application of the regularized generalized identification algorithm ...provides and improves the computational stability of the procedure for estimating the parameters of neuro-fuzzy models.
Three-dimensional dynamic modeling and numerical simulation of manipulating deformable linear objects (DLOs) such as hoses, wire and cables are studied in this paper. Several coordinate systems are ...introduced first to describe the deformation of DLOs. The dynamic three-dimensional deformation of an inextensible linear object is formulated using finite element method (FEM) and lagrange motion equations. The proposed method provides more efficient simulation of dynamic manipulating DLOs with faster computation speed. Two simulation examples are illustrated to confirm the effectiveness of our methods. One is a comparison simulation study and the other shows the three-dimensional dynamic character of the DLO.
In map generalization, it is important to simplify closed linear objects. In fact, an area object can be represented by its boundary in vector, such as lakes, land use, etc. So shape simplification ...of area objects can make full use of the algorithms for simplifying closed linear objects. According to the purpose of simplification for boundary line of land use, the standards on map production and basic principles of ant colony optimization algorithm, the simplification of closed linear objects is explained as a kind of combinatorial optimization problem. The algorithm of automated simplification using ant colony optimization algorithm is put forward and the key steps are given. The algorithm not only sets up objective function, constraints, heuristic information, pheromones and so on, but also introduces long-term taboo list with the idea of tabu search algorithm is merged into the simplification of linear objects. In order to demonstrate that the proposed algorithm to the automatic simplification of closed linear objects is feasible and effective, the algorithm will be tested for boundary line of forest land, and the results will be compared with the Douglas algorithm.
The presence of Deformable Linear Objects (DLOs) such as wires, cables or ropes in our everyday life is massive. However, the applicability of robotic solutions to DLOs is still marginal due to the ...many challenges involved in their perception. In this letter, a methodology to generate datasets from a mixture of synthetic and real samples for the training of DLOs segmentation approaches is thus presented. The method is composed of two steps. First, key-points along a real-world DLO are labeled by employing a VR tracker operated by a user. Second, synthetic and real-world datasets are mixed for the training of semantic and instance segmentation deep learning algorithms to study the benefit of real-world data in DLOs segmentation. To validate this method a user study and a parameter study are conducted. The results show that the VR tracker labeling is usable as other labeling techniques but reduces the number of clicks. Moreover, mixing real-world and synthetic DLOs data can improve the IoU score of a semantic segmentation algorithm by circa 5%. Therefore, this work demonstrates that labeling real-world data via a VR tracker can be done quickly and, if the real-world data are mixed with synthetic data, the performances of segmentation algorithms for DLOs can be improved.
Validating DLO Models from Shape Observation Palli, Gianluca; Pirozzi, Salvatore
2021 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM),
2021-July-12
Conference Proceeding
Odprti dostop
In this paper, the problem of fitting the model of deformable linear objects from the observation of the shape under the effect of known external forces like gravity is taken into account. The model ...of the deformable linear object is based on dynamic splines, allowing to obtain a reliable prediction of the object behavior while preserving a suitable efficiency and simplicity of the model. The object shape is measured by means of a calibrated vision system, and a fitting between the observed shape and the theoretical model is defined for validation. Experiments are executed in different conditions, showing the reliability of the proposed spline-based model.
This letter presents DLO3DS , an approach for the 3D shapes estimation and tracking of Deformable Linear Objects (DLOs) such as cables, wires or plastic hoses, using a cheap and compact 2D vision ...sensor mounted on the robot end-effector. DLO3DS can be applied in all those scenarios in which the perception and manipulation of DLO-like structures are needed, such as in the case of switchgear cabling, wiring harness manufacturing and assembly in the automotive and aerospace industries, or production of hoses for medical applications. The developed procedure is based on a pipeline that first processes the images coming from the 2D camera extracting key topological points along the DLOs. These points are then used to model each DLO with a B-spline curve. Finally, the set of splines obtained from all the images is matched by exploiting a multi-view stereo-based algorithm. DLO3DS is validated both on a real scenario and on simulated data obtained by exploiting a rendering engine for photo-realistic images. In this way, reliable ground-truth data are retrieved and utilized for assessing the estimation error achievable by DLO3DS , which on the employed test set is characterized by a mean reconstruction error of 0.82 mm.