Abstract
Bin picking is a challenging problem that involves using a robotic manipulator to remove, one-by-one, a set of objects randomly stacked in a container. In order to provide ground truth data for evaluating heuristic or machine learning perception systems, this paper proposes using simulation to create bin picking environments in which a procedural generation method builds entangled tubes that can have curvatures throughout their length. The output of the simulation is an annotated point cloud, generated by a virtual 3D depth camera, in which the tubes are assigned with unique colors. A general metric based on micro-recall is proposed to compare the accuracy of point cloud annotations with the ground truth. The synthetic data is representative of a high quality 3D scanner, given that the performance of a tube modeling system when given 640 simulated point clouds was similar to the results achieved with real sensor data. Therefore, simulation is a promising technique for the automated evaluation of solutions for bin picking tasks.

Abstract
Bin picking is a highly researched topic, due to the need for automated procedures in industrial environments. A general bin picking system requires a highly structured process, starting with data acquisition, and ending with pose estimation and grasping. A high number of bin picking problems are being presently solved, through deep learning networks, combined with distinct procedures. This study provides a comprehensive review of deep learning approaches, implemented in bin picking problems. Throughout the review are described several approaches and learning methods based on specific domains, such as gripper oriented and object oriented, as well as summarized several methodologies, in order to solve bin picking issues. Furthermore, are introduced current strategies used to simplify particular cases and at last, are presented peculiar means of detecting object poses.

Abstract
Bin picking based on deep learning techniques is a promising approach that can solve several analytical methods problems. These systems can provide accurate solutions to bin picking in cluttered environments, where the scenario is always changing. This article proposes a robust and accurate system for segmenting bin picking objects, employing an easy configuration procedure to adjust the framework according to a specific object. The framework is implemented in Robot Operating System (ROS) and is divided into a detection and segmentation system. The detection system employs Mask R-CNN instance neural network to identify several objects from two dimensions (2D) grayscale images. The segmentation system relies on the point cloud library (PCL), manipulating 3D point cloud data according to the detection results to select particular points of the original point cloud, generating a partial point cloud result. Furthermore, to complete the bin picking system a pose estimation approach based on matching algorithms is employed, such as Iterative Closest Point (ICP). The system was evaluated for two types of objects, knee tube, and triangular wall support, in cluttered environments. It displayed an average precision of 79% for both models, an average recall of 92%, and an average IOU of 89%. As exhibited throughout the article, this system demonstrates high accuracy in cluttered environments with several occlusions for different types of objects.

Abstract
Bin picking is a challenging task involving many research domains within the perception and grasping fields, for which there are no perfect and reliable solutions available that are applicable to a wide range of unstructured and cluttered environments present in industrial factories and logistics centers. This paper contributes with research on the topic of object segmentation in cluttered scenarios, independent of previous object shape knowledge, for textured and textureless objects. In addition, it addresses the demand for extended datasets in deep learning tasks with realistic data. We propose a solution using a Mask R-CNN for 2D object segmentation, trained with real data acquired from a RGB-D sensor and synthetic data generated in Blender, combined with 3D point-cloud segmentation to extract a segmented point cloud belonging to a single object from the bin. Next, it is employed a re-configurable pipeline for 6-DoF object pose estimation, followed by a grasp planner to select a feasible grasp pose. The experimental results show that the object segmentation approach is efficient and accurate in cluttered scenarios with several occlusions. The neural network model was trained with both real and simulated data, enhancing the success rate from the previous classical segmentation, displaying an overall grasping success rate of 87.5%.

Abstract
Human-Robot Collaboration is a critical component of Industry 4.0, contributing to a transition towards more flexible production systems that are quickly adjustable to changing production requirements. This paper aims to increase the natural collaboration level of a robotic engine assembly station by proposing a cognitive system powered by computer vision and deep learning to interpret implicit communication cues of the operator. The proposed system, which is based on a residual convolutional neural network with 34 layers and a long -short term memory recurrent neural network (ResNet-34 + LSTM), obtains assembly context through action recognition of the tasks performed by the operator. The assembly context was then integrated in a collaborative assembly plan capable of autonomously commanding the robot tasks. The proposed model showed a great performance, achieving an accuracy of 96.65% and a temporal mean intersection over union (mIoU) of 94.11% for the action recognition of the considered assembly. Moreover, a task-oriented evaluation showed that the proposed cognitive system was able to leverage the performed human action recognition to command the adequate robot actions with near-perfect accuracy. As such, the proposed system was considered as successful at increasing the natural collaboration level of the considered assembly station.

Abstract
Machine vision systems are widely used in assembly lines for providing sensing abilities to robots to allow them to handle dynamic environments. This paper presents a comparison of 3D sensors for evaluating which one is best suited for usage in a machine vision system for robotic fastening operations within an automotive assembly line. The perception system is necessary for taking into account the position uncertainty that arises from the vehicles being transported in an aerial conveyor. Three sensors with different working principles were compared, namely laser triangulation (SICK TriSpector1030), structured light with sequential stripe patterns (Photoneo PhoXi S) and structured light with infrared speckle pattern (Asus Xtion Pro Live). The accuracy of the sensors was measured by computing the root mean square error (RMSE) of the point cloud registrations between their scans and two types of reference point clouds, namely, CAD files and 3D sensor scans. Overall, the RMSE was lower when using sensor scans, with the SICK TriSpector1030 achieving the best results (0.25 mm +/- 0.03 mm), the Photoneo PhoXi S having the intermediate performance (0.49 mm +/- 0.14 mm) and the Asus Xtion Pro Live obtaining the higher RMSE (1.01 mm +/- 0.11 mm). Considering the use case requirements, the final machine vision system relied on the SICK TriSpector1030 sensor and was integrated with a collaborative robot, which was successfully deployed in an vehicle assembly line, achieving 94% success in 53,400 screwing operations.