1,721,020 research outputs found
Remote Sensing of Target Object Detection and Identification II
Full text linkThe ability to detect and identify target objects from remote images and acquisitions is paramount in remote sensing systems for the proper analysis of territories [...
A Smart Monitoring System for Automatic Welding Defect Detection
No full textThis work introduces an intelligent system able to perform quality control assessment in an industrial production line. Deep learning techniques have been employed and proved successful in a real application for the inspection of welding defects on an assembly line of fuel injectors. Starting from state-of-the-art deep architectures and using the transfer learning technique, it has been possible to train a network with about 7 millions parameters using a reduced number of injectors images, obtaining an accuracy of 97,22%. The system has also been configured in order to exploit new data, collected during operation, to extend the existing dataset and to improve further its performance. The developed system showed that deep neural networks can successfully perform quality inspection tasks which are usually demanded to humans
Dynamics Modeling of an Encountered Haptic Interface for Ball Catching and Impact Tasks Simulation
No full textThis paper deals with a model-based control strategy implemented on an encountered haptic interface developed for the simulation of ball catching tasks. A dynamical model of a reference device has been developed and validated by experimental results. This model was applied to increase the control performance and to simulate realistic impacts. The control strategy to generate the haptic interface trajectories consistent with the simulation of ballistic motion of virtual objects has been defined. At the impact instant the perceptively correct kinetic energy is transferred from the device end-effector to the user hand adopting a velocity scaling rule. Experimental results confirm control accuracy in fast dynamics trajectory tracking
A wireless haptic data suit for controlling humanoid robots
No full textIn this paper we present a novel wearable suit for haptic feedback capabilities at user's hands combined with upper body motion tracking. In the work we present both the system design and the algorithms used for motion tracking and haptic rendering. The overall system was applied to the co-located tele-operation of the Baxter research robot to perform manipulative tasks usually carried out by human personnel in the industry
Haptic Rendering of Juggling with Encountered Haptic Interfaces.
No full textHaptic interaction in a virtual world can be tool mediated or direct; and, among direct interactions, the encountered haptic interfaces provide physical contact only when there is contact with a virtual object. This paper deals with the haptic rendering of the catching and throwing of objects by means of this type of interface. A general model for the rendering of the impact is discussed with the associated formalism for managing multiple objects and multiple devices. Next, a key parameter for simulating the impact is selected by means of a psychophysical test. Finally, a working system is presented with the application of the rendering strategy to the case of haptic juggling, showing the possibility of effectively performing basic juggling patterns with two balls
CEPB dataset: a photorealistic dataset to foster the research on bin picking in cluttered environments
No full textSeveral datasets have been proposed in the literature, focusing on object detection and pose estimation. The majority of them are interested in recognizing isolated objects or the pose of objects in well-organized scenarios. This work introduces a novel dataset that aims to stress vision algorithms in the difficult task of object detection and pose estimation in highly cluttered scenes concerning the specific case of bin picking for the Cluttered Environment Picking Benchmark (CEPB). The dataset provides about 1.5M virtually generated photo-realistic images (RGB + depth + normals + segmentation) of 50K annotated cluttered scenes mixing rigid, soft, and deformable objects of varying sizes used in existing robotic picking benchmarks together with their 3D models (40 objects). Such images include three different camera positions, three light conditions, and multiple High Dynamic Range Imaging (HDRI) maps for domain randomization purposes. The annotations contain the 2D and 3D bounding boxes of the involved objects, the centroids’ poses (translation + quaternion), and the visibility percentage of the objects’ surfaces. Nearly 10K separated object images are presented to perform simple tests and compare them with more complex cluttered scenarios tests. A baseline performed with the DOPE neural network is reported to highlight the challenges introduced by the novel dataset
Autonomous exploration of indoor environments with a micro-aerial vehicle
No full textThe paper presents a system designed for micro-aerial-vehicles capable of autonomously explore an indoor environment, detect objects in the environment and build a map of the environment structure with reference to objects locations within the map. The found objects are saved in an internal database containing all previously recognized objects. The system allows fast exploration time and it is characterized by lightweight computation algorithms for the localization, map building and navigation components. The environment map is built as a set of 2D feature map layers where each layer corresponds to an environment floor. For each framework component a simulated testing scenario is presented to evaluate the capabilities of the designed algorithms. The introduced system is efficient from the computational cost point of view and allows fast exploration time that is critical for battery powered systems
Visual navigation of mobile robots for autonomous patrolling of indoor and outdoor areas
No full textIn many applications, robots should be able to move autonomously in semi-structured or unstructured environments. Autonomous robots can be employed for instance in area patrolling tasks in order to perform surveillance of sites. To autonomously navigate in an unknown outdoor scenario, a robot should be able to acquire sensible information about the environment by means of its own sensors and at the same time perform some reasoning to decide where and how to move. In this paper, we present a vision-based solution for the decision making and a behavior based low-level control for the navigation. Three different testing scenarios have been employed to assess the capabilities of the proposed approach: a computer simulated scenario, an indoor test on a real robotic platform and finally an outdoor test in a city park
Autonomous navigation of mobile robots: From basic sensing to problem solving
No full textAutonomous navigation is a complex task that requires both sensing capabilities to react to sudden environmental changes or map the environment and reasoning to schedule the next action to perform. Starting from basic sensing technology used in the majority of mobile robotic systems, the introduction of sensor fusion techniques allows to obtain useful information to solve the localization, mapping and navigation problems. Applications of these methods to achieve specific robot capabilities will be presented starting from object detection and recognition, passing to scene classification and ending with an industrial related application: the visual inspection of industrial facilities by means of a flying vehicle
- …
