Results 1 - 10
of
13
Deep Learning for Detecting Robotic Grasps
"... Abstract—We consider the problem of detecting robotic grasps in an RGB-D view of a scene containing objects. In this work, we apply a deep learning approach to solve this problem, which avoids time-consuming hand-design of features. This presents two main challenges. First, we need to evaluate a hug ..."
Abstract
-
Cited by 22 (6 self)
- Add to MetaCart
(Show Context)
Abstract—We consider the problem of detecting robotic grasps in an RGB-D view of a scene containing objects. In this work, we apply a deep learning approach to solve this problem, which avoids time-consuming hand-design of features. This presents two main challenges. First, we need to evaluate a huge number of candidate grasps. In order to make detection fast, as well as robust, we present a two-step cascaded structure with two deep networks, where the top detections from the first are re-evaluated by the second. The first network has fewer features, is faster to run, and can effectively prune out unlikely candidate grasps. The second, with more features, is slower but has to run only on the top few detections. Second, we need to handle multimodal inputs well, for which we present a method to apply structured regularization on the weights based on multimodal group regularization. We demonstrate that our method outperforms the previous state-of-the-art methods in robotic grasp detection. I.
Robust Reconstruction of Indoor Scenes
"... We present an approach to indoor scene reconstruction from RGB-D video. The key idea is to combine geomet-ric registration of scene fragments with robust global opti-mization based on line processes. Geometric registration is error-prone due to sensor noise, which leads to alias-ing of geometric det ..."
Abstract
-
Cited by 2 (0 self)
- Add to MetaCart
(Show Context)
We present an approach to indoor scene reconstruction from RGB-D video. The key idea is to combine geomet-ric registration of scene fragments with robust global opti-mization based on line processes. Geometric registration is error-prone due to sensor noise, which leads to alias-ing of geometric detail and inability to disambiguate dif-ferent surfaces in the scene. The presented optimization ap-proach disables erroneous geometric alignments even when they significantly outnumber correct ones. Experimental re-sults demonstrate that the presented approach substantially increases the accuracy of reconstructed scene models. 1.
Large-scale and drift-free surface reconstruction using online subvolume registration
- In CVPR, 2015. 8
"... Depth cameras have helped commoditize 3D digitization of the real-world. It is now feasible to use a single Kinect-like camera to scan in an entire building or other large-scale scenes. At large scales, however, there is an inherent chal-lenge of dealing with distortions and drift due to accumu-late ..."
Abstract
-
Cited by 1 (0 self)
- Add to MetaCart
(Show Context)
Depth cameras have helped commoditize 3D digitization of the real-world. It is now feasible to use a single Kinect-like camera to scan in an entire building or other large-scale scenes. At large scales, however, there is an inherent chal-lenge of dealing with distortions and drift due to accumu-lated pose estimation errors. Existing techniques suffer from one or more of the following: a) requiring an expensive offline global optimization step taking hours to compute; b) needing a full second pass over the input depth frames to correct for accumulated errors; c) relying on RGB data alongside depth data to optimize poses; or d) requiring the user to create explicit loop closures to allow gross alignment errors to be resolved. In this paper, we present a method that addresses all of these issues. Our method supports online model correction, without needing to reprocess or store any input depth data. Even while performing global correction of a large 3D model, our method takes only minutes rather than hours to compute. Our model does not require any explicit loop closures to be detected and, finally, relies on depth data alone, allowing operation in low-lighting conditions. We show qualitative results on many large scale scenes, high-lighting the lack of error and drift in our reconstructions. We compare to state of the art techniques and demonstrate large-scale dense surface reconstruction “in the dark”, a capability not offered by RGB-D techniques. 1.
Building a grid-point cloud-semantic map based on graph for the navigation
"... of intelligent wheelchair ..."
(Show Context)
Improving Localization Accuracy for an Underwater Robot With a Slow-Sampling Sonar Through Graph Optimization
"... Abstract — This paper proposes a novel localization algorithm for an autonomous underwater vehicle equipped with a mechanical scanning sonar that has a slow frequency of data sampling. The proposed approach incrementally constructs a pose graph and conducts graph optimization to correct the robot po ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract — This paper proposes a novel localization algorithm for an autonomous underwater vehicle equipped with a mechanical scanning sonar that has a slow frequency of data sampling. The proposed approach incrementally constructs a pose graph and conducts graph optimization to correct the robot poses. The construction of a pose graph has three stages: 1) scan generation which incorporates an extended Kalman filter-based dead reckoning algorithm that takes the robot motion into account while eliminating the sonar scan distortion caused by the motion; 2) data association which is based on Mahanalobis distance and shape matching for determining loop closures; and 3) scan matching which calculates constraints constructs pose graph. The constructed pose graph is then fed into a graph optimizer to find the optimal poses corresponding to each scan. A trajectory correction module uses these optimized poses to correct intermediate poses during the process of scan generation. Both simulation and practical experiments are conducted to verify the viability and accuracy of the proposed algorithm. Index Terms — Autonomous underwater vehicle, localization, mechanical scanning sonar, graph optimization.
Dense Continuous-Time Tracking and Mapping with Rolling Shutter RGB-D Cameras
"... We propose a dense continuous-time tracking and map-ping method for RGB-D cameras. We parametrize the camera trajectory using continuous B-splines and optimize the trajectory through dense, direct image alignment. Our method also directly models rolling shutter in both RGB and depth images within th ..."
Abstract
- Add to MetaCart
(Show Context)
We propose a dense continuous-time tracking and map-ping method for RGB-D cameras. We parametrize the camera trajectory using continuous B-splines and optimize the trajectory through dense, direct image alignment. Our method also directly models rolling shutter in both RGB and depth images within the optimization, which improves track-ing and reconstruction quality for low-cost CMOS sensors. Using a continuous trajectory representation has a num-ber of advantages over a discrete-time representation (e.g. camera poses at the frame interval). With splines, less vari-ables need to be optimized than with a discrete represen-tation, since the trajectory can be represented with fewer control points than frames. Splines also naturally include smoothness constraints on derivatives of the trajectory es-timate. Finally, the continuous trajectory representation al-lows to compensate for rolling shutter effects, since a pose estimate is available at any exposure time of an image. Our approach demonstrates superior quality in tracking and re-construction compared to approaches with discrete-time or global shutter assumptions. 1.
Using Augmented Reality to Tutor Military Tasks in the Wild
"... Intelligent Tutoring Systems (ITSs) have been shown to be effective in training a variety of military tasks. However, these systems are often limited to laboratory settings on standard PCs and laptops which focus on exercising cognitive skills (e.g., decision-making and problem solving) and may pote ..."
Abstract
- Add to MetaCart
(Show Context)
Intelligent Tutoring Systems (ITSs) have been shown to be effective in training a variety of military tasks. However, these systems are often limited to laboratory settings on standard PCs and laptops which focus on exercising cognitive skills (e.g., decision-making and problem solving) and may potentially limit the learning and retention of the dismounted Soldiers and Marines training to master physical tasks. Augmented reality presents the possibility of combining intelligent tutoring with hands-on applications in realistic physical environments. In this paper, we examine the use of an augment-reality based adaptive tutoring system for instruction in the wild, locations where no formal training infrastructure is present, and identify the challenges that arise when developing such a system. We began the transition from desktop tutoring to the wild by exploring an existing real life mockup of a market scene along with low cost commercial-off-the-shelf devices (e.g., HMDs coupled with depth cameras) and a 3D model of the environment. The goal of our canning approach is to use “human in the loop ” 3D scene acquisition via augmented reality so that the scene can be scanned efficiently and with complete coverage. Using this 3D model, intelligent tutoring systems can adaptively manage instruction while being aware of the physical and augmented objects in the scenario. Furthermore, with this awareness of the physical environment, we hope to provide augmented effects and objects (e.g., virtual humans) that register to the physical environment and respond realistically to interactions with
Dense RGB-D Visual Odometry using Inverse Depth∗
, 2015
"... Abstract – In this paper we present a dense visual odometry system for RGB-D cameras performing both photometric and geometric error minimisation to estimate the camera motion between frames. Contrary to most works in the literature, we parametrise the geometric error by the inverse depth instead of ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract – In this paper we present a dense visual odometry system for RGB-D cameras performing both photometric and geometric error minimisation to estimate the camera motion between frames. Contrary to most works in the literature, we parametrise the geometric error by the inverse depth instead of the depth, which translates into a better fit of the distribution of the geometric error to the used robust cost functions. To improve the accuracy we propose to use a keyframe switching strategy based on a visibility criteria between frames. For the comparison of our approach with state-of-the-art approaches we use the popular datasets from the TUM for RGB-D benchmarking as well as two synthetic datasets. Our approach shows to be competitive with state-of-the-art methods in terms of drift in meters per second, even compared to methods performing loop closure too. When comparing to approaches performing pure odometry like ours, our method outperforms them in the majority of the tested datasets. Additionally we show that our approach is able to work in real time and we provide a qualitative evaluation on our own sequences showing a low drift in the 3D reconstructions. We have implemented this method within the scope of PCL (Point Cloud Library) as a branch of the code for large scale KinectFusion, where the original ICP system for odometry estimation has been completely substituted by our method. A PCL fork including the modified method is available for download. 1
Article RGB-D SLAM Combining Visual Odometry and Extended Information Filter
, 2015
"... sensors ..."
(Show Context)
Article A Fast Robot Identification and Mapping Algorithm Based on Kinect Sensor
, 2015
"... sensors ..."
(Show Context)
