In this paper, we focus on the problem of directing the gaze of a vision system mounted to the body of a high-degree-of-freedom (DOF) legged robot for active perception deployments. In particular, we consider the case...
详细信息
ISBN:
(纸本)9781538660263
In this paper, we focus on the problem of directing the gaze of a vision system mounted to the body of a high-degree-of-freedom (DOF) legged robot for active perception deployments. In particular, we consider the case where the vision system is rigidly attached to the robot's body (i.e., without any additional DOF between the vision system and robot body) and show how the supernumerary DOFs of the robot can be leveraged to allow independent locomotion and gaze control. Specifically, we augment a workspace central pattern generator (CPG) with omnidirectional capabilities by coupling it with a body pose control mechanism. We leverage the smoothing nature of the CPG framework to allow online adaptation of relevant locomotion parameters, and obtain a stable mid-level controller that translates desired gaze orientation and body velocity directly into joint angles. We validate our approach on an 18-DOF hexapod robot, in a series of indoor and outdoor trials, where the robot inspects an environmental feature or follows a pre-planned path relative to a visually-tracked landmark, demonstrating simultaneous locomotion and directed vision.
3D measurement and reconstruction techniques are widely used in many fields, such as automated production, computer vision, and medical diagnostics. Fast and accurate three-dimensional reconstruction of the surface of...
详细信息
ISBN:
(纸本)9781450371834
3D measurement and reconstruction techniques are widely used in many fields, such as automated production, computer vision, and medical diagnostics. Fast and accurate three-dimensional reconstruction of the surface of the object for three-dimensional reconstruction is an important technical problem [1]. Based on the monocular vision measurement technology, this paper proposes a single-camera single-light source system for 3D reconstruction. the experiment uses a hemisphere of known radius to make the center of the hemisphere coincide withthe target center. the circular target center is derived by the algorithm proposed in this paper and three-dimensional information of the location and error analysis.
Aiming to enhance features discrimination for pose variation in person re-identification, reinforcing soft part-aligned features is proposed in this paper. By adding a channel-wise module in two-stream re-identificati...
详细信息
ISBN:
(纸本)9781728133263
Aiming to enhance features discrimination for pose variation in person re-identification, reinforcing soft part-aligned features is proposed in this paper. By adding a channel-wise module in two-stream re-identification network, the network can be more robust to person's pose changes. Firstly, the pre-trained GoogLeNet and OpenPose is loaded to extract pedestrian appearance features and pose feature of pre-processed images. Secondly, pose features are fed into the designed channel-wise module. Finally, the two branches are trained jointly by applying bilinear pooling. From the experimental results on the Market-1501 dataset and the DukeMTMC-reID dataset, it shows that reinforcing soft part-aligned features are more discriminative and achieves higher identification accuracy over the state-of-the-arts.
As a light-weight and efficient stereo disparity estimation algorithm, ELAS (Efficient Large-scale Stereo) has gained a lot of attention in the community of vision-based robotics. However, ELAS directly generate the d...
详细信息
ISBN:
(纸本)9781728133263
As a light-weight and efficient stereo disparity estimation algorithm, ELAS (Efficient Large-scale Stereo) has gained a lot of attention in the community of vision-based robotics. However, ELAS directly generate the disparity map from the Delaunay triangulation mesh of detected support points, which ignores the fact that these points may reside on different physical planes and consequently results in big errors on planar surfaces. In this paper we introduce the plane constraint into ELAS for better performance and robustness. We first cluster the selected support points using mean shift segmentation before triangulation, and then use the classified point sets belonging to different planes to triangulate separately, which avoids the errors caused by the cross-plane triangulation neglecting possible depth discontinuities in ELAS algorithm. Moreover, our algorithm also gets parametric plane models by fitting the large flat planes to rectify some mismatched support points. We evaluated our algorithm on boththe Middlebury dataset and our own captured indoor scenes, and the results show that our approach behaves better in occlusion and weak texture areas with faster speed.
In the RGBD-SLAM system, the feature point's uncertainty plays an important role in back-end optimization of the entire system. By analyzing advantages and disadvantages of existing uncertainty models, we propose ...
详细信息
ISBN:
(纸本)9781728133263
In the RGBD-SLAM system, the feature point's uncertainty plays an important role in back-end optimization of the entire system. By analyzing advantages and disadvantages of existing uncertainty models, we propose a scale-space-based uncertainty model. the feature point's uncertainty in disparity image space is determined simultaneously by boththe scale-space where the point locates and its disparity value. the pyramid layer where the feature point locates corresponds to the uncertainty of both itself and its pixel position. And the uncertainty of the feature point's depth is also related to its disparity. that is, as feature's disparity increases, the uncertainty of the feature point's depth also increases. Compared with traditional models, our model performs better in public dataset.
In this paper, a mobile robot equipped with an onboard computing unit and a stereo camera for autonomous target detection and tracking is introduced. this system can figure out an interesting target and track it robus...
详细信息
ISBN:
(纸本)9789897583803
In this paper, a mobile robot equipped with an onboard computing unit and a stereo camera for autonomous target detection and tracking is introduced. this system can figure out an interesting target and track it robustly in real time. It is based on the ROS framework and can handle multi-resource information, such as RGB images, depth information, and IMU data. To balance the performance of the machine learning based object detection algorithm and the algorithm for object tracking, the Hamming distance and the intersection over union are selected as criteria. the performance of the system is verified in a hardware experiment in two typical scenarios.
Recently, human-robot collaboration has attracted many research interests. Unlike traditional industrial robots, human-robot collaboration combines the advantages of robots withthe flexibility and cognitive skills of...
详细信息
ISBN:
(纸本)9781728133263
Recently, human-robot collaboration has attracted many research interests. Unlike traditional industrial robots, human-robot collaboration combines the advantages of robots withthe flexibility and cognitive skills of human worker to work together in a shared manufacturing environment. However, to achieve an efficient human-robot collaboration, the communication channels between human workers and robots needs to be tackled. vision-based hand gesture recognition has been effectively applied as the interface between humans and computer for long time. So, this paper presents a survey of the vision-based hand gesture recognition research in human-robot collaborative manufacturing. We first propose an overall model of vision-based hand gesture recognition for human-robot collaboration. then, we review three essential technical components of this model: sensor technologies, hand gesture detection and segmentation and hand gesture classification. In addition, the relevant techniques for vision-based hand gesture recognition are compared. Towards the end of this paper, future research trends are discussed.
Fast implementation of stereo depth estimation algorithms on low-cost processors is demanding both insight of the algorithmic architecture and efficient parallelization techniques. this paper proposes an ARM-base Open...
详细信息
ISBN:
(纸本)9781728133263
Fast implementation of stereo depth estimation algorithms on low-cost processors is demanding both insight of the algorithmic architecture and efficient parallelization techniques. this paper proposes an ARM-base OpenMP parallel implementation method of ELAS (Efficient Large-Scale Stereo Algorithms), which is very promising for fast computation on embedded systems. Our experiments were conducted under multi-core environments, both on an 8-core computer and an 8-core ARM ODROID-XU4. the parallel performance was examined in two ways: single image's parallel processing and batch processing of an image set containing seven pictures. Experimental results demonstrate the efficiency of our acceleration method, which can achieve a speeding-up for more than four times compared withthe original sequential implementation.
this paper takes the planar formation tracking control problem of second-order agents with model uncertainties into consideration. A novel adaptive neural network (ANN) estimation method based on neighborhood informat...
详细信息
ISBN:
(数字)9781728177090
ISBN:
(纸本)9781728177106
this paper takes the planar formation tracking control problem of second-order agents with model uncertainties into consideration. A novel adaptive neural network (ANN) estimation method based on neighborhood information is designed to estimate model uncertainties. By using the tools of the back-stepping method and ANN, a robust formation tracking control law is given. the law can guarantee this formation tracking errors uniformly bounded under connected topologies. Numerical simulation results verify the validity of the analysis results.
In this paper, we provide a tutorial introduction to a topological approach for the computation of some supremal sublanguages, often specified by language equations, that arise in the study of the supervisory control ...
详细信息
ISBN:
(数字)9781728177090
ISBN:
(纸本)9781728177106
In this paper, we provide a tutorial introduction to a topological approach for the computation of some supremal sublanguages, often specified by language equations, that arise in the study of the supervisory controltheory. As an illustration, we show that the supremal sublanguages for properties such as normality, prefix-closedness, trace-closedness, L-closedness and prefix-closed controllability are supremal open subsets of certain topologies. thus, the supremal sublanguages correspond to the interiors and can be directly expressed.
暂无评论