scholarly journals Implementation of Robust Tracking Algorithm on Nano-Computer

2018 ◽  
Vol 11 (2) ◽  
pp. 446
Author(s):  
Khaled Hammemi ◽  
Mohamed Atri
Keyword(s):  
Video Surveillance ◽  
Visual Servoing ◽  
Active Vision ◽  
Field Of View ◽  
Video Sequences ◽  
Partial Occlusion ◽  
Robust Tracking ◽  
Arm Processor ◽  
Probabilistic Process ◽  
Require Data

<p>In this work, we developed the NSSD-DT method, which allows us to track a target in a robust way. This method effectively overcomes the problems of geometrical deformation of the target, partial occlusion and allows recovery after the target leaves the field of view. The originality of our algorithm is based on a new model, which does not depend on a probabilistic process and does not require data-based beforehand. Experimental results on several difficult video sequences have proven performance benefits. The algorithm is implemented on a BCS 2835 system based on a quad core ARM processor, it is also compared to the software solution. NSSD-DT can be used in several applications such as video surveillance, active vision or industrial visual servoing.</p>

Robust tracking via monocular active vision for an intelligent teaching system

2016 ◽  
Vol 32 (11) ◽  
pp. 1379-1394 ◽  
Author(s):  
Rui Wang ◽  
Hao Dong ◽  
Tony X. Han ◽  
Lei Mei
Keyword(s):  
Active Vision ◽  
Robust Tracking ◽  
Teaching System

scholarly journals Augmenting ViSP’s 3D Model-Based Tracker with RGB-D SLAM for 3D Pose Estimation in Indoor Environments

2016 ◽  
Vol XLI-B1 ◽  
pp. 925-932
Author(s):  
J. Li-Chee-Ming ◽  
C. Armenakis
Keyword(s):  
Pose Estimation ◽  
Visual Servoing ◽  
3D Model ◽  
Field Of View ◽  
Indoor Environments ◽  
Camera Motion ◽  
3D Pose Estimation ◽  
Building Models ◽  
Outdoor Environments ◽  
Model Features

This paper presents a novel application of the Visual Servoing Platform’s (ViSP) for pose estimation in indoor and GPS-denied outdoor environments. Our proposed solution integrates the trajectory solution from RGBD-SLAM into ViSP’s pose estimation process. Li-Chee-Ming and Armenakis (2015) explored the application of ViSP in mapping large outdoor environments, and tracking larger objects (i.e., building models). Their experiments revealed that tracking was often lost due to a lack of model features in the camera’s field of view, and also because of rapid camera motion. Further, the pose estimate was often biased due to incorrect feature matches. This work proposes a solution to improve ViSP’s pose estimation performance, aiming specifically to reduce the frequency of tracking losses and reduce the biases present in the pose estimate. This paper explores the integration of ViSP with RGB-D SLAM. We discuss the performance of the combined tracker in mapping indoor environments and tracking 3D wireframe indoor building models, and present preliminary results from our experiments.

Robust Image-Based Visual Servoing With Prescribed Performance Under Field of View Constraints

2019 ◽  
Vol 35 (4) ◽  
pp. 1063-1070 ◽  
Author(s):  
Charalampos P. Bechlioulis ◽  
Shahab Heshmati-alamdari ◽  
George C. Karras ◽  
Kostas J. Kyriakopoulos
Keyword(s):  
Visual Servoing ◽  
Field Of View ◽  
Prescribed Performance ◽  
Robust Image

Feasibility Study of Force Measurement for Multi-digit Unconstrained Grasping via Fingernail Imaging and Visual Servoing

2020 ◽  
Vol 1 (2) ◽  
Author(s):  
Navid Fallahinia ◽  
Stephen A. Mascaro
Keyword(s):  
Visual Servoing ◽  
Degrees Of Freedom ◽  
Force Measurement ◽  
Field Of View ◽  
Imaging Method ◽  
Shear Forces ◽  
Maximum Value ◽  
6 Degrees Of Freedom ◽  
Finger Pad ◽  
Rms Errors

Abstract A fingernail imaging has been shown to be effective in estimating the finger pad forces along all three directions simultaneously in previous works. However, this method has never been used for the purpose of force measurement during a grasping task with multiple fingers. The objective of this paper is to demonstrate the grasp force-sensing capabilities of the fingernail imaging method integrated with a visual servoing robotic system. In this study, the fingernail imaging method has been used in both constrained and unconstrained multi-digit grasping studies. Visual servoing has been employed to solve the issue of keeping fingernail images in the field of view of the camera during grasping motions. Two grasping experiments have been designed and conducted to show the performance and accuracy of the fingernail imaging method to be used in grasping studies. The maximum value of root-mean-square (RMS) errors for estimated normal and shear forces during constrained grasping has been found to be 0.58 N (5.7%) and 0.49 N (9.2%), respectively. Moreover, a visual servoing system implemented on a 6-degrees-of-freedom (DOF) robot has been devised to ensure that all of the fingers remain in the camera frame at all times. Comparing unconstrained and constrained forces has shown that force collaboration among fingers could change based on the grasping condition.

Prescribed performance image based visual servoing under field of view constraints

2014 ◽  
Author(s):  
Shahab Heshmati-alamdari ◽  
Charalampos P. Bechlioulis ◽  
Minas V. Liarokapis ◽  
Kostas J. Kyriakopoulos
Keyword(s):  
Visual Servoing ◽  
Field Of View ◽  
Prescribed Performance

Full Leg Control of a Quadruped Robot Using Linear Visual Servoing with a Stereo Omnidirectional Camera

2011 ◽  
Vol 5 (2) ◽  
pp. 241-246
Author(s):  
Yukinari Inoue ◽  
◽  
Noriaki Maru ◽  
Keyword(s):  
Visual Servoing ◽  
Visual Space ◽  
Quadruped Robot ◽  
The Body ◽  
Field Of View ◽  
Transformation Equation ◽  
Stereo Camera ◽  
Omnidirectional Camera ◽  
Quadruped Robots ◽  
Omnidirectional Image

The authors have previously proposed foot tip control for quadruped robots using linear visual servoing (LVS) with a normal stereo camera. However, a normal stereo camera has a narrow field of view and is incapable of seeing all four legs simultaneously. Consequently, it has been a problem that the control of all the legs have required that the rotatation of the camera be controlled. This article proposes a method by which a stereo omnidirectional camera is provided at a position low on the body to control all four legs through LVS. In this article, we at first present a transformation equation from an omnidirectional image to a binocular visual space, and we develop a servo equation of LVS in which an omnidirectional image is used. Then, through simulation, we confirm trajectories with the LVS applied to foot tip control. We also conduct an experiment using TITAN-VIII to demonstrate the efficacy of the proposed method.

Sign in / Sign up

Export Citation Format

Share Document