Papers by Keyword: Robot Vision

Abstract: An automated robot vision application for work piece manipulation is described. The purpose is to pick up work pieces like bearings from a pile and store them in a specific location. The necessary image processing algorithms developed for this application and the camera calibration process are realized by using Matlab software. Based on these algorithms the operation for detecting, identifying, sorting and manipulating the bearings is realized. Also the mathematical model for determining the position and orientation of the bearing with respect to the world reference system attached to the robot’s base is presented.

Abstract: Algorithm encryption of wavelet image which has excellent localization property in time domain and frequency domain can locate graphic information having different directional characteristics into any precision level in certain extent. It has feature matched with human visual characteristics because of the disappearance of blocking effect and the noise. The article first analyzes the wavelet transformation principle of image model in feature extraction process and the second iteration process. Dynamic image feature extraction model can be established using this method in robot vision system. The result of envelope is more realistic and the effect is obvious. The application of the diversity training to project and contrast, the comparison precision is high and relative error is low.

Abstract: This Paper will introduce methods used to resolve self-location problem of humanoid robots. There are two method that will be mentioned in this paper. A traditional method which uses marks and flags on the soccer field for robot location will be introduced at first. This paper will also discuss how to solve the singularity problem in traditional self-location method. The optimization self-location method which uses sidelines on the soccer field will be introduced at the same time. This paper will introduce how to mapping the field sidelines in vision to the field sidelines of soccer field. This method is implemented and tested in Robocup 3D Soccer Simulation League. The results show that the accuracy of self-location method will be increased by taking full use of soccer field sidelines in vision information.

Abstract: Target precise positioning by vision system is one of key techniques in robot vision system. In target positioning and selection with robot vision technique, the camera lens distortion must be calibrated. In this paper, a calibration method based on segment slope is used to calibrate the camera and the radial lens distortion coefficient is obtained. The distortion coefficient is used in calculating target position coordinates, and the robot end-exceutor is led to position the target with the use of the coordinates. The experimental results show the effectiveness of the research work. Keywords:robot vision;camera calibration;radial distortion;target positioning

Abstract: This article took the 2011 ABU Robocon Contest as background, and introduced the design and implementation of a vision system for visual servoing navigation and object tracking in this contest. For visual servoing navigation part, it had a series of simple image processing and then added to a previous control law; tests showed only 3cm distance and 1°angle error occur when robot ran one line of 6m. For object tracking part, article introduced a combined algorithm for the detecting of a black and white semi-circle object, which moved quickly and uncertainty in a colorful background and shining environment; tests showed the proposed algorithm work better than other exist algorithms in complex environment.

Abstract: This paper presents the use of a fuzzy logic controller combined with Machine Vision to improve the accuracy to identify ball and speed to approaching ball in Robot Soccer. The conventional robot control consists of methods for path generation and path following. When a robot moves away the desired track, it must return immediately, and while doing so, the obstacle avoidance behavior and the effectiveness of such a path are not guaranteed. So, motion control is a difficult task, especially in real time and high speed control. To achieve good control performance and requirements for timeliness, accuracy to identify and approach the ball, the paper designs a sugeno fuzzy control system, which could make robot complete the whole football performance, including finding, approaching and shooting ball. At the end of the paper, to validate the proposed fuzzy algorithm, the experiment on METALFIGHTER-2 humanoid robot has been run, which showed that the fuzzy controller is effectiveness and robustness.

Abstract: An automated object recognition application type, implemented on a robotic platform, is described. It is based on an image processing algorithm that enables: the detection of objects like bearings from a pallet conveyor, the identification of bearings type and the calculus of position and orientation, in order to be manipulated by a SCARA robot. The identification algorithm developed in Matlab software consists in HSV image segmentation and a matching technique based on comparison with previous given set of bearing prototypes. The methods used for camera and robot calibration are also described.

Abstract: This paper introduces calibration technique for robot vision system, aiming to enhance the robot’s ability to adapt to various environments. In accordance with a calibration plate and car body, we figured out the relationship among the coordinates of car body, camera, and robot by image acquisition, processing, analysis, and calculation. An algorithm was developed to deduce the coordinate transformation between robot frame and calibration plate frame with 4 points. Its feasibility was validated by the experiment.

Abstract: Timeliness and accuracy is a key to be resolved on robot binocular measurement. In this paper, a kind of robot vision projection has been completely established. It analyzes the principle of binocular ranging in three aspects, makes the calculation concise and easy to understand, and expands the range of effective distance. On binocular image processing, we have proposed a gray-scale computing for, firstly, generating characteristic area, then, executing template matching in the area, finally, extracting feature points and matching them in the templates. It ensures certain robustness to noise spots and tries its best to avoid mismatches. The experiments show that the robot vision system has a better accuracy and a low time complexity, and the robot can react in real time.

Abstract: In this paper, the robot vision systems are studied. Through the analysis of the visual tracking process, this paper classifies and introduces several commonly track. The features affecting the quality of target tracking, such as block, rotation, translation deformation and others, are analyzed and discussed. At last, some further directions of target tracking algorithm are also shortly addressed.