Research On Stereo Vision Technology Of Construction Robot

Stereo vision, the most active branch of robotic technology, is an important attribute of intelligent robot. Binocular stereovision can be defined as the process of tri-dimensional reconstruction through the abstraction, identification, matching and interpretation of twain images gathered from identical target, mostly including: video frequency trapping, camera calibration, image preprocessing, feature extraction, stereo matching and tri-dimensional reconstruction as well. It is through further examination of binocular stereovision that we can bring important enlightenment to multi-eye vision system and present supporting technology for visual problem of intelligent locomotive robot, industry assembling robot, domestic robot and community servicing robot. This article based on the system of a all-directions robot hand of the electromechanical laboratory of Jilin University, as the theoretical basis of stereo vision technology, as a means of head to three cameras and DIGI-HIBW-60 three-dimensional imaging system, established a engineering robot systems with visual function, the research concerning the image division technology in the stereo vision in an effort to construct a vision system for the robot hand of all directions so that it can fulfill the task of carrying and moving the target object automatically. It's no doubt that this research carries much weight in the successful realization of engineering robots' complete automation.This article begins with the principle of the paper-computer stereo vision, and then gives a completely explanation of its origin, improving and principle. Then the six modules, which are the collection of vision, the pointing of camcorder, the extraction of character, the matching of stereo, rebuilding of 3D and the dealing of the final stage, are introduced briefly and the division technique of vision is emphasized.The image division is a focal part of the research which is also important in stereo vision. As a fundamental aspect of digital image processing technique, the image division technique enjoys an extending popularity in academic research and application. In terms of image techniques it is intermediate between image processing and image analysis and different from the general image processing. Due to its fundamental role in image analysis it is entitled to special priority which has a direct impact on the analysis conclusions. This article first gives a full account of the definition and classification of the image division and then briefs the concept and requirements for margin testing with a respect of marginal image division. The theoretical analysis on some frequently used margin testing calculation including Roberts, Prewitt, Sobel, Robinson, Kirsch, Laplace, LOG, and Canny have been conducted and the conclusions are made regarding the typical calculations. In addition, according with the most complex and difficult part of the 3D vision-stereo matching, we acquire the data after the feature-extraction, and then establish the relationship between them. We correspond the point in the same space to the image point in different vision, this way we acquired the corresponding vision. For this aim, we use the method of the matching of 3D based on the division of vision, and then establish the corresponding mathematical model.By the end, we establish a 3D vision experiment system which constituted by three parts: camcorder DIGI-HIBW-60, card IEEE 1394 and picture-disposing computer. With it we make the procedure of eliminating lens distortion and edge detection. Finally we got the 3D picture and depth map of the objects. Meanwhile, the measurement of the distance between the actual measurement of distance is compared. This paper build a good foundation for the auto of the robot hand which can move object automatically and independently.