| With the rapid development of science and technology, the traditional manualmeasurement test model has been replaced by the advanced image processingtechnology and computer technology, using the computer realization of visualinformation processing, thus forming a new subject-computer vision technology. Thenature of computer vision technology is used binocular camera to replace human visualfunction, Use of computer technology for processing to realize the understanding of theobjective world of3d scene, analysis and understanding.Binocular vision is a non-contact measurement technology, binocular visionmeasurement system has a simple structure, easy to operate, online real-time3Dmeasurements and other advantages. Currently, it is attracting many researchers indifferent subject to carry out in-depth research, and in many areas it has broadapplication prospects.How to get a clear image pair is the foundation of all other image processing.Therefore, this first image acquisition-autofocus technology conducted in-depthresearch, analysis of the evaluation function, on this basis, a new focusing evaluationmethod based on local2/8center windows is proposed for an auto-focusing system inthis paper. We have proposed a focusing evaluation function which improved Robertsfunction with eight-neighborhood or four-neighborhood. The approximate optimalcriteria window is selected by2/8rule. Comparative experiments implemented withother methods have shown that our proposed approach not only meet the basicperformance focusing evaluation function, but also has unique advantages.Secondly, this paper describes the calibration method used and the calibrationprinciple, pushed to the relationship between the coordinate system calibration, and thengives a structural model of binocular vision system and then perform precision analysis.Based on binocular vision stereo matching technology, first introduced the stereomatching constraints and matching strategies, and then introduced regional-based,characteristics-based and phase-based matching method and the difficulties of thematching technology and future research directions.Then introduced the common methods of visual-based three-dimensionalreconstruction and three-dimensional reconstruction of the principle of binocular vision,depth study of the key technologies of three-dimensional reconstruction based on binocular vision techniques: feature extraction, feature matching and depth informationcomputing; using the least squares method to achieve a Euclidean space-basedthree-dimensional point cloud computing of scene objects and to achieve athree-dimensional display method based on OpenGL, Finally, on OpenCV and VS2010platform to develop three-dimensional reconstruction and measurement of the system.Finally, in the development of three-dimensional reconstruction software systemsand hardware platforms, experimental results are given stereo calibration, stereocorrection, three-dimensional stereo matching and coordinate calculation, triangulationand three-dimensional display, and analysis of the experimental results. |
