Research And Application On 3D Reconstruction Of Computer Vision

One of the main components of the computer vision is 3D reconstruction. Processing of the nonlinear distortion for 2D images and Calibration of camera's interior and external parameters are key techniques of 3D reconstruction of the computer vision. With development of the computer and digital imaging technique, new system scheme and many new techniques are required. Advanced apparatus should be adequately used. Applicability and facility of the computer vision should be enhanced in different occasion. In the dissertation, theory and application technique of 3D reconstruction are deep studied, based digital camera and in the developing process of management system of the paying for highway road-equipment.A system scheme of 3D reconstruction is brought forward after characteristic of digital camera in the dissertation, that is based digital camera imaging. Solving of nonlinear distortion and calibration of digital camera's interior and external parameters are respectively modeled and solved. Portions of them are solved in the laboratory.Nonlinear distortion of images by digital camera is made of combined errors of lens in the camera. It is confirmed through theoretic analysis and experimental proof. Nonlinear distortion of the images is relative to camera state when it shoots the images. At present, radial distortion of image affect to image processing in imaging by digital camera.A method that is used to expressing magnitude of the image nonlinear distortion is brought forward, after relation of the magnitudes of nonlinear distortion and the image pixels is studied.Mathematical model of image nonlinear distortion is built. Standard graphics method is brought forward that is used to solving coefficient of nonlinear distortion of the images. The method is made in the laboratory. Solving plan of the method is designed. The calculating means and steps of the method are studied in the here.Parameters subgroup calibration of the digital camera is brought forward. Interior and external parameters of digital camera are separately calibrated in the method. The interior parameters can be calibrated in laboratory. Two techniques of the interior parameters calibration are studied in here. A technique is quick and facility solving method. Another is precision solving method. A technique that is called perspective matrix calibration is studied for solving external parameters of the digital camera.Another calibration technique that is called right angle and three points is studied for digital camera external parameters. Calibration device is simple and only 3 points constitute it in the method. A right-angled triangle that is similar to the calibration deice is built in 3 light lines that are imaging lines of 3 point of the calibration device. The relation can solve the external parameters of the digital camera. Theory analysis and process design are made detailed in here.Management system of the paying for highway road-equipment is developed and applied. The project has approved validity and practicability of the studies in the dissertation.