Research On 3D Reconstruction System For Dental Model Based On Stereo Vision

With the improvement of the people's life nowadays, people pay more and more attention on their appearance . Abnormal teeth are the important factor which influences people's appearance. The dental model is an important instrument for diagnosing in orthodontic treatment. In order to make a cure scheme and assess the curative effect exactly. we need to make analysis according to the dental parameters. The traditional methods of measuring parameters of dental model used vernier and foot angle .this method can't get comprehensive information of dental model and take much time and labor. At the same time, as a result of the impact of keeping environment, the dental model will have shape change. So, we need to do three-dimensional reconstruction for dental model.We do three-dimensional reconstruction for dental model based on stereo vision in this paper. The CCD digital camera is adopted to capture digital image of dental model. and Visual c++ 2005 and QT and Open CV in windows platform is utilized to perform investigation and development on reconstruction system. Modularized programming is adopted in system development and design. The reconstruction system include image acquisition, camera calibration, image processing, stereo matching and 3D reconstruction.In the stage of calibration, we introduce the theory of camera calibration and develop the calibration program and acquire matrix of inside parameters, matrix of external parameters, fundamental matrix and distortion parameters. in the stage of image processing, we develop the image processing program. in the stage of stereo matching , we use the matching method based area matching, and introduce the theory of area matching and constraint condition. we do a lot of research on area matching method and compare the effect caused by different matching operators and matching windows. In this stage we can get the parallax image of dental model. in the last stage, based on the results of camera calibration and stereo matching, we reconstruct the matching points according to the mathematical model of space points and use OpenGL to make the model looks more vivid.