Research On Stereo Reconstruction System Of Gum-line Based On Stereo Vision

In recent years,the orthodontic technique in the Department of Stomatology has been rapidly developed.At present,the orthodontic technique is roughly divided into the bracketed orthodontic technique and the bracketless orthodontic technique.The invisible orthodontics belongs to the bracketless orthodontics.The invisible braces used need to be separated from the patient’s gum line by a 3D printed thermocompression model.At present,the cutting of the hot pressing model is generally performed by manual cutting.In the face of the strong market,the artificial cutting efficiency is low,and it is not enough to rely solely on manual cutting;and the transparent hot pressing film after manual cutting is rough and needs Secondary processing;since 2014,the annual use of invisible orthodontic technology in China is almost 100% growth trend,artificial cutting hot pressing model can only meet small batch production,it is difficult to meet the increasingly large-scale market.In order to improve the cutting efficiency and cutting quality of the hot pressing model,this paper combines the binocular stereo vision technology and the actual engineering application requirements,and designs a stereoscopic reconstruction system of the gum line based on binocular vision to realize the automatic cutting hot pressing model.The threedimensional reconstruction system of the gingival line was studied.The key problems of camera selection,camera calibration and image processing of the three-dimensional reconstruction system of the gingival line were studied in depth.Finally,a three-dimensional reconstruction system of the gingival line was designed.The experimental results show that the measurement accuracy of this system is 0.1 mm.The main research contents of this paper are as follows:(1)Study the imaging principle of binocular camera,discuss the camera selection;establish the camera linear model and nonlinear model to obtain the camera internal and external parameter solving method,use MATLAB to solve the binocular camera parameters and save;for the shooting process The environmental interference problem is smoothed by bilateral filtering.The principle of the refinement algorithm is expounded.According to the problems existing in the current refinement algorithm,the Zhang refinement algorithm is improved to obtain the gingival line image of single row pixel arrangement.(2)This paper proposes a Harris-SURF-RANSAC algorithm to detect feature points on the gingival line image.By comparing the SIFT algorithm with the SURF algorithm in Open CV,the Harris-SURF-RANSAC algorithm can be compared.Good matching stable feature points,compared with SIFT algorithm to improve 10 pairs of matching points;introduced the stereo matching principle and constraint criteria,focusing on the SAD algorithm in the local matching algorithm and the SGBM algorithm in the semi-global matching algorithm.In the range of points,different window size SAD algorithms are used to perform stereo matching on the gum line,and finally the feature point matching map about the gum line is obtained.(3)A three-dimensional reconstruction system based on binocular vision is established.The system is simple in operation and low in cost.The parallax map of the gingival line is extracted to obtain the three-dimensional coordinates of the gum line.Experiments show that the accuracy of the whole stereo reconstruction system is satisfied.Design requirements.