Research On 3D Reconstruction Algorithm Of Denture Based On Point Cloud Data

Reconstruction based on denture model of digital dental model,which can effectively solve the traditional model of dental malpractice,the reduction of denture model surface through 3D reconstruction technology information,convenient for doctors diagnose patients,and in combination with 3D technology,can produce high precision denture model quickly,significantly reduce patient treatment cycle.Combined with the actual application of the denture model,the focus is on the feasibility of a series of operations such as point cloud simplification and 3D model reconstruction.The specific research contents are as follows:In view of the incomplete extraction of feature information when the existing point cloud simplification algorithm processes the denture data,and the obvious holes appear in relatively flat areas due to the uneven data distribution,this paper proposes a denture point cloud simplification algorithm based on feature retention.By comparing different topology structure construction methods,Kd Tree was selected to search the K neighborhood of point cloud.Through multi-feature extraction,the feature discriminant function was constructed,and the detailed features of the denture model were retained in the way of multi-feature discrimination,so that the simplified data could better highlight the features of point cloud model and thus improve the accuracy of the algorithm.For the non-characteristic region of the model,a Kmeans clustering subdivision method based on adaptive octree was proposed.K value and initialization clustering center were provided for the Kmeans clustering algorithm through octree,and then further iterative subdivision was carried out according to whether the maximum normal Angle of data between clustering clusters met the threshold value.This method could effectively improve the efficiency of the algorithm.Uniform noncharacteristic region simplification results are obtained.Finally,the extracted features were fused with the results of non-characteristic regions to obtain the final simplified denture data.The experimental results show that the proposed algorithm can obtain the simplified results with good smoothness,simplicity and precision while retaining the detailed features of the denture model.In order to further meet the requirements of digital oral cavity,high precision reconstruction of simplified denture model is needed.By studying the problems existing in Poisson reconstruction algorithm,a feasibility study scheme is proposed.Firstly,based on the original Poisson algorithm,a shielding factor is introduced to increase the reconstruction accuracy of the algorithm.Aiming at the ambiguity of normal vector direction,the global consistency adjustment of normal vector is realized by quadratic direction adjustment.Secondly,an improved moving cube algorithm is used to solve the ambiguity problem.In view of the problem that the traditional algorithm is easy to generate error surface at the edge of the non-closed model surface,this paper designs a Poisson error surface clipping algorithm based on density threshold.Experiments show that the proposed reconstruction algorithm can restore the original model well,maintain the details well,and solve the problem of too many wrong surfaces in the Poisson reconstruction algorithm.The reconstruction effect is good,and the accuracy of the algorithm can meet the needs of practical application.