Optimization Technology Of 3D Model And Its Application

Three-dimensional scanners and three-dimensional data acquisition are commonly used equipment and methods in the three-dimensional field,not only in the field of reverse engineering,but also in many different industrial applications.The most advanced 3D scanners,especially optical scanners,are characterized by high resolution and fast speed.The point cloud data extracted by the three-dimensional scanning device is very dense.These point clouds carry rich information about the surface of the scanned object,so the amount of data obtained is very large.On the one hand,a large amount of point cloud data will cause difficulty in processing scanned data,high computational cost and timeconsuming,so it will consume a lot of computing resources.On the other hand,a large amount of point cloud data brings high data transmission delays and additional difficulties to cloud services and the process of transmitting 3D data to Web clients.At the same time,in the process of acquiring point cloud data by the 3D scanning device,due to the operation error and the influence of the device itself,the point cloud data is noisy and lost.This article also uses a 3D scanning device as the basis for acquiring point cloud data,and further studies the above problems in the scanned point cloud data.The main work and research content of this article are divided into the following four aspects:1.Since two different 3D point cloud data acquired by the 3D scanner have a point cloud with a different geometrical plane missing.In order to obtain complete 3D clay point cloud data,these two different point clouds need to be registered.In this paper,the ICP algorithm is used to register the obtained point cloud,and according to the analysis of the experimental results,the expected registration effect is achieved.2.In the process of using the 3D scanning equipment in this article,the outliers due to occlusion and operation errors need to be removed.Aiming at this problem,two different filtering algorithms,Gaussian filtering and statistical filtering,are proposed for point cloud denoising.According to the experimental results,the two methods have little effect deviation when processing discrete points,but statistical filtering is processing.After the discrete point cloud data is passed,the original point cloud information can be retained more completely,so the statistical filtering is more in line with the expected effect of this article.3.This paper proposes an algorithm based on Gaussian kernel function to support regression vector machine.For the large amount of point cloud data obtained by the 3D scanner,it usually contains a lot of redundant data.These data have little or no additional information about the geometry of the scanned object.The problem of information is simplified by point cloud,and compared with the effect of K-means mean algorithm.It is verified that the point cloud simplification method proposed in this paper can retain more features and geometric information of the initial point cloud under the same simplification.4.In this paper,the processed point cloud data is used for 3D model reconstruction.The obtained 3D clay model is combined with leap motion to develop a 3D clay model system,which realizes the functions of rotation,zoom in and zoom out and scene switching of the 3D clay model.It builds a bridge between culture and technology between resources and media,content and technology,inheritance and innovation,supports the digital inheritance and dissemination of cultural resources,and adapts culture to the constantly updated information technology to meet the changing needs of people.