Research On Model Feature Extraction Method Based On VSA And Its Application

Triangular mesh processing has always been a hot issue in computer graphics.The main reason is that the triangular mesh model consists of points,lines,and polygons.It describes the topological relationship among vertices.It is relatively intuitive and easy to quickly prototype and integrate.Triangular mesh processing can be used to simulate the surface of complex objects.With texture and shader techniques,realistic scenes and objects can be rendered realistically.The main research contents include mesh denoising,segmentation and reconstruction,parameterization and editing.After decades of research,many efficient algorithms have been proposed and widely used in industrial production design.The development of triangular mesh processing technology has important significance for virtual reality and industrial design.This thesis mainly studies the three problems of feature extraction,mesh segmentation and mesh denoising of mesh models.Try to solve some of the problems involved(such as the interference of noise on feature extraction,the instability of mesh segmentation,the unprotected features of mesh denoising,etc.),The main research contents of the thesis are as follows:1)A noise mesh feature extraction method based on mesh segmentation is proposed.This thesis analyzes and compares the current mainstream mesh feature extraction methods and their application scopes,and discusses the extraction feature methods applicable to noise grids.Aiming at the problem that noise affects the differential properties of mesh models,a noise mesh feature extraction method based on mesh segmentation is proposed.Combining the advantages of mesh segmentation at moderate noise levels,it cleverly avoids the analysis of differential properties.The problem of extracting features on a noise grid is well solved.2)A method for dividing the contour of mesh features based on variational shape approximationA method of mesh feature contour partitioning is proposed,and the geometric analysis process of segmentation region is added.The instability of mesh segmentation algorithm based on k-means clustering is solved to some extent,and further division is obtained.Better mesh feature contour information can well describe the contour information of the mesh.3)A joint filtering mesh denoising algorithm with guaranteed features is proposed.Aiming at the problem that it is difficult to maintain features in the process of mesh denoising,this thesis proposes a joint filtering denoising algorithm with guaranteed features.The algorithm is divided into three steps:the first step is to use the variational shape approximation algorithm to segment the mesh,and the model is subjected to block noise reduction preprocessing.In the second step,the variational shape approximation algorithm is used to extract the feature information in the segmentation boundary,and the mesh is divided into feature regions and non-feature regions.Combine the filter face normal vectors with different filters for the two regions.The third step updates the vertex position by using a non-iterative mesh vertex update method based on the filtered surface normal vector.