Anti-noise three-dimensional grid optimization method based on Mumford-Shah algorithm framework

Abstract

The invention provides an anti-noise three-dimensional grid optimization method based on a Mumford-Shah algorithm framework, and the method specifically comprises the steps: constructing a Mumford-Shah model, and discretizing the Mumford-Shah model; obtaining a vertex set, an edge set and a surface set of a triangular mesh model of the CAD model; calculating domain information of the vertex set, the edge set and the surface set; calculating a surface normal vector of a triangular mesh model of the CAD model; setting an optimization target of a surface normal vector according to the discretizedMumform-Shah model; converting the optimization target into an N-subproblem and a v-subproblem; adopting a direction alternating method and a variable fission method to iteratively solve the N-subproblem and the V-subproblem to obtain an optimized surface normal vector; updating the vertex position of the triangular mesh model of the CAD model by adopting a mesh reconstruction algorithm to obtaina denoised triangular mesh model of the CAD model; the method provided by the invention has the beneficial effects that the geometrical characteristics of the triangular mesh of the CAD model can bereserved to the greatest extent while various types of noises are processed.

Description

technical field [0001] The invention relates to computer geometric feature processing in the field, in particular to an anti-noise three-dimensional grid optimization method based on the Mumford-Shah algorithm framework. Background technique [0002] Anti-noise 3D mesh optimization is a typical inversion problem in geometry processing, and its main goal is to restore a noise-free mesh while maintaining the basic geometric characteristics of the mesh surface. With the rapid development of scanner equipment, people can easily obtain more and more meshes from the real world. However, due to local measurement errors and the scanning process, a certain degree of noise will inevitably be generated in the 3D mesh model, which will not only degrade the quality of the mesh, but also cause errors in the lower-level 3D mesh processing applications. . Therefore, the task of removing noise from lossy meshes while preserving geometric features becomes increasingly important, especially ...

AI Technical Summary

Problems solved by technology

Although the isotropic filtering method can effectively optimize the quality of the 3D mesh model, it will blur the geometric features of the model at the same time; the anisotropic filtering method will also make the geometric characteristics of the model blurred when dealing with high-frequency noise. features become blurred

[0005] 2. The method based on sparse optimization. The basic idea of ​​this type of method is to optimize the quality of the 3D mesh model by optimizing the global energy function. Although this method can well maintain the geometric characteristics of the model while optimizing the quality, it will produces a staircase effect in smooth areas

[0006] 3. Based on the data-driven method, although the method can effectively optimize the model quality without considering the geometric characteristics and noise types of the underlying surface, the performance of this method depends on the integrity of the training data set, and it is very expensive Time

Images