Quadrilateral mesh generation method based on model segmentation and splicing

Abstract

The invention discloses a quadrilateral mesh generation method based on model segmentation and splicing. The method comprises the following steps of: 1) inputting a mesh model; 2) segmenting the model into sub-models according to the requirements of users on a target quadrilateral mesh; analyzing and comparing each sub-model to find out the target sub-models having semantic information; 3) respectively generating a direction field and a density field on all the sub-models according to the requirements of the users on the target quadrilateral mesh and the semantic information of each target sub-model; 4) establishing an energy equation according to the semantic information of each target sub-model and the generated direction field and density field, and obtaining the optimized density field and a degree of freedom of boundary subdivision; and 5) independently performing quadrilateral meshing on each sub-model according to the obtained optimized density field and the degree of freedom of boundary subdivision, and then, splicing the sub-models to form the target quadrilateral mesh. The method can perform flexible semantic control on the generation of the quadrilateral mesh, and also can robustly and effectively process a large-scale model.

Description

technical field [0001] The invention relates to a method for generating a quadrilateral grid, in particular to a method for generating a quadrilateral grid based on model segmentation and splicing. technical background [0002] Quadrilateral meshes are widely used in many fields, such as texture mapping, finite element simulation, B-spline surface reconstruction, etc. The main difficulty of quadrilateral mesh generation is how to deal with complex and diverse user requirements and controls, such as patch quality, edge orientation, feature alignment, and singular point distribution, etc. The final mesh quality depends heavily on the global consideration and optimization of these requirements . [0003] At the same time, it is noticed that there are some semantic information in many natural and man-made models, such as repetition information, symmetry information or quadrilateral grid structure information. These high-level semantic information put forward higher requirement...

AI Technical Summary

Problems solved by technology

[0005] At present, quadrilateral mesh generation algorithms are mainly divided into the following types. These methods fail to consider the above semantic information in the algorithm, so they cannot maintain the semantic information in the model; at the same time, none of the existing methods can handle large-scale models:

Images