Research on Boundary Extraction of STL Models based on Genetic Algorithm

Abstract

To efficiently decompose a large complex STL model, an improved boundary extraction method is proposed based on genetic algorithm. Three curvature parameters (dihedral angle, perimeter ration and convexity) were used to estimate the surface curvature information. Genetic Algorithm (GA) is used to determinate the threshold of feature edge. The discrete feature edges are grouped and filtered using the best-fit plane (BFP), which is calculated by Least Square Method (LSM). Several experimental results demonstrate that the amount of feature edges is about half of the preset threshold method, and useful feature edges were reserved. The extracted feature boundaries can be directly used to decompose large complex models.