一种基于图像的几何简化方法

Image Based Geometry Simplification Method

在传统QEM算法的基础上,引入图像处理的方法,根据一种新的简化算子来对几何体进行简化。因为在简化之前,通过对原始纹理图像进行边缘检测,给每条边赋以视觉效果上重要性的权值,这个成为边折叠的一个基本依据,然后再利用这种算子时候,增加了对曲率因素的考虑,使得每次都折叠时同时保留了视觉上重要的边和曲率变化剧烈的边,使得简化剧烈的模型仍能保留较好的外观和形状特征。

Based on the QEM algorithm,we present a new error metric descriptor using image processing method to simplify the geometry mesh.Before simplification,we detect the edge of texture image to get visual importance weight value for each edge.This can be a suppletment measurement of error metric in geometry simplification.Also,we take into account the curvature of each edge to make sure a correct edge collapse order to preserve similar appearance.In the case,some sharply simplified models can keep good visual and shape feature.