Extracting features including crease,corner and boundaiy from triangulated surface is of important significance in both theory and application~[16,35] .The feature extraction problem is closely related to surface reconstruction ,which has important application in laser range scanning, scientific computing, computer vision, and medical imaging~[3,4] ;the feature extraction can be used in surface simplification, we can segment the unstructured meshes of a geometric 3D model into several parts according to the features, and then simplifying each part, of the meshes iteratively.Therefore, after simplification the shape and feature of the original model can be preserved'5'; another important application of features is to enrich the point cloud around feature lines in order to ensure a sufficient sampling density~[6].In this paper, we first introduce the mesh and its application background and the research of feature detection. Then we expatiate the theory and methods of feature extraction particularly; summarize the curvature estimation and feature extraction methods of 3D points and triangulated surfaces. At last, we define the "Hierarchical Feature Detection" , associated the features with the observation scale to get an algorithm to extract " Hierarchical Features".
|