Edge extraction for adaptive mesh construction

Abstract

We propose the method of extracting sharp edges of an object from range data in order to create an adaptive mesh model. So far, when an adaptive mesh model is created from range data, we extract points that present important features such as high curvature points of an object. The human vision system relies heavily on sharp edges to recognize objects. So the location of sharp edges of an object should be retained as well when creating an adaptive mesh model. Sharp edges are expressed by the connectivity of points in a mesh model. Our method handles directly the information of the connectivity to create the mesh model.