COMPUTATION OF NORMAL VECTORS OF DISCRETE 3D OBJECTS: APPLICATION TO NATURAL SNOW IMAGES FROM X-RAY TOMOGRAPHY

Abstract

Estimating the normal vector field on the boundary of discrete 3D objects is essential for rendering and image measurement problems. Most of the existing algorithms do not provide an accurate determination of the normal vector field for shapes that present edges. We propose here a new and simple computational method to obtain accurate results on all types of shapes whatever their degree of local convexity. The presented method is based on the analysis of the gradient vector field of the distance map of the object. Some results on simulated data and snow images from X-ray tomography are presented and discussed.