Medial Axis Extraction for Complex 3D Objects

摘要

A novel methodology for computing the medial axis/skeleton of a discrete binary object using a ‘divide and conquer’ algorithm, in which any 3D object is first sliced into a series of 2D images in X, Y and Z directions. Then, a geometric (Voronoi) algorithm is applied on each 2D image to extract the respective medial axis. This information is then combined to reconstruct the medial axis of the original 3D object using an intersection technique. An optional 3D interpolation step to achieve continuous connected skeletons uses Delaunay triangles and a spherical search to establish the nearest neighboring points in 3D space to interpolate between. Test results show that the proposed 3D Voronoi and optional interpolation algorithms are able to accurately and efficiently extract medial axes for complex 3D objects as well. Finally, an axis-smoothing algorithm using the same Delaunay triangle and spherical test is operable to remove unwanted noise from the extracted medial axis.