Bilkent University
Department of Computer Engineering

An Automatic Pixel-Wise Mapping Procedure for 3D Digital Image Models

Jürgen Friedrich, Ph.D.

Department of Geodesy and Photogrammetry, Karadeniz Technical University


Digital image processing has seen great progress in the past years but still; there are customer demands for more user-friendliness by an extended automation of labor-intensive processes. One area of research are automatic mapping procedures for digital images of very complex, dense and detailed information like in archaeology, architecture or medicine where a 3D image model should be mapped into a 2D image plan similar to an ortho-photo in order to measure surface positions, angles, distances and areas, for example. The task at hand is more difficult because any type of non-singular body surface should be processed. This work introduces a general approach based on differential geometry where a body given in a 3D image model is split into differential surface parts depending on its curvature. Each part can then be automatically mapped pixel by pixel after computing its inner surface properties. The derived solution method is illustrated by first test results taken from archaeology and architecture.


DATE: October 09, 2002, Wednesday @ 15:40