Reconstruction of 3D urban model using range image and aerial image

Abstract

This paper introduces a method that utilizes airborne laser range images and aerial images for the 3D reconstruction of urban models. The method consists of three procedures: object extraction and object modeling using range images, and texture mapping using aerial images. Objects are extracted from range images by edge analysis and classification through the use of vertical geometric patterns. Our approach also uses the automatic registration of images onto 3D models for texture mapping. The method is based on edge pair extraction as matching features and displacement determination by voting. Experimental results show that 3D reconstruction in a test area is generally successful.