Image-Based Surface Measurement for Close-Range Heritage Documentation

Abstract

Recent new developments in the electronics have leaded to the manufacturing of new high-resolution digital sensors (up to 39 mega pixel) for terrestrial cameras. These technological innovations, together with new processing algorithms of image matching, allow to obtain image-based surface models in an almost automatic way with an accuracy and a detail level that can besurely compared with the one achievable with active sensors. With this in mind, we present new developments in terrestrial 3D surface reconstruction and object modeling using digital images, reporting some tests conducted using low cost digital cameras and commercial or in-house software. The tests were conducted selecting different kinds of objects (low relief, statue, buildings, small heritage find, etc.), with different camera configurations and different image resolutions. Different image matching procedures are analyzed, comparing area-based, feature-based or an integration of both. A critical examination of the results is presented, with particular attention to the daily archaeological or heritage documentation purpose.