M. Ioannides et al. (Eds.): EuroMed 2012, LNCS 7616, pp. 40–54, 2012. © Springer-Verlag Berlin Heidelberg 2012 Low-Cost and Open-Source Solutions for Automated Image Orientation – A Critical Overview Fabio Remondino 1 , Silvio Del Pizzo 2 , Thomas Kersten 3 , and Salvatore Troisi 2 1 3D Optical Metrology (3DOM), Bruno Kessler Foundation (FBK), Trento, Italy remondino@fbk.eu http://3dom.fbk.eu 2 Parthenope University of Naples, Dept. of Applied Science, Naples, Italy {silvio.delpizzo,salvatore.troisi}@uniparthenope.it 3 Photogrammetry & Laser Scanning Lab, HafenCity University Hamburg, Germany thomas.kersten@hcu-hamburg.de http://www.hcu-hamburg.de/geomatik/kersten Abstract. The recent developments in automated image processing for 3D reconstruction purposes have led to the diffusion of low-cost and open-source solutions which can be nowadays used by everyone to produce 3D models. The level of automation is so high that many solutions are black-boxes with poor repeatability and low reliability. The article presents an investigation of automated image orientation packages in order to clarify potentialities and performances when dealing with large and complex datasets. Keywords: photogrammetry, computer vision, orientation, low-cost, open- source. 1 Introduction 3D model generation of artifacts, monuments or large environments is becoming a common practice for applications like documentation, digital restoration, visualization, inspection, planning, AR/VR, gaming, entertainment, etc. 3D modeling should be intended as the entire procedure which produces a three-dimensional product starting from surveyed data (reality-based approach) or other sources of information. Data can be recorded with digital cameras or active sensors leading to the well-known image-based [1] or range-based [2] approaches, respectively. The image-based approach is generally considered a low-cost method (in particular for terrestrial applications), flexible, portable and capable of reconstructing lost scenarios simply using archives images [3]. In the recent months different solutions have become available for the automated processing of images and the derivation of 3D information and models. The processing mainly includes image orientation and dense 3D reconstruction with an incredible level of automation. The article investigates the performances and reliability of some low-cost commercial and open-source packages able to automatically process large blocks of images and retrieve the unknown camera poses. Different datasets are used comparing the software outcomes in terms of visual and metric analyses.