Metric comparison of automatic orientation techniques

Abstract

Automatic orientation of image sequences in close range photogrammetry is becoming more and more important and a large number of strategies and implementations of Structure from Motion (SFM) have been developed. There are several commercial and freeware products which implement automatic orientation procedures but often inexperienced users regard as equivalent these different solutions.

Object of this work is a comparative study between two 3D modelling systems based on the principles of Structure from Motion, which allows to identify interior and exterior orientation parameters of the camera and the reconstruction of a point cloud using a collection of images. The purpose is to identify the variability of the results obtained in terms of reliability and accuracy reconstruction of the final three-dimensional model. The first software package (EyeDEA) is based on a rigorous photogrammetric approach while the second system (Bundler), developed mainly in the context of computer vision, uses algorithms and analytical approaches different usually not as rigorous as photogrammetric techniques.