Taking Computer Vision Aloft

Abstract

Structure From Motion (SFM) algorithms are known for their ability to reconstruct a sparse point cloud of scenes that were imaged by a series of overlapping photographs. When complemented by stereo matching algorithms, detailed 3D models can be built from such photo collections in a fully automated way. Since a few years, such algorithms are embedded in several free and low-cost computer vision software packages. Using a mixture of archaeological case studies, it will be shown that those computer vision applications produce excellent 3D data sets from aerial imagery with little effort needed. Moreover, imagery can be generated in both controlled and uncontrolled situations: from low-cost platform such as balloons and kites equipped with simple compact cameras to highly professional medium-format imagers operated during a conventional aerial photographic survey. Besides serving the purpose of a pleasing 3D visualisation for virtual display or publications, the 3D output additionally allows to extract accurate metric information about the archaeology under study.