Interactive Domitilla catacomb exploration

Abstract

In this paper we present an approach for interactive visualization and manipulation of huge point clouds. Archaeological monuments like the Domitilla Catacomb in Rome lead to data sets surpassing 1 Billion (109) points or 20GB of storage space, which makes standard techniques like mesh conversion or in-core point-based rendering infeasible. Our system uses an out-of-core octree structure and a number of interactive editing tools to enable many archaeological tasks to be carried out on the whole point cloud that would not be possible using traditional methods. We allow fast selection, insertion and deletion of points, and through out-of-core rendering, the frame rate always stays above 20 frames per second on a fast workstation. To the best of our knowledge, this is the first interactive visualization of the complete data set of a large subterranean catacomb, and we show that direct point cloud visualization on the complete data set of a scan campaign is an important tool in archaeological practice.