CAA2016

Best practices to re-use remote sensing data coming from marine geophysical surveys for the 3D reconstruction of underwater archaeological deep-sites

Manuela Ritondale, Gaia Pavoni, Roberto Scopigno, Marco Callieri, Matteo Dellepiane

Abstract
Since several decades, underwater archaeology takes advantage of the tools available for marine geophysics to map, document and monitoring the seabed. Impressive levels of details have been reached in underwater photogrammetry. Nonetheless, a huge amount of data has been collected in the past with several different purposes and with rather different technologies. The aforementioned materials, particularly those coming from commercial archaeology, is underexplored and might be re-used and optimized by using digital technologies to enhance the accessibility of underwater sites through virtual reconstructions.
Our goal is, on one hand, to investigate the technical and the legal/bureaucratic constraints that prevent the re-use of remote sensing data coming from geophysical marine surveys. On the other side, we propose a solution to reuse stills images and video data acquired with ROVs-based surveys of archaeological sites (characterized by non-homogeneous quality and resolution), aimed at obtaining 3D image-based models. 
Our approach uses standard photogrammetric solutions, but is based on specific pre-processing and enhancements of the input raw data. In the first pre-processing phase, we propose an automatic frame extraction algorithm working on video streams, able to reject the damaged or non-useful frames and to select the more proper for 3D reconstruction, taking into account all the best practices for an optimal photogrammetric reconstruction. In the second phase, we apply underwater-specific image enhancement filters to either images or video frames to sharpen useful details and to correct undesirable aspects for a good reconstruction like color absorption and blur. 
Practical examples and first results will be discussed.