Towards semi-autonomous underwater documentations at “See am Mondsee”
Marco BLOCK-BERLITZ | Dennis WITTCHEN | Benjamin GEHMLICH | Sven ZEISBERG
(HTW Dresden, Dresden, Germany)
Keywords: UUV, 3D reconstruction, SfM, underwater archaeology, autonomous
Abstract:
Archaeologists have highly profited from developments in UAV and progress in camera technology. However, the recording methods currently used in underwater archaeology are still complex and expensive. The importance of the photogrammetric approach in this area is therefore analogous to that in aerial documentation, and the use of underwater photogrammetry is accordingly on the rise. In our experience with recording data while moving, videogrammetry is the more fault-tolerant, more cost-effective and easier-to-use approach. To pursue the aims of documenting archaeological sites and exploring unknown areas, the small submarine “Eckbert-II”, based on the OpenROV, was developed in the project Archaeonautic, with an aim towards semi-autonomous underwater documentations.
A key technological challenge in this context is providing positional data for underwater navigation and georeferencing. Localization in known or unknown environments is a well known problem, which has been studied for decades in the field of robotics as a building block of simultaneous localization and mapping (SLAM). In the last two decades, camera as the only sensor became a more and more popular means of solving the SLAM problem, because of their extremely low size, weight and power footprint, as well as the growing computational capabilities of recent hardware to calculate camera poses of consecutive images (e.g. video stream) in real-time. In the field of archaeology, visual (V-)SLAM is a feasible method for repeatable (semi-)autonomous monitoring of small or large scale sites, where Photogrammetry or Videogrammetry is used to create 3D-models of the excavation area. Whereas the underlying SfM processing is done offline and can take from hours to days for calculating a complete 3D-model, V-SLAM can give quick feedback to the user by providing a sparse representation of the processed area in real-time.
Relevance conference / Relevance session:
The conference is focused on cultural heritage and new technologies as we do in our paper.
Innovation:
Low-cost, easy-to-use UUV documenting underwater situations semi-autonomously.
References:
Figure:
