Skip to main content

Full text: Fusing ROV-based photogrammetric underwater imagery with multibeam soundings for reconstructing wrecks in turbid waters

DOI: 10.23784/HN116-03 
Underwater photogrammetry 
HN 116 — 06/2020 
23 
Fusing ROV-based photogrammetric 
underwater imagery with multibeam 
soundings for reconstructing 
wrecks in turbid waters 
An article by ROBIN ROFALLSKI, PATRICK WESTFELD, JEAN-GUY NISTAD, ANNETT BÜTTNER and THOMAS LUHMANN 
Observation and monitoring ofwrecksarean integral part of the duties of hydrographic 
offices such as BSH. A common practice consists of first surveying wrecks using vessel- 
based multibeam echo sounding systems and subsequently having divers visually in 
spect them. In order to provide an objective procedure and seta baseline for monitoring 
wrecks, unmanned underwater vehicles equipped with imaging systems can be used to 
inspect wrecks and other obstructions in more details. This paper presents a workflow 
for combining multibeam soundings and photogrammetric point clouds generated by a 
ROV-based camera system. Structure from motion and image enhancement are used to 
obtain a colour-coded point cloud, which is then fused and scaled with the multibeam 
soundings, resulting in data densification on wrecks. Finally, the feasibility of integrating 
this fused data to common hydrographic practice is demonstrated. 
ROV | underwater photogrammetry | multibeam echo sounder | point cloud fusion 
ROV | Unter-Wasser-Photogrammetrie | Fächerecholot | Punktwolkenfusion 
Wracks zu suchen und zu überwachen gehört zu den Aufgaben von Hydrographischen Diensten wie 
dem BSH. In der Praxis Ist es gängig, Wracks zunächst mit schiffsgestützten Fächerecholotsystemen zu 
vermessen und anschließend von Tauchern visuell Inspizieren zu lassen. Um ein objektives Verfahren be 
reitzustellen und eine Ausgangsbasis für die Überwachung von Wracks zu schaffen, können unbemann 
te, mit blldgebenden Systemen ausgerüstete Unter-Wasser-Fahrzeuge eingesetzt werden, mit denen 
Wracks und andere Hindernisse genauer Inspiziert werden. Dieser Beitrag stellt einen Arbeltsablauf zur 
Kombination von Fächerecholotpellungen und photogrammetrischen Punktwolken vor, die von einem 
ROV-baslerten Kamerasystem erzeugt werden. Strukturen aus Bewegung und Bildverbesserung werden 
verwendet, um eine farbcodlerte Punktwolke zu erhalten, die dann mit den Fächerecholotpellungen ver 
schmolzen und skaliert wird, was zu einer Datenverdichtung bei Wracks führt. Schließlich wird gezeigt, 
dass die Integration dieser verschmolzenen Daten In der hydrographischen Praxis machbar Ist. 
i Motivation and state of the art 
Mapping underwater obstructions (e.g. wrecks, 
rock fields) Is a crucial mandate for the German 
Federal Maritime and Hydrographic Agency (Bun 
desamt für Seeschifffahrt und Hydrographie, BSH) 
and other Institutions Involved In safety of naviga 
tion around the world. BSH alone Is responsible for 
monitoring more than 2,500 underwater obstruc 
tions In German territorial waters, regarded as po 
tential hazards to shipping and fisheries (BSH 2020). 
BSH presently carries out surveys of wrecks In 
a two-stage process. Hydrographic vessels first 
survey the wrecks using multibeam echo sound 
ing systems (MBES) to obtain a georeferenced 3D 
point cloud, I.e. sounding set. This data set pro 
vides essential preliminary Information (e.g. posi 
tion, shape, height above seafloor) to professional 
divers, who subsequently Inspect the wrecks with 
the aim of providing an accurate portrayal of 
their state. This process Involves a detailed visual 
and tactile Inspection as well as an Independent 
measurement of the wrecks' shoalest point using 
a pneumatic depth sensing hose pipe. Remotely 
operated vehicles (ROV) equipped with camera 
systems sometimes supplement or partially re 
place the work of divers. Any potential change 
(e.g. collapse, sedimentation, drift) from the pre 
vious known state of a wreck Is subsequently re 
ported. 
Vessel-mounted MBES are widely established 
for collecting bathymetry and detecting objects 
underwater (Brisette et al. 1997). Wide swaths and 
numerous narrow beams allow for efficient object 
detection and Identification, even In turbid envl- 
Authors 
Robin Rofallski is a research 
assistant at Jade University in 
Oldenburg. 
Dr. Patrick Westfeld, Jean-Guy 
Nistad and Annett Buttner are 
employed at BSH in Rostock. 
Prof. Dr. Thomas Luhmann 
teaches at Jade University in 
Oldenburg. 
robin.rofallski@jade-hs.de
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.