|
Titel |
Image seafloor mosaics: Acquisition, processing and role on deep-sea observatory planning and implementation |
VerfasserIn |
Javier Escartin, Rafael Garcia, Ricard Prados, Ricard Campos, Thibaut Barreyre, Bathyluck'08 Science Party |
Konferenz |
EGU General Assembly 2010
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 12 (2010) |
Datensatznummer |
250040928
|
|
|
|
Zusammenfassung |
Near-bottom, high-resolution geophysical surveys carried out by remotely operated vehicles
(ROVs) provide unprecedented resolution of the seafloor structure, either from
multibeam acoustic systems or from imagery (video and electronic still images).
Limitations in image processing techniques and capabilities have greatly restricted the
exploitation of these data to date. We present a set of seafloor image mosaics of the
Lucky Strike hydrothermal field (Mid-Atlantic Ridge, 37.25Ë N), which extends
over a surface of ~1 km2. The image mosaics have a pixel resolution of ~10 mm,
and are built from >60,000 digital images acquired during near-bottom surveys in
1996 (ARGO-II towed system, Lustre’96 Cruise), 2006, 2008 and 2009 (VICTOR
ROV with Module Route, Momareto’06, MoMAR’08 and Bathyluck’09 cruises).
These data provide a coverage that varies from ~25% to practically 100% of the
seafloor, depending on the survey and the zone of the study. Individual images
are corrected for uneven illumination, they are then globally aligned using both
the vehicle navigation and the correspondences among images identified through
image processing, and different surveys are geographically co-registered. Image
processing is thus critical to overcome the limitations of underwater acoustic positioning
systems, which have location errors (up to 10 m) that are much larger than the
resolution of the acquired geophysical data (10 mm pixels in case of imagery).Â
To generate the final mosaic we use graph cuts and gradient-domain processing
to obtain a single seamless image. As the resulting mosaics are large, reaching
>100.000x100.000 pixels, this imagery can be viewed and exploited using a dedicated
viewer which allows also the extraction of information (digitization, extraction of
georeferenced images, etc.). The method developed here is applied to the Lucky Strike
hydrothermal field, where a pilot seafloor observatory will be installed in 2010
(MoMAR node of the ESONET program). Â It demonstrates that it is now feasible to
systematically map with digital cameras large areas of the seafloor (in the order
of 1 km2) in hydrothermal systems, cold seeps, and other areas of interest, and
to obtain georeferenced images. These can then be used to identify and quantify
geological and biological processes, and to guide and map the installation of seafloor
instrumentation and infrastructures associated with deep-seafloor observatories.
Repeated surveys can be now precisely co-registered, and used to study temporal
variability of natural systems or of human impact, including scientific activities. |
|
|
|
|
|