Bio-optical diversity on cold water coral habitats

Abstract

This study aimed to elucidate the bio-optical diversity of cold water coral habitats and to further improve the accuracy and efficiency of seafloor mapping with the use of an Underwater Hyperspectral Imager (UHI). Comparison between a spectrometer and a hyperspectral imager were investigated, where the former indented to serve as a benchmark for the UHI. Both laboratory and in situ measurements with UHI, deployed on a Remotely Operated Vehicle (ROV), were conducted. Further, UHI raw data were corrected for the effect of the Inherent Optical Properties (IOP) in water and light source spectral output, to provide reflectance spectra. Their signatures were later used for supervised classification, based on optical fingerprints from each pixel in the hyperspectral image. The results were put into context of using a spectral library of verified optical fingerprints , to enable automatic mapping of the seafloor.

A high diversity of bio-optical signatures were found, comprising all colours in the visible spectrum. New knowledge towards a more automated identification, mapping and monitoring of taxa diversity and health status are shown. With a potential for new tools for statistical information. Distance to object of interest (OOI), light source, water constituents (IOP), sensor sensitivity, spectrograph calibration and background, were all factors found to alter the optical signatures of OOI. Controlling for these factors are crucial to enable automatic seafloor mapping by use of a spectral library, but obtainable trough state of the art technology.