Landscape video mosaic from a mesophotic coral reef

Abstract

Landscape video mosaics (Lirman et al. 2007) were acquired at four sites in the Hind Bank Marine Conservation District, US Virgin Islands (Smith et al.In press), one of which is shown here (Fig. 1). This mosaic covers 31 m 2 with a resolution of 1.8 mm/pixel. Live stony coral cover, estimated from 400 random points on the mosaic, was high (42.5%). The community was dominated by plating or encrusting forms of colonies of the genus Montastraea. Other coral genera identified from the video at lower abundance included Agaricia, Colpophyllia, Helioseris, Mycetophyllia, and Porites.Dive time and diver training were not obstacles to video acquisition for landscape mosaics, which in this case took only 25 min per mosaic by a diver who had only one practice acquisition. Processing each mosaic took about 1 h of analyst time and 24 h of computer time. The ease of data acquisition, ability to use off-the-shelf hardware (i.e., a video camera and possibly lights), and potential suite of indicators (e.g., benthic cover, coral sizes, disease prevalence, and bleaching) that can be obtained from the images suggest that landscape mosaics could be applied widely for surveys of mesophotic coral ecosystems.

Theme section on “Mesophotic Coral Ecosystems: Characterization, Ecology, and Management”

“…Advances in technical diving methods and instrumentation, such as mixed gas diving, rebreathers, and autonomous underwater vehicles, as well as in imaging techniques, e.g., see Gleason et al (2010), are increasingly providing easier access to study coral ecosystems in the intermediate depth realm (Pyle et al 2008). Kahng et al (2010) review the geographic distribution of studies of MCEs and conclude that although some generalizations may be made about community structure and distribution of MCE in the Caribbean, these generalizations cannot be made for the vastly understudied Indo-Pacific regions.…”

Underwater hyperspectral imagery to create biogeochemical maps of seafloor properties

“…The great advantage for imaging and detection purposes is that each image pixel and its highly detailed spectrum has the potential to discriminate subtle differences in spectral features, thus providing a means of classifying different OOI. Current work on airborne hyperspectral refl ectance of corals, algae, seagrass, and sediment indicates that a spectral resolution of <10 nm might improve classifi cation accuracy of unknown targets signifi cantly relative to the RGB bands available on commercial cameras (Mobley et al , 2005;Klonowski et al ., 2007;Gintert et al ., 2009;Gleason et al ., 2010;Fearns et al , 2011). Numerical/statistical analysis of the OOI spectra provides an increased understanding of the natural variability in spectral characteristics of an OOI, beyond that which the 'eye' can see, which in turn aids in developing techniques to classify OOI automatically.…”

Landscape-Level Imaging of Benthic Environments in Optically-Deep Waters

“…Gleason et al ( 2007 ) use a 10 × 10 m twodimensional video mosaic with a resolution of 1-2 mm per pixel to assess hurricane damage to a shallow reef. Using the same technique, a 31 m 2 mosaic of a coral reef site at the MCD was created by Gleason et al ( 2009 ) from an altitude of 2 m. The resolution of these images (1.8 mm per pixel) was comparable to the resolution obtained by the AUV Prosilica camera (2.2 mm per pixel) but from an altitude of 3 m from the bottom.…”