A simple, binary classification algorithm for the detection of Trichodesmium spp. within the Great Barrier Reef using MODIS imagery

McKinna, Lachlan I.W., Furnas, Miles J., and Ridd, Peter V. (2011) A simple, binary classification algorithm for the detection of Trichodesmium spp. within the Great Barrier Reef using MODIS imagery. Limnology and Oceanography: Methods, 9. pp. 50-66.

[img] PDF (Published Version) - Published Version
Restricted to Repository staff only

View at Publisher Website: http://dx.doi.org/10.4319/lom.2011.9.50
 
27
3


Abstract

A binary classification algorithm to detect the presence of high surface concentrations of the nitrogen-fixing cyanobacterium Trichodesmium spp. was developed for high spatial resolution (250 m) imagery of the Moderate-resolution Imaging Spectroradiometer (MODIS). Above-water hyperspectral radiometric measurements of dense Trichodesmium surface aggregations (>10 µg L Chlorophyll a) showed that the water-leaving radiance Lw at wavelengths greater than 700 nm were much higher in magnitude (>0.05 W m2sr-1) relative to the visible wavelengths 400-700 nm (<0.03 W m2sr-1). The binary classification algorithm is based on three criteria. The first criteria relied on the difference in magnitude between the MODIS normalized water-leaving radiance (nLw) at the 859 and 678 nm wavebands. The magnitude of the nLw at the 555 and 645 nm wavebands relative to nLw 678 nm waveband formed the second and third criteria, respectively. The classification algorithm was tested on a small subset of 13 MODIS images with corresponding Trichodesmium sea-truths and yielded an 85% accuracy. Fine-scale features consistent with dense Trichodesmium surface aggregations, such as eddy swirls and windrows, appear to be well represented with the algorithm results. The algorithm was also found to be robust in the presence of highly reflective, potential confounding effects.

Item ID: 15643
Item Type: Article (Research - C1)
ISSN: 1541-5856
Related URLs:
Date Deposited: 07 Jun 2011 07:04
FoR Codes: 04 EARTH SCIENCES > 0405 Oceanography > 040503 Physical Oceanography @ 100%
SEO Codes: 96 ENVIRONMENT > 9611 Physical and Chemical Conditions of Water > 961102 Physical and Chemical Conditions of Water in Coastal and Estuarine Environments @ 100%
Downloads: Total: 3
More Statistics

Actions (Repository Staff Only)

Item Control Page Item Control Page