Remote sensing of benthic habitats in southwestern Puerto Rico

Abstract

Different remote sensing techniques were employed to study benthic habitats in La Parguera, Puerto Rico. These include the comparison of two sensors with different spatial and spectral resolution, IKONOS (1 m, 4 bands) and Hyperion (30 m, 220 bands). Image processing of IKONOS included atmospheric, sun glint, water column corrections, and supervised classifications for the characterization of sea grass, sand and coral. Hyperion data analysis included destriping, atmospheric correction, sun glint correction and classifications. Field data collection was performed by the establishment of three transects with ten quadrants for each habitat class. The best results for image classification in Ikonos imagery were obtained after deglinting of the image with 84 % accuracy and the best result with Hyperion were obtained with the spectral subset in the visible range with an accuracy of 75 %. These results showed that IKONOS had the best results with some limitations on the characterization of the composition of the benthic communities. Hyperspectral shows promise, but the coarse spatial resolution and poor signal to noise of the Hyperion instrument resulted in lower classification accuracy compared to IKONOS.