期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2006
卷号：XXXVI Part 4
出版社：Copernicus Publications
摘要：Coral reefs are considered one of the most diverse, productive and sensitive ecosystems among the various ecosystems existing today. They are usually found in clear waters and often amenable to optical remote sensing. But the submerged and heterogeneous coral reef habitats, the intervening atmospheric path radiance, the contribution of water surface and the effects of variation in water column depth pose great challenge in processing of the measured signal. The present study involves the mapping of coral reefs at certain selected islands of Mahatma Gandhi Marine National Park, South Andaman using multi-spectral data sets of varying resolution and a comparative analysis. Three main aspects considered in the study are spatial resolution, spectral resolution and reef system configuration. Data sets used in the study are IRS P6 AWiFs, IRS P6 LISS 3, Landsat ETM and IKONOS multi-spectral data that represents a range of spatial and spectral resolutions. As this exercise involves a multi-sensor multi-resolution approach, the water leaving reflectance signal were isolated from the influence of atmospheric effects and the same image classification procedure had been adopted for all the data sets. The overall classification accuracy obtained for IKONOS MSS and LANDSAT ETM reveals that blue band (0.45μm-0.53 μm) can contribute greatly to the accuracy of digital classification with respect to coral reef mapping. Depth invariant bands developed from suitable pairs of visible bands helped in discerning six classes of reef habitat from IKONOS MSS. Further analysis with respect to accuracy by taking more reference points falling on the edges or confluence zones of classes from Quick Bird data set is in progress
关键词：Coral Reefs; Multispectral Data Sets; Depth Invariant Bottom Index
