Mapping of sea bottom features using high resolution satellite data

Abstract

Satellite remote sensing techniques have been found to be useful in many areas of applications. This thesis reports on a study using high resolution satellite imageries for mapping of different sea bottom features in tropical waters of Sibu Island and Merambong, Malaysia. The satellite data used for Sibu Island are ALOS AVNIR-2, Quickbird and Landsat-5 TM data, while the satellite data for Merambong is ALOS AVNIR-2 data. Two techniques, namely the depth invariant index and bottom reflectance index were applied in order to map different types of sea bottom features. The depth invariant index technique was derived from the measured radiance, deep water radiance and ratio of attenuation coefficients. However, the bottom reflectance index technique used the measured radiance together with water depths and attenuation coefficients. The sea truth data were used to verify the existence of particular sea bottom features. A direct comparison of overall accuracy, kappa coefficient and Tau coefficient between both techniques revealed that depth invariant index technique using Quickbird satellite data provides accurate mapping for Sibu Island which are 85%, 0.78 and 0.79 respectively while bottom reflectance index technique using ALOS AVNIR-2 data gives better results for Merambong which are 85%, 0.30 and 0.30 respectively. Based on the accuracy of both techniques, it can be concluded that different technique is suitable for different water conditions. The waters in Merambong can be classified as moderately clear and the sea bottom cannot be seen directly from the boat unless it is low tide since the sea bottom is covered by mud and dark sand. The waters in Sibu Island can be classified as clear and the sea bottom can directly be seen from the boat for shallow waters. It can be concluded that depth invariant index technique is suitable for clear waters while bottom reflectance index is suitable for areas with moderate clear waters.