SHALLOW WATER BATHYMETRY MAPPING USING HYPERSPECTRAL DATA

Using airborne hyperspectral images and field survey data of Yamada Bay, northeast Japan, we examined to improve the water depth accuracy by enhancing benthic type classification accuracy and generalize the method. Also, the method was applied to airborne hyperspectral data of Akajima Island, southwest Japan and the versatility of the method was verified. Compared to field survey data in Yamada Bay, the determination coefficient of water depth improved from 0.19 to 0.50, however, RMSE increased from 2.7 m to 3.3 m. In Aka-jima Island, the determination coefficient improved from 0.34 to 0.74, and RMSE decreased from 2.6 m to 1.3 m.