An Adaptive Blended Algorithm Approach for Deriving Bathymetry from Multispectral Imagery

The log-ratio method (LRM) proposed by Stumpf et al. has been widely used to map bathymetry from multispectral imagery for oligotrophic waters, while the selection criteria of bands LRM have subject tradeoffs between maximum detectable depth and sensitivity. In this article, we first applied a global sensitivity analysis semianalytical forward model optically shallow waters with WorldView-2 band-set. results show that sensitive wavelength band in water-leaving reflectance water varies longer shorter increasing depth. Then, developed an adaptive blended algorithm approach (ABAA) seamlessly shallower region deeper region. different combinations was selected sub-algorithms ABAA. subalgorithms range each subalgorithm ABAA were automatically determined applicable considers logarithmic regression LRM. Landsat-8 Xisha Qundao. When situ data are available, compared blue green bands, significantly improves accuracy estimated depth, especially than 6 m (root-mean- square error (RMSE) = 0.31 0.94 data, RMSE= 0.25 1.42 data). absent, performs better single ratio optimization-based overall.