Assessment of hyperspectral classification methods for benthic cover type mapping

Though several research papers exploring the possibilities of benthic cover type classification have been published, little emphasis has been given to comparison of the performance of the various spectral classification methods. In this study four different classification methods were tested with and without water column correction. The comparison of the methods was done using a simulated data set. The data set was generated using spectral in situ measurements. The results showed significant differences in mapping accuracy among the tested classification methods. The best overall mapping accuracy (75.0%) without water column correction was achieved using SCM classification while the ED classification produced best results (84.5%) when the correction was applied.