Abstract:
The objective of this research are two main issues, the first is to find the suitable data type (i.e. EO-1 Hyperion and EO-1 ALI) and the appropriate method to estimate Leaf Area Index (LAI) of dense canopy mangrove. Another issue is the classification of tropical mangrove at species level by band selection methods. The study site is at the Talumpuk cape, Pak Phanang District, Nakorn Sri Thammarat Province, Thailand In the first issue, the results suggest that the Partial Least Square Regression (PLSR) improved the accuracy compared with other methods. The Root Mean Square Error (RMSE) of PLSR methods are 0.192 and 0.337 when used the Hyperion and ALI respectively. Moreover, the hyperspectral data helped improve RMSE 43% compared multispectral data when used PLSR. This study demonstrated the capability of the hyper-dimensional remote sensing data for discriminating diversely-populated tropical mangrove species. It was found that five different tropical mangrove species were correctly classified. The Genetic Algorithm based bands selection helped improve the overall accuracy from 86% to 92% despite the remaining confusion between the two members of the Rhizophoraceae family and the pioneer species.