ABSTRACT
Tropical tree species discrimination using hyperspectral data is an emerging field in which progress is closely linked to the availability and advancement of remote sensing technology. Initial interest in the optical properties of tropical plants was expressed in the latter half of the 1980s [1], and the potential for using these properties for tree classification was recognized at least a decade later [2,3]. While studies to that point focused on leaf-to-branch scale reflectance, the first attempt to classify tree crowns using hyperspectral imagery from airborne sensors was made only recently [4]. This small body of research to date has approached tree identification over multiple scales and has been largely exploratory in nature.
