Analysing detection bands of two-spectral reflection method to identify forest species composition

Abstract

The optical reflection method is considered for detection of the forest areas where coniferous or broadleaved trees are dominant. Statistical modelling of correct detection and false alarm probabilities for identifying dominant (coniferous or broadleaved) tree species by the two-spectral reflection method has been conducted. It has been shown that monitoring enables us to identify dominant (coniferous or broadleaved) tree species with correct detection probability close to 1 and false alarms probability ~ second decimal places for the temperate climate zone at the wavelengths of 532 and 1540 nm or 532 and 1480 nm. As to the subtropical climate zone, due to a great variety of reflection spectra of vegetation, a selection of the spectral detection bands for reliable identification of dominant coniferous or broadleaved tree species is possible only for specific forestlands where the number of evergreen broadleaved and coniferous tree species is relatively small.