USE OF ARTIFICIAL NEURAL NETWORKS FOR FORESTRY BIODIVERSITY ESTIMATION
Abstract
This work aims to estimate unequal forest stand characteristics such as biodiversity, biomass, carbon content and forest restoration indexes, as well as estimate the population size of threatened species in forest fragments from remote sensing data (Band and Index Values) and artificial intelligence tools. For this study we used floristic inventory data provided by AMBINOVA Company. For each plot, we estimated the number of species, Shannon-Weaver floristic diversity index, and the input values for landsat 5 bands 1, 2, 3, 4, 5 and 7 and NDVI, SAVI and EVI indices. the use of artificial neural networks. The results showed a correlation between observed and estimated values of 90%, demonstrating the feasibility of the tool for estimating biodiversity indexes.