Um estudo de predição de volumes de clones de eucalipto utilizando redes neurais artificiais

Abstract

Forest inventory is an important element for the effective management of forest resources. Through it is possible, for example, to quantify trees, identify species of a settlement and obtain the total volume to be explored. Volume is one of the most important elements for the exploration of a given area. Therefore, there is the challenge of methods that can precisely calculate the volume of trees without raising costs, including the use of artificial neural networks. This paper aims to present an approach with artificial neural networks for diameter prediction and volume calculation of eucalyptus clones. Models were proposed with and without total tree height, which is costly to obtain in the field. The results achieved were quite promising in relation to traditional methods, besides minimizing the parameters used for the estimation of volumes, thus presenting a path for forest inventory automatio.