ARTIFICIAL NEURAL NETWORKS FOR ESTIMATING VOLUME OF TREES IN CERRADO
Main Article Content
Abstract
Considering the importance of studies that improve the volume and biomass estimative for the Cerrado biome trees, this work has as objective use artificial neural networks to estimate the volume of trees from different species of cerrado sensu stricto and compare these estimates with ones obtained from volumetric equations traditionally used for the same aim. The data was obtained from 15 squared samples with 400 m² in an area with 29.6 ha. In each plot the diameter at breast height (DBH) (diameter at 1.30 m from soil) and height (H), both total (Ht) and commercial (Hc) height, of all individuals with DBH equals or higher than 3.0 cm were measured. Then each tree was felled in order to obtain their volume. Was used the Huber method considering measurement along the stem up to diameter equals to 3.0 cm. The data obtained of the measurement of the all individuals was used to train artificial neural networks (ANN) and adjust volumetric equations in order to estimate the total volume and commercial volume of trees. The results allowed the following conclusions: ANN can be used to estimate the total and commercial volume; both ANN and regression models are efficient in obtaining the estimated volume of trees in Cerrado biome, presenting low errors and artificial neural networks that consider the specie as a categorical input variable and trained with all data present better results than the ones that are trained for each specie in separately and without the categorical input.
Article Details
The published articles are freely distributed among researchers and social media, and all authors transfer the copyright to Cerne. The research findings can also be used in classroom teaching, conferences, dissertations/theses, and other applications without any restriction. We strongly recommend citing the article to reach a wider audience. The Author also declares that the work is original and free of plagiarism. The authors agree with the publication and are responsible for the accuracy of the information.