Please use this identifier to cite or link to this item:
Title: Estimador geoestatístico e modelos de efeito misto para modelagem hipsométrica e volumétrica de povoamentos florestais
Other Titles: Geostatistical estimator and mixed models for hypometric and volumetric modeling of forest stands
Authors: Calegario, Natalino
Calegario, Natalino
Acerbi Júnior, Fausto Weimar
Carvalho, Samuel de Pádua Chaves
Keywords: Redes neurais artificiais
Modelos lineares
Modelos não-lineares
Modelos de efeitos mistos
Artificial neural networks
Linear models
Nonlinear models
Mixed models
Issue Date: 17-Apr-2019
Publisher: Universidade Federal de Lavras
Citation: DANTAS, D. Estimador geoestatístico e modelos de efeito misto para modelagem hipsométrica e volumétrica de povoamentos florestais. 2019. 120 p. Dissertação (Mestrado em Engenharia Florestal)-Universidade Federal de Lavras, Lavras, 2019.
Abstract: Volume information in a forest stand is essential for the forest management, taking into consideration that most of the productive processes in a company are dependent on the productivity data that are obtained through forest inventories. In this context, the measurement procedures and data processing are fundamental elements and deserve attention, since they are the basis for determining the volume in an area. The forest sector is constantly searching for techniques that enable it to achieve excellence in the estimation of timber production, optimizing processes and maximizing profits. In this work was proposed and evaluated the performance of a model extracted from the training of an artificial neural network and of new models to estimate the total height of eucalyptus trees; of mixed models with adoption of structure of the variance and covariance matrix for hypsometric and volumetric estimates; and a geostatistical estimator for volume prediction per hectare, considering different sample intensities. The techniques presented satisfactory performances, resulting in the gain in accuracy in the estimates of total height, individual volume and stand volume. The results of this study allow the reduction of the sample intensity in pre-cut forest inventories, indicating the potential and applicability of these techniques in the solution of problems of measurement and forest management.
Appears in Collections:Engenharia Florestal - Mestrado (Dissertações)

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.