作者: Eder Pereira Miguel , Alba Valéria Rezende , Fabrício Assis Leal , Eraldo Aparecido Trondoli Matricardi , Ailton Teixeira do Vale
DOI: 10.1590/S0100-204X2015000900012
关键词:
摘要: The objective of this work was to evaluate the effectiveness regression models and artificial neural networks (ANNs) in predicting wood volume aboveground biomass arboreal vegetation area tall Cerrado (a forest, savanna‑like vegetation). Wood were estimated with allometric equations developed for studied area. indices, as predictor variables, from LISS‑III sensor imagery, basal determined field measurements. Equation precision verified by correlation between observed values (r), standard error estimate (Syx), residual plot. total bole (0.96 0.97 r, 11.92 9.72% Syx, respectively), as well (0.91 0.92 22.73 16.80% respectively) showed good adjustments. also adjustments both (0.99 0,99 4.93 4.83% Syx) (0.97 0.98 8.92 7.96% respectively). Basal indices effective estimating cerrado vegetation. Measured did not differ statistically predicted (χ² ns ); however, ANNs are more accurate.