Estimation of forest volume based on LM-BP neural network model
Dasheng Wu1, 2
COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(4) 131-137
1 Key Laboratory of Zhejiang Province about Forestry Intelligent Monitoring and Information Technology Research, Zhejiang A & F University, Lin'an 311300, Zhejiang, China
2 School of Information Engineering, Zhejiang A & F University, Lin'an 311300, Zhejiang, China
Since cost factors are of primary importance, continuously searching for more efficient and reliable estimation models that could integrate or, in some cases, substitute the traditional and expensive measuring techniques for forest investigation is necessary. The evaluation indexes set, which included 10 factors: elevation, slope, aspect, surface curvature, solar radiation index, topographic humidity index, tree ages, the depth of soil layer, the depth of soil A layer, and coarseness, was established. Then, using the integration data of the administrative map, Digital Elevation Model (DEM), and forest resource planning investigation data of the key forestry city of Longquan, Zhejiang Province, PRC, the membership of each factor was empirically fitted by polynomials, and the forest volume was estimated via an improved back propagation (BP) neural network(NN) model with Levenberg-Marquardt(LM) optimization algorithm(LM-BP). The results show that the individual average relative errors (IARE) were from 23.29% to 47.87% with an average value of 33.06%; The groups relative errors (GRE) were from 0.38% to 9.31% with an average value of 3.65%, this meant that groups estimation precision was more than 90% which is the highest standard of overall sampling accuracy about volume of forest resource inventory in china.