Artificial neural network model of Pinus armandi landscape pattern in Huoditang Forest Region
2006
Wang Bin | Zhang Shuoxin | Lei Ruide
Chino. 以火地塘林区森林资源二类调查资料为主要数据源，选取华山松林小班的面积、坡度、坡向、海拔、地形指数、灌丛盖度、草本盖度、土层厚度、小班蓄积和优势木年龄10个调查因子作为输入变量，以小班分维数作为输出变量，建立了火地塘林区华山松景观格局的BP神经网络模型，并应用该模型对部分华山松景观格局进行了预测。结果表明，该模型网络收敛效果理想，泛化能力强，为林区森林景观分析提供了一种新的研究方法。
Mostrar más [+] Menos [-]Inglés. Based on class Ⅱ investigation of forest resources, the Back-Propagation Network model was first introduced to landscape pattern in Huoditang forest region. This paper applied the area, aspect, slope, elevation, density, thicket cover, herbage cover, status grade, patch total cumulation and predominance tree species of forty random samples as the model’s input variables and fractal dimension index of each sample as the model’s output variables to train the network. The result showed that the model had a good performance, hence it gave a new study method to analyze forest landscape pattern.
Mostrar más [+] Menos [-]Palabras clave de AGROVOC
Información bibliográfica
Este registro bibliográfico ha sido proporcionado por Institute of Agricultural Information, Chinese Academy of Agricultural Sciences