基于GIS与人工神经网络的广州森林景观生态规划

应用GIS空间分析构建各景观要素的专题图层,建立了森林景观生态模拟的人工神经网络(ANN)模型,并应用分维数、破碎化指数、多样性指数和聚集度指数进行了系统分析,对广州市森林景观进行规划和评价.结果表明:人工神经网络模型泛化能力强,对于研究自然、社会驱动因素与城市森林景观结构变化的非线性对应关系提供了一条有效的途径,是森林景观格局预测、模拟分析及景观规划的新方法.

GIS and ANN-based Ecological Planning of Forest Landscape in Guangzhou

The subject maps of each dominant landscape factor were built under spatial analysis of the geography information system (GIS), and forest landscape simulative ecological model was established by an artificial neural network (ANN). Shape index, fractal dimension, Shannon's diversity index, contagion index were chosen as output of the model and system analysis was performed based on the model for assessment and planning of forest landscape of Guangzhou. The results revealed that the ANN model could be appropriate to study driving mechanism of dynamics of forest landscape pattern under nonlinear relationship among the natural and social factors. It could be a new simulative method for analysis of forest landscape dynamics.