基于分形理论和神经网络模型的土地利用分析——以闽清县为例

A Study of Land Use Based on Fractal Theory and Neural Network Model——Taking Minqing County as a Study Case

根据闽清县2000年和2005年的土地利用现状,结合分形理论对土地利用格局变化特征进行分析,并在此基础上运用神经网络模型对土地利用变化的空间差异进行聚类.结果表明:土地利用类型空间分布具有明显的规律性:耕地主要位于5条河流沿岸,而建设用地则位于平坦地带;人类活动对土地利用格局的影响逐步加强,且大多数地类的稳定性增强;闽清县土地利用变化情况可以分为4类,其中南部、中部、北部在土地利用变化上具有各自的特色.

Based on land use actualities of Minqing in 2000 and 2005, combining fractal theory, characteristics of land use changes were analysed. Moreover, clustered spatial differences in land use changes with the adoption of neural network model SOM. The results indicated that there was an obvious rule of land use spatial distribution. Arable lands were mostly along the five rivers, whereas constructional lands were in flat areas. Human activities had increasing influences on land use pattern and the stabilities of most land types were toning up. The complexion of land use changes had four classes, and the South, the Middle part and the North had their characteristics of land use changes respectively.