基于多时相遥感影像的龙岩市景观格局变化分析

A Study of the Landscape Pattern Change of Longyan City Based on Multi-temporal Remote Sensing Images

基于龙岩市多时相Landsat TM/ETM+数据,应用最大似然、决策树及支持向量机分类方法对龙岩市景观类型进行分类。从3种分类方法的比较得知,支持向量机分类方法表现出较高的性能,分类精度明显高于其他的分类方法。因此选择最佳的支持向量机分类结果,并结合景观生态学方法,分析了1992—2008年龙岩市新罗区景观格局及其动态变化信息。结果表明,1992—2008年新罗区主城区农业用地大幅度降低,相应转化为建筑用地,但是却保持着良好的森林覆盖率。同时城市景观组分经历了由扩散式增长过程到粘合式集聚增长过程的转变,城市形态由不稳定形态逐步向稳定形态演化。整体上而言,城市景观呈现出破碎度变小、多样性降低和聚集度升高的发展趋势,建筑用地是龙岩市新罗区的主要景观类型。

Landsat TM /ETM＋ images of Longyan city were employed for landscape pattern classification by traditional MLC（maximum likelihood classification）,DTC（decision tree classification） and SVM（support vector machine） techniques.A comparison between the three classification methods shows that the SVM classification method has outstandingly higher classification accuracy than the other methods.In combination with the theory of landscape ecology,therefore,the authors used the classification results of SVM to analyze the landscape pattern dynamic change of Xinluo district,Longyan city,from 1992 to 2008.The experimental results indicate that,from 1992 to 2008,the agricultural land of the main city proper of Xinluo district was reduced greatly due to its transformation into the built-up land,but the forest coverage remained good.Meanwhile,urban landscape component changed from the process of diffusion growth into agglutinate clustering process and the urban pattern was converted from unstable into stable.Overall,the urban landscape showed the tendency of less fragmentation,lower diversity and higher concentration,and the built-up land had become the main landscape type of Longyan city.