利用SVM提取赣州市主城区不透水面信息

Using SVM to Extract Impervious Surface Information in the Main Urban Area of Ganzhou City

随着城市化对全球环境变化的影响越来越深刻,及时、准确地监测不透水面的动态变化对环境保护具有重要意义。而遥感技术以其庞大的监控范围及快速的信息获取功能,已成为提取城市不透水面信息的主要技术。但当前的利用遥感技术提取城市地区不透水面信息主要是根据光学遥感影像的光谱信息进行地物分类,往往忽视了重要的地物纹理信息。结合归一化植被指数(NDVI)、归一化建筑指数(NDBI),以及采用灰度共生矩阵(GLCM)提取地物的纹理信息,通过图层叠加的方式进行特征信息组合,最后基于支持向量机(SVM)提取不透水面信息。实验结果表明,该方法的分类总体精度为93.41%,Kappa系数为0.9117,其中不透水面的提取精度为90.64%。通过精度分析,认为该模型提取结果可靠,适用于大范围的城市不透水面信息提取。

With the increasingly profound impact of urbanization on global environmental changes,timely and accurate monitoring of the dynamic changes of impervious surfaces is of great significance for environmental protection.Remote sensing technology,with its large monitoring range and fast information acquisition function,has become the main technology for extracting impervious surface information in cities.However,the current remote sensing technology mainly extracts impervious surface information based on the spectral information of optical remote sensing images for terrain classification,which often ignores the important texture information.In this paper,combing Normalized Vegetation Index(NDVI)and Normalized Building Index(NDBI),and extracting texture information of features using Gray Level Coevolution Matrix(GLCM),combing feature information by layer superposition,and finally the impervious surface information is extracted based on Support Vector Machine(SVM).The experimental results show that the overall accuracy of the classification of the method is 93.41%,the Kappa coefficient is 0.9117,and the extraction accuracy of impervious surface is 90.64%.Through accuracy analysis,the model extraction results are reliable and applicable to the extraction of large scale urban impervious surface information.