基于遥感和GIS的土地利用分类方法研究

The Research on the Land Use Classification Method Based on Remote Sensing and GIS

本文基于遥感卫星数据,在遥感软件ENVI 5.1和地理信息系统软件Arc GIS 10.0的支持下,通过最大似然法(MLC)、支持向量机(SVM)、人工神经网络(ANN)三种方法,对研究区土地利用/覆盖(LUCC)分类进行信息提取,并对不同分类方法的结果进行比较分析和精度检验.研究表明:使用支持向量机进行遥感图像分类,精度优于最大似然法和人工神经网络,且学习速度也较快,可更好地区分土地利用类型,提高土地利用信息的精度,适用于不同地貌单元,能够作为小尺度范围内遥感影像LUCC分类研究的有效工具.

Based on the remote sensing satellite data and the support of the remote sensing software ENVI5.1 and geographic information system software Arc GIS10. 0,land use / cover classification information extraction and the comparative analysis and precision inspection of the results of different classification methods is studied through the maximum likelihood method（ MLC）,support vector machine（ SVM）,and artificial neural network（ ANN）. The results show that in terms of the accuracy of remote sensing image classification,support vector machine method is superior to the maximum likelihood method and artificial neural network,and the learning speed is faster. Support vector machine method can better distinguish land use types,improve the accuracy of the information of land use,and be suitable for different geomorphic units. Therefore,it can be used as the effective tool of a small scale of study of remote sensing image LUCC classification.