基于Landsat-8 OLI的西安市土地利用类型遥感分类研究

Remote Sensing Classification of Land Use in Xi’an Based on Landsat-8 OLI

近年来,由于区域人口的增加和社会经济的快速发展,西安市的土地利用类型发生了明显变化。土地利用分类可为生态系统模型、水资源模型和气候模型等提供重要信息,遥感技术为土地利用分类提供了有效的工具。本文以西安市2016年Landsat-8卫星的OLI多光谱数据为基础资料,参考国家土地利用分类标准和西安市土地利用现状,将西安市的土地类型分为建设用地、裸地、水体、草地、耕地、林地6类,采用监督分类中常用的最大似然分类法和决策树分类方法对研究数据进行解译,利用总体分类精度和Kappa系数等指标对各分类精度加以评价,并结合实际用地情况对分类结果进行了总结分析。

With the rapid increase of population and the development of social economy in the region,land use types change evidently in recent years in Xi'an city.The land use classification can provide ecosystem model,water resources model,climate models and others with important information,and remote sensing technology provides a powerful tool for land use classification.The paper used the Landsat-8 satellite OLI multi spectral data in Xi'an city in 2016 as basic materials,classified Xi'an city land types into 6 kinds,which are construction land,bare land,water,green land,cultivated land and woodland referring to the national standard of land use classification and recent situations of the land use in the city Xi'an.The study interpreted images in maximum likelihood classification method that is used commonly in the supervised classification and decision tree classification method,evaluated the classification accuracy by overall classification accuracy and Kappa coefficient,summarized and analyzed the classification results by combining with local situations of land uses.