基于LandsatTM数据的土地利用/覆盖分类方法比较研究  

Comparative Study of Land Use/Cover Classification Methods Based on Landsat TM Data

作　　者：梁铭[1] LIANG Ming(Huizhou City Land and Resources Bureau Huicheng District Branch, Huizhou Guangdong 516001, China)

机构地区：[1]惠州市国土资源局惠城区分局,广东惠州516001

出　　处：《北京测绘》2018年第12期1512-1516,共5页Beijing Surveying and Mapping

摘　　要：本文以惠州市作为研究区域,对Landsat 5TM数据进行处理,利用遥感影像的光谱特征进行分类试验,从训练样本数据中分析归纳出分类规则,建立决策树分类,并与传统的监督分类和非监督分类方法进行比较。结果表明,决策树分类方法的总体精度达到88.68%,与传统的分类方法相比有一定的提高,而且算法复杂性低、效率高。This paper takes HuiZhou City as the research area to process the Landsat5TM data,uses the spectral features of remote sensing images to conduct classification experiments,analyzes and summarizes the classification rules from the training sample data,establishes the decision tree classification,and compares with the traditional supervised classification and unsupervised classification methods for comparison.The results show that the overall accuracy of the decision tree classification method is88.68%,which is a certain improvement compared with the traditional classification methods.Moreover,the algorithm has low complexity and high efficiency.

关 键 词：专题制图仪 土地利用/覆盖 决策树分类 监督分类 非监督分类 

分 类 号：P237[天文地球—摄影测量与遥感]