摘要
为了准确掌握半干旱地区的土地利用情况,以辽宁省彰武县为研究区,分析多时相的HJ-1B遥感影像。依据影像各时相的归一化植被指数(NDVI)及其波谱特性,获得对影像数据的整体把握。提出基于NDVI、光谱和纹理信息相结合的最大似然分类方法,通过此方法探讨该区的土地利用信息,并将分类结果与基于NDVI和光谱的最大似然分类结果进行对比分析。研究结果表明,纹理特征参与的分类精度较高,Kappa系数达到0.80,总分类精度为83.56%。因此,该方法更适合于半干旱区的遥感影像信息提取。
In order to accurately know the land use situation of the semi-arid areas,this paper analyzes the multi-temporal remote sensing images with Zhangwu Country,Liaoning Province,as the research area,obtains an integral grasp of the image data based on each band characteristic and the normalized difference vegetation index （NDVI）,and advances the maximum likelihood classification method with the combination of synthesized NDVI,spectral and texture information to explore the land use information and to contrast the classification results with those based on NDVI and maximum likelihood classification of spectrum.The results show that the classification accuracy of the first method is much higher,the Kappa coefficient reaching 0.80 and the total classification accuracy 83.56%.Therefore,this method is more suitable for the information extraction of remote sensing image in semi-arid areas.
出处
《国土资源科技管理》
北大核心
2014年第4期61-67,共7页
Scientific and Technological Management of Land and Resources
基金
辽宁省科学事业公益研究基金项目(2011005002)
农业部公益性行业科研专项经费项目(200903007)
关键词
遥感
半干旱
土地利用
纹理
分类
remote sensing
semi-arid
land use
texture
classification
