基于Landsat 8遥感影像的土地利用分类研究——以四川省红原县安曲示范区为例

The land use classification based on Landsat 8 remote sensing image——A case study of Anqu demonstration community in Hongyuan County of Sichuan Province

以四川省红原县安曲示范社区2013年7月23日的Landsat 8OLI影像为数据源,参照国家土地利用分类标准,结合示范区的野外实际调查资料、数字高程模型、草地资源类型等辅助信息,建立了较为准确的解译标志。将目视解译结果视为地表真实状况,对非监督分类、监督分类和基于专家知识的决策树分类的结果进行精度验证。结果表明,监督分类方法对红原安曲示范区的土地利用分类结果与实际类别吻合最好,其分类总体精度可达85.02%,其中面积占示范区总面积约99%以上的草地、湿地、道路及水体的分类精度分别达88.66%、83.42%、83.73%和85.26%。

According to the national standard of land use classification,based on the Landsat 8OLI remote sensing data on July 23,2013,combined with the ground observed data,digital elevation model,grassland resource type data and other auxiliary information,the visual interpretation signature of land use types were established in Anqu demonstration area in Hongyuan County of Sichuan Province.Regarding the results of the visual interpretation as the true values,we analyzed and compared the accuracy of unsupervised and supervised classifications and expert knowledge decision tree classification.The results show that,the supervised classification result has higher accuracy(85.02%)which is close to the actual situation comparing with other methods.The classification accuracy of the land use types of grassland,wetland,road and water,accounting for 99% of the entire study region,are 88.66%,83.42%,83.73% and85.26%,respectively.