波密县冰川遥感判识与信息提取

Remote sensing recognition and extraction of glacier debris flow in Bomi County

依据Landsat 8 OLI影像与Sentinel-1A反演的DEM数据,结合波段比值与面向对象两种方法的优点,采用基于规则的面向对象方法对波密县冰川信息进行判识与提取,利用混淆矩阵评价分类结果的精度。研究结果表明:该方法提取的总体精度98.66%,Kappa系数0.97;提取无表碛覆盖冰川7657.11 km^(2),占波密县冰川总面积的98.72%;表碛覆盖型冰川99.28 km^(2),占波密县冰川总面积的1.28%。

Based on the DEM data of Landsat 8 oli image and sentinel-1A inversion,combined with the advantages of band ratio and object-oriented method,the rule-based object-oriented method is used to identify and extract the glacier information in Bomi County,and the confusion matrix is u sed to evaluate the accuracy of classification results.The results show that:the overall accuracy of the method is 98.66%,and the Kappa coefficient is 0.97;the total area of glaciers without moraine cover is 7657.11 km^(2),accounting for 98.72%of the total area of glaciers in Bomi County;the surface moraine covered glacier is 99.28 km ^(2),accounting for 1.28%of the total glacier area in Bomi County.