地表温度和NDBI增强VIIRS的城市建成区提取  

VlIRS with LST and NDBI Enhanced Urban Built-up Area Extraction Method

作　　者：边子俊 王佩瑜 BIAN Zijun;WANG Peiyu(Jiangxi University of Science and Technology,Ganzhou 341000,China)

机构地区：[1]江西理工大学,江西赣州341000

出　　处：《长江信息通信》2024年第7期43-46,共4页Changjiang Information & Communications

摘　　要：现有城市建成区提取研究存在湿地、裸地、水体、植被与建筑混淆的问题,导致建成区面积的低估和高估。提出了一种地表温度和NDBI增强VIIRS的城市建成区遥感提取方法,对VIIRS进行数据预处理得到校正后的夜间灯光,利用辐射传输模型估算大气温度,由波谱运算得到NDBI。对不同数据使用几何平均值消除像元异常值,使用Otsu算法根据最佳阈值提取城市建成区。结果显示,本文方法在不同的城市提取建成区结果的F1分数分别为88%、85%、81%,总体精度分别为92.55%、96.15%、93.9%。The existing research on the extraction of urban built-up area has the problem of con-fusion between wetland,bare land,water body,vegetation and buildings,which leads to the un-derestimation and overcstimation of thc built-up area.A remote scnsing extraction method of ur-ban built-up area enhanced by surfacc temperature and NDBI is proposed.The data of VIIRS is preprocessed to get the corrected night light,the atmospheric temperature is estimated by radia-tion transmission model,and NDBI is obtained by spectral operation.The geometric mean is used to eliminate the abnormal values of pixels for different data,and the Otsu algorithm is used to extract the urban built-up area according to the optimal threshold.The results show that the F1 scores of this method in different cities are 88%,85%and 81%respectively,and the overall accuracy is 92.55%,96.15%and 93.9%respcctivcly.

关 键 词：数据融合 地表温度 阈值法 城市建成区 

分 类 号：P237[天文地球—摄影测量与遥感]