摘要
不同的辐射校正顺序会引起各波段均值变化,进而影响影像融合的效果和遥感指数特征。为选择适合的数据预处理流程,文章研究比较了PC、NNDiffuse和GS三种影像融合方法在不同辐射校正顺序下的效果,并分析了归一化差异植被指数(NDVI)和归一化差异水体指数(MNDWI)特征。结果显示:不同处理形式下的PC融合影像均呈现失真现象,GS融合影像的光谱保真度均较高,NNDiffuse融合影像在先校正后融合顺序下光谱保真度最高;同时PC融合影像提取的NDVI和MNDWI反映的植被和水体信息效果均较差,而NNDiffuse、GS融合影像提取的以上两种指数特征都能有效获取植被和水体信息,两者各有优势。据此,最佳提取流程为先大气校正后影像融合,同时在地物分类和水体信息提取方面应用NNdiffuse融合方法,而在植被信息提取方面应用GS融合方法。
Different radiometric correction sequences can cause changes in the mean values of each band,which can in turn affect the effectiveness of image fusion and remote sensing index features.To select an appropriate data preprocessing flow,this study compared the effects of three image fusion methods(PC,NNDiffuse,and GS)under different radiometric correction sequences,and analyzed the features of the normalized difference vegetation index(NDVI)and the normalized difference water index(MNDWI).The results showed that under all radiometric correction sequences,the PC fused image was prone to spectral distortion,while the spectral fidelity of the three GS fused image was high,and the NNDiffuse fused image had higher spectral fidelity when the correction was performed before fusion.Therefore,it is recommended to use the NNDiffuse method for land classification and water information extraction by performing atmospheric correction before image fusion,which can enhance image details and improve MNDWI water accuracy.For vegetation information extraction,the GS method should be used by performing atmospheric correction before image fusion to enhance the monitoring ability of NDVI vegetation.
作者
张斌
刘新星
刘冰
茹曼
ZHANG Bin;LIU Xinxing;LIU Bing;RU Man(Henan Aero Geophysical Survey and Remote Sensing Center,Zhengzhou 450053,China;Mineral Resources Exploration Center of Henan Geological Bureau,Zhengzhou 450016,China;School of Earth Science,Hebei GEO University,Shijiazhuang 050031,China)
出处
《航天返回与遥感》
CSCD
北大核心
2023年第5期72-82,共11页
Spacecraft Recovery & Remote Sensing
基金
国家自然科学基金(41702352)
河南省青年人才托举工程(2022HYTP003)。
关键词
影像融合
辐射校正
光谱保真度
遥感指数
数据预处理
遥感图像
image fusion
radiometric correction
spectral fidelity
remote sensing index
data pre-processing
remote sensing image
