摘要
针对我国首个0.5m高分辨率遥感星座高景一号缺少最优融合算法评价的问题,基于ENVI、ERDAS、PCI、ArcGIS四大常用专业软件提供的14种影像分辨率融合方法,开展高景一号影像融合试验,结合目视观察进行主观评价,并选择均值、标准差、信息熵、平均梯度、平均偏差、相关系数等6个统计指标进行客观分析。结果显示,所有的融合方法都能够在提高空间清晰度的同时从总体上保留光谱信息。从遥感人工解译和定量反演2个角度出发,可以考虑选择HPF、Pansharp、Subtractive 3种方法。通过农用地、水体和建筑物3种主要地类的对象提取对比,认为Pansharp融合的结果更为均衡。该文为高景一号影像的深入应用提供参考。
GJ-1is the first satellite constellation with 0.5mhigh resolution in China,but lacking of evaluation for best fusion methods.Provided by four commonly used remote sensing softwares which are ENVI,ERDAS,PCI and ArcGIS,14 different image resolution fusion algorithms were used in experiment for merging the images.Combined with the subjective analysis by visual observation,objective evaluation was implemented through six statistical indicators,composed of mean, standard deviation,information entropy,average gradient,average deviation and correlation coefficient.The results show that all the fusion images can preserve the spectral information while improving the spatial resolution.From the perspectives of artificial interpretation and quantitative inversion of remote sensing,HPF,Pansharp and Subtractive can be choosed for working,and the analysis for the effect of object extraction shows Pansharp is likely a more balanced method,which can provide a reference for the GJ-1application in future.
作者
卢刚
高磊
王彦敏
LU Gang;GAO Lei;WANG Yanmin(Survery Engineering Institute of Jiangsu Province,Nanjing210013,China;Key Laboratory of Satellite Mapping Technology and Application,National Administration of Surveying, Mapping and Geoinformation,Nanjing210013,China)
出处
《遥感信息》
CSCD
北大核心
2018年第6期124-131,共8页
Remote Sensing Information
基金
江苏省测绘地理信息科研项目(JSCHKY201804)
关键词
高景一号
影像融合
定性评价
定量分析
对象提取
GJ-1
image fusion
qualitative evaluation
quantitative analysis
object extraction
