WorldView-2影像数据融合方法比较研究——以在四川省地理国情监测项目中应用为例

Comparative Study for WorldView-2 Image Data Fusion Method——a case of general survey project of geographic national conditions in Sichuan

WorldView-2影像数据是四川省地理国情监测项目的重要数据源。项目通过使用Brovey、Ehlers、HPF、HIS、PCA、Wavelet、PanSharpen、Gram-Schimdt八种不同的数据融合方法对其全色和多光谱影像进行融合;并从清晰度、纹理和色调进行定性分析。从标准差、信息熵、平均梯度、偏差指数、相关系数和光谱扭曲程度进行定量评价;并通过FeatureStation GeoEX解译对比,研究融合结果的可用性。结果表明:基于PanSharpen变换融合方法产生的遥感图像失真较小,同时很大程度地保持了高分辨率全色波段的空间纹理细节信息,是一种适合于WorldView-2影像融合的较好方法;同时,与原始多光谱影像相比,具有更好的解译效果。

WorldView-2 image datian importandatsource fogeneral survey projecof geographic： national conditionin Sichuan. In thiresearch, eighmethods, namely Brovey, Ehlers, HPF, HIS, PCA, Wavelet, Pan- Sharpen and Gram-Schimdwere adopted to fuse panchromatiand muhi-spectral imageof WorldView-2. Qualita- tive analysifrom the aspectof clarity, texture and tone and quantitative evaluation from the aspectof standard deviation, information entropy, average gradient, deviation index, correlation coefficientand spectral distortion degl^e had been done, and the availability of fusion resultare researched by FeatureStation GeoEX interpreted contrast. The resultshow thathe image distortion degree based on the transformation method of Pan-Sharpen ismaller, and meanwhile, the fused image can largely keep the spatial texture detailof high-resolution panchromatiband, so iisuitable method fofusion of WorldView-2 images. Athe same time, the fused image habettein- terpreted effeccompared with the original muhispectral image.