摘要
随着高分辨率影像研究的发展,影像融合成为遥感影像处理领域中一个重要的研究方向。目前对于影像的融合已有很多的方法,如IHS法、HPF法等。本文将以SPOT-7高分辨率影像作为研究对象,探讨经PCA法、Pansharpening法、Gram-Schmidt法、NNDiffuse法这4种融合方法,融合后的影像在保持地物的波谱特征,以及信息识别效果等多方面的各自优势,并通过定性和定量评价来确定适合SPOT-7高分辨率影像的融合方法,实验结果表明:NNDiffuse方法的融合结果最为理想。
With the development of high spatial resolution satellite images, the fusion of multi - resolution images has been an impor- tant research topic. Current methods for fusion of remote sensing images include IHS transform, wavelet transform, muhiplicative transform, HPF transform, Brovey transform and so on. In this study, four different image fusion methods including PCA, Pansharp- ening, Gram - Schmidt, NNDiffuse Pan Sharpening were applied to a SPOT - 7 high - resolution remote sensing image, and their per- formance in retaining the spectral characteristics of surface features, and improving image interpretation and analysis were compared by use of qualitative and quantitative evaluation methods. Results show that NNDiffuse Pan Sharpening performed best, implying that SPOT- 7 images processed with appropriate image fusion algorithms can have more applications.
出处
《测绘与空间地理信息》
2017年第1期75-78,共4页
Geomatics & Spatial Information Technology
基金
黑龙江省自然科学基金重点项目(ZD201308)
黑龙江省自然科学基金青年项目(QC2016050)资助
国家自然科学基金面上项目(41571199)
关键词
SPOT7影像
融合方法
定性评价
定量评价
SPOT - 7
image fusion
qualitative evaluation
quantitative evaluation
