面向地质应用的高分二号卫星影像融合方法评价

Evaluation on Fusion Method for GF-2 Satellite Images in Geological Application

为了探讨不同影像融合方法在地质应用上的适用性,以高分二号(GF-2)影像为数据源,采用彩色标准化变换(Brovey)、超球体色彩空间变换(HCS)、相位恢复(Gram-Schmidt)、最邻近扩散(NNDiffuse)和高通滤波变换(HPF)等5种融合方法对全色和多光谱数据进行融合实验,并对融合结果进行质量评价。首先从定性和定量两方面进行评价,然后对融合影像进行岩性分类。结果表明:HPF和Gram-Schmidt方法无论是在定性、定量还是岩性分类应用上均具有较好的效果,NNDiffuse和HCS方法融合效果适中,Brovey方法色彩失真明显,该结论可为地质部门提高GF-2卫星的适用性以及对国产卫星数据的推广应用提供参考。

To discuss the applicability of different image fusion methods applied in geology,five fusion methods,Brovey,HCS,Gram-Schmidt,NNDiffuse and HPF transform,are adopted to fuse the panchromatic and multispectral images of GF-2.In order to assess the quality of the fused images,the results were examined through subjective evaluation and quantitative analysis,and then conduct lithological classification for the fused images.The result shows that HPF and Gram-Schmidt transform perform well in all three aspects discussed above.The HCS and NNDiffuse transform have modest performance,and Brovey transform has serious color distortion.This conclusion can provide a good reference for the application of GF-2 images in geological field and promote the popularization of domestic satellite data.