The theory of fusion method and the feature of satellite sensor are different, so there is no certain fusion method which is applicable for all types of images. In this paper, it uses the panchromatic and multispectral data of GF-2 as the data source, and adopts five image fusion algorithms Brovey, PCA, GS, HPF and NNDiffuse to process PAN and MSS data. And then a comprehensive evaluation of the effectiveness of image fusion is made through subjective evaluation and quantitative analysis to explore a suitable fusion method for GF-2 satellite image. The results show that fusion image with HPF method has minimum spectral distortion and keep the best spectral information. NNDiffuse method has the best effect in general and the effect of fusion in visible bands is very outstanding. While in the near-infrared band, PCA method and GS method can be used. The results of this research provides the reference for fusion processing and application of GF-2 satellite data.%由于融合处理方法的原理以及获取影像的传感器特性不同,目前尚没有某种融合方法能适用于所有类型的影像.为了探究适用于"高分二号"(GF-2)卫星影像的融合方法,文章以GF-2卫星影像全色与多光谱数据为数据源,采用Brovey、PCA、GS、HPF、NNDiffuse五种融合方法对全色和多光谱数据进行融合处理,以主观评价和定量分析相结合的方式对影像融合结果进行综合对比和评价.试验结果表明,HPF变换法的融合影像光谱畸变最小,光谱信息保持最好;NNDiffuse变换法的综合效果最好,其在可见光波段的融合效果方面很突出;而在近红外波段的融合上,可以考虑PCA法和GS法.文章的研究成果可对GF-2卫星影像数据的融合处理及应用提供参考.
展开▼
