CBERS-02B卫星图像的融合试验研究

STUDY ON FUSION ALGORITHMS OF CBERS-02B SENSING IMAGE

遥感影像数据融合是目前遥感界研究的新热点,作为在各个领域应用日益广泛的CBERS-02B卫星影像,其全色波段空间分辨率达到2.36m,多光谱波段信息丰富,为了既保留丰富的光谱信息又能提高空间分辨率,本文在对主成分变换、Brovey变换、Mutiplicative变换融合3种传统像素级融合方法分析、归纳与算法实现研究的基础上,对CBERS-02B卫星多光谱波段与全色波段进行了图像融合试验研究,从光谱特征与空间纹理特征两个方面对融合效果进行定性与定量评价.分析结果表明,Brovey变换融合法光谱退化最小,同时也最大程度地保持了高几何分辨率全色波段的空间信息.

In recent years, the merging of remote sensing image has become the new hotspot. CBERS - 02B sensing image has been widely applied in each domain. Its panchromatic image has high space resolving power and its multi - spectral image has abundant spectrum information. In order to improve the multi - spectral image＇ s space resolving power, we firstly analyze the principles and methods of three traditional piexl mergings, which are principal components transformation, Brovey transformation and Mutiplicative transformation. Then three traditional pixel mergings are used in CBERS - 02B HR and CCD images. The fused images are evaluated from the spectral quality and from the spatial quality. Analyses show that Brovey transformation has the best quality beth from the spectral quality and the spatial quality.