Landsat 8 OLI多光谱与全色影像融合算法的比较

Comparisons on fusion algorithms of Landsat 8 OLI multi-spectral and panchromatic images

以Landsat8 OLI卫星遥感影像为数据源,以大伙房水库为研究区,采用Brovery融合法、NNDiffuse Pan Sharpening融合法和Gram-Schmidt融合法对比研究了多光谱和全色数据的融合问题。经过定性和定量分析发现,Gram-Schmidt融合法和NN融合法拥有更丰富的光谱信息,且光谱保真度最优的是Gram-Schmidt融合法。针对Gram-Schmidt融合法,选取了水体边界线处的影像像元进行光谱分析,研究结果发现,融合后的水体边界线呈现得更为清晰,有助于细化水体提取边界。

Three image fusion algorithms of Brovery,NNDiffuse Pan Sharpening（ NND） and Gram-Schmidt（ GS） were compared when the multi-spectral and panchromatic images of Landsat 8 OLI were merged,taking the Dahuofang Reservoir as the study area. From the qualitative and quantitative analyses,it was concluded that GS and NND had richer spectral information and GS had the best spectrum fidelity. Based on the GS method,the image elements at the borderline of the water body were selected for spectral analysis. The results showed that the boundary line of the water after fusion appeard more clearly,which helped to refine the water extraction boundary.