基于Landsat 8遥感影像融合方法和评价

Research on Remote Sensing Image Fusion Method and Evaluation Based on Landsat 8

不同卫星所拍摄的遥感影像对融合算法具有不同的适应性,人们需要根据不同的需求选择不同的融合算法。本文以Landsat 8遥感影像为研究对象,进行全色影像和多光谱影像的融合,同时选择4种影像融合算法,对校正后的Landsat 8多光谱影像和全色影像进行融合处理,并对融合后结果进行定性和定量评价。试验结果表明,Brovey融合算法均值最大为168.266737,融合效果相对较差;PC融合算法的均值最小为161.863477,影像光谱保真度最好;HSV融合算法信息熵和平均梯度依次为5.286542、18.826124,其表达空间细节的能力最佳,但其光谱保真度较差;NNDiffuse融合算法各项评价指标综合效果最好。

The remote sensing images taken by different satellites have different adaptability to the fusion algorithm,people need to select different fusion algorithms according to different needs.Taking Landsat 8 remote sensing image as the research object,this paper carried out the fusion of panchromatic image and multispectral image,and choosed 4 image fusion algorithms to fuse the corrected Landsat 8 multispectral image and panchromatic image,and qualita⁃tively and quantitatively evaluated the results after fusion.The experimental results show that average value of Brovey fusion algorithm is 168.266737,and the fusion effect is relatively poor;the minimum value of the PC fusion algorithm is 161.863477,and the image spectral fidelity is the best;the information entropy and average gradient of the HSV fusion algorithm are 5.286542 and 18.826124 in order,its ability to express spatial details is the best,but its spectral fidelity is poor;the comprehensive effect of various evaluation indexes of NNDiffuse fusion algorithm is the best.