P6影像数据融合及其最佳融合方法在漓江两岸土地利用调查中的应用

Application of P6 Image Data Fusion and Its Optimal Fusion Method in Remote Sensing Investigation of Land Utilization along Both Sides of Lijiang

遥感影像融合对影像分类、特征提取和目标识别具有重要的意义。针对印度P6卫星LISS-4全色数据和LISS-3多光谱数据进行多种融合实验,融合方法包括HIS变换、Brovery变换、PCA变换和Gram-Schmidt变换。根据融合结果的最终应用目的——目视解译和屏幕矢量化,采用信息熵、标准差、相关系数和视觉效果对融合方法进行定量和定性评价。分析结果表明,Brovey变换后的图像空间信息详细程度最差;HIS变换后的图像光谱保真能力最差;PCA变换和Gram-Schmidt变换方法融合效果图纹理信息清晰,光谱保真度好,适合P6卫星影像融合。最后选择2%线性拉伸后的PCA融合影像为工作底图,在此基础上对漓江两岸土地利用状况展开遥感调查。

The fusion of remote sensing images is very significant for image classification,feature Abstraction and target recognition.In this article,normal image fusion algorithms（HIS transform,Brovey transform,PCA transform,Gram-Schmidt transform） were used to fuse the multispectral images of LISS-3 and panchromatic images of LISS-4.According to the final application purposes of the image fusion which were visual interpretation and screen vector,we used the information entropy,standard deviation,correlation coefficient and visual effect to quantitatively and qualitatively evaluate the fusion algorithms.The results indicate that Brovey has the lowest details of spatial information and HIS has the lowest spectral fidelity,but the PCA and Gram-Schmidt transform fusion images which are suitable for P6 image fusion have more clear texture information and better spectral fidelity.So,we chose the two percent of linear extension PCA fusion image for the working map for remote sensing survey of land utilization along both sides of Lijiang.