全极化合成孔径雷达(SAR)具有丰富的极化信息, 对地物识别具有显著优势, 提出了一种顾及极化特征的SAR与中分光学影像融合的方法, 对全极化SAR影像进行极化目标分解, 采用改进的色度、饱和度、明度(HSV)变换方法融合极化特征波段与中分光学影像, 并基于面向对象的方法对融合影像进行地物分类。结果表明, 该融合方法优于传统单极化SAR与中分光学影像的HSV融合方法, 能够有效利用全极化SAR的极化纹理信息。面向对象分类方法能够降低SAR对融合影像的斑点噪声影响, 地物总体分类精度优于高分光学影像, 且对于极化信息敏感的地物, 其分类精度明显优于高分光学影像。

Full polarimetric synthetic aperture radar (SAR) possessed rich polarization information, it has a significant advantage for coverings recognition. A fusion method which took into account polarization characteristics of full polarimetric SAR is proposed based on SAR and medium resolution optical image. Full polarimetric SAR is carried out polarimetric target decomposition, polarization characteristics and optical image is fused with the improved hue, saturation, value (HSV) transform method. The fusion image is classified based on the object-oriented method. The results show that the proposed fusion method is superior to the traditional HSV fusion method for effectively using the polarimetric information and texture information of full polarimetric SAR. Object-oriented classification method can reduce the speckle noise of fusion image from SAR. The overall classification accuracy is better than that of high resolution optical image, and the classification accuracy of coverings which is sensitive to the polarization information is obviously better than that of high resolution optical image.