基于联合稀疏模型的高分辨率SAR与光学遥感影像融合研究

发布时间：2018-05-27 14:19

本文选题：稀疏表示 + 联合稀疏模型　；参考：《现代测绘》2016年03期

【摘要】：遥感影像融合技术能通过综合不同卫星传感器获取影像互补信息达到对地物进行提取的效果。针对合成孔径雷达(Synthetic Aperture Radar,SAR)与光学影像融合的光谱扭曲和空间细节信息丢失问题,提出一种基于联合稀疏模型的高分辨率SAR与光学影像融合方法。该方法借鉴信号联合稀疏表示思想,假定SAR与光学影像均可由共有和专有稀疏表示部分组成,分别对应影像冗余信息中的有效部分与影像互补信息:其中共有部分用离散余弦字典表示,专有部分用离散余弦字典和学习字典组成的混合字典表示。在求解出融合影像在联合稀疏模型下的稀疏系数之后重建出融合影像。实验结果表明,该方法能同时保持高分辨率SAR影像空间细节信息和光学影像光谱信息,提高了高分辨率SAR与光学遥感影像的融合效果。
[Abstract]:Remote sensing image fusion technology can obtain complementary image information by synthesizing different satellite sensors to extract ground objects. Aiming at spectral distortion and loss of spatial detail information of synthetic Aperture Radar (SAR) fusion with optical image, a high resolution SAR and optical image fusion method based on joint sparse model is proposed. The method uses the idea of signal union sparse representation for reference and assumes that both SAR and optical image can be composed of shared and proprietary sparse representation. The effective part and the image complementary information are respectively corresponding to the redundant information of the image: the total part is represented by the discrete cosine dictionary and the exclusive part is represented by the mixed dictionary composed of the discrete cosine dictionary and the learning dictionary. After solving the sparse coefficients of the fusion image under the joint sparse model, the fusion image is reconstructed. The experimental results show that the proposed method can maintain both spatial detail information and spectral information of high-resolution SAR images and improve the fusion effect between high-resolution SAR and optical remote sensing images.
【作者单位】：中南大学地球科学与信息物理学院;有色金属成矿预测与地质环境监测教育部重点实验室(中南大学);
【基金】：国家973计划资助项目(2012CB719903) 国家自然科学基金资助项目(41401402,41301453,51479215) 中国博士后科学基金资助项目(2014M552161) 湖南省自然科学基金资助项目(2015JJ3150) 湖南省地理国情监测项目(HNGQJC201503)
【分类号】：P237

【相似文献】