【摘要】 提出了一种基于形态成分分析的多源图像融合方法。为了将源图像中不同形态结构的卡通-纹理成分分离,把图像的分解问题转化为图像的分类问题,设计了卡通纹理判别字典学习模型。考虑到图像分解不仅与字典有关,还与分解的策略有关,设计了一种新的图像分解模型。在模型中,将纹理成分看成叠加在源图像卡通成分上的噪声,引入非局部均值相似性的一致性正则项,来约束稀疏编码系数的解空间。根据对应成分的编码系数l1范数值最大来选取融合图像的编码系数。实验结果表明,无论是在视觉效果还是在客观指标上,方法都具有更好的融合性能。更多还原

【Abstract】 A multi-source image fusion method based on morphological component analysis is proposed.In order to separate the cartoon texture components of different morphological structures in source images,the image decomposition problem is transformed into the image classification problem and a cartoon-texture discriminant dictionary learning model is designed.Consiering that image decomposition is not only related to dictionary but also to decomposition strategy,a new image decomposition model is designed.In this model,the texture component is considered as noise superimposed on the cartoon component of source image,and introduces the consistency regular term of nonlocal mean similarity to constrain the solution space of sparse coding coefficients.According to the maximum values of corresponding component coding coefficients l1 to select the coding coefficients of fusion images.The experiments have verified that the proposed method has better fusion performance in both subjective visual quality and objective indicator.更多还原