自相似性和边缘保持分解的超分辨率重建算法

Super-resolution reconstruction method based on self-similarity and edge-preserving decomposition

针对目前遥感图像超分辨率重建中存在边缘细节信息重建效果不佳的问题,本文提出了—种自相似性特征和边缘特征保持分解的超分辨率重建方法。首先,为了充分利用原始低分辨率图像自身的相似性信息,通过局部自相似性重建方法得到图像的初始重建结果;然后,为进一步增加不同尺度的边缘信息,采用加权最小二乘法对初始重建结果进行多尺度边缘保持分解,并对分解的细节层进行加权线性组合;最后,通过优化计算,得到融合多尺度边缘、细节信息及局部相似性特征的超分辨率重建图像。利用多组仿真和遥感卫星图像进行对比试验。结果表明,该方法可有效提升遥感图像的边缘信息和细节信息。

In view of the poor effect of edge detail information reconstruction in remote sensing image super-resolution reconstruction,a super-resolution reconstruction method based on self similarity feature and edge feature preserving decomposition is proposed.Firstly,in order to make full use of the similarity information of the original low resolution image,the local self similarity reconstruction method is used to obtain the initial reconstruction results of the image.In order to further increase the edge information of different scales,the weighted least square method is used to decompose the initial reconstruction results,and the decomposed detail layers are weighted linear combined.Finally,the super-resolution reconstructed image integrating multi-scale edge,detail information and local similarity features is obtained through optimization calculation.The results show that this method can effectively improve the edge information and detail information of remote sensing images.