摘要
如何更好地保持重建图像的边缘信息是当今超分辨率重建技术的一个重要研究课题。针对基于全变分模型的超分辨率重建方法容易产生阶梯效应的不足,利用图像局部模糊熵信息,设计一个表征图像区域结构特征的局部自适应度量函数。利用该函数对全变分模型和超数字化全变分模型进行耦合,进而提出一种基于自适应耦合局部数字全变分模型的超分辨率重建方法。实验结果表明,该方法在保持图像几何边缘结构和消除平坦区域阶梯效应方面能力较强,重建效果较好。
It is an important research topic in super-resolution reconstruction technique that how to better preserve the edges information of the reconstructing image. Total variation model-based super-resolution reconstruction method is prone to generating staircase effect. In light of this deficiency, by using image＇ s local fuzzy entropy information, we design a local adaptive metric function which represents regional structure features of the image. Then we use this function to couple the total variation model and the ultra digital total variation model, and further propose a super-resolution reconstruction algorithm, it is based on adaptive and coupled local digital total variation model. Experimental results show that this algorithm has higher capability in preserving the geometrical edge structures of the image and eliminating the staircase effect in flat regions. Its reconstruction effect is also better.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第6期185-187,227,共4页
Computer Applications and Software
基金
国家自然科学基金项目(1014ZGA033)
甘肃省自然科学基金项目(1112RJZA015)
关键词
超分辨率
全变分
局部结构特征
模糊熵
自适应耦合
Super-resolution Total variation Local structure feature Fuzzy entropy Adaptive coupling
