摘要
针对传统超分辨率重建方法计算复杂度高、重建效果差等问题,提出一种基于稀疏表示的图像超分辨率重建模型。该模型利用稀疏表示方法,结合自回归原理将原始图像表示为若干个图像块的线性组合,并根据图像边缘特征将图像块进行划分,以提高算法效率,最后结合分治思想、变量分离技术以及增广拉格朗日方法对模型进行求解。实验结果表明,与传统插值算法相比,该算法对图像重建效果更好。
Aiming at the problem that the traditional super-resolution reconstruction method has high computational complexity and poor reconstruction effect,this paper proposes a super-resolution reconstruction model based on sparse representation.The model uses the sparse representation method and combines the autoregressive principle to represent the original image as a linear combination of several image blocks.In order to improve the efficiency of the algorithm,this model divides the image into blocks according to the edge features of the image.Finally,we use the variable separation technique and the Augmented Lagrangian method to solve the problem.Experimental results show that compared with the traditional interpolation algorithm,this algorithm has better effect on image reconstruction.
出处
《软件导刊》
2017年第11期225-229,共5页
Software Guide
基金
国家自然科学基金项目(61502356)
湖北省教育厅科研项目(B2016590)
湖北省教育厅科学规划项目(2016GB123)
关键词
稀疏表示
超分辨率
变量分离技术
增广拉格朗日
sparse representation
super-resolution
variable separation technology
augmented lagrangian
