为了提高图像超分辨率重建的质量,采用离线双字典学习算法.首先图像块建立字典稀疏模型,确定字典中原子数量;然后使用基于离线字典学习对图像稀疏编码,同时把稀疏编码统一到一个框架中进行优化编码;接着对字典进行分解多个子字典,将图像块中像素点的列向量在子字典展开;最后双字典与超分辨率重构中不同分辨率的异构数据进行同构化,确定控制残差条件,给出了算法实现过程.实验仿真显示本文算法重建效果清楚,峰值信噪比最大,BIQI 最小.

In order to improve the quality of image super-resolution reconstruction,offline double dictionary learning algorithm is established.First,dictionary sparse representation model was built by image block,and the number of atoms in dictionary is determined.Second,sparse code of image is used based on offline dictionary learning and put into frame to optimize code.Third,double dictionary is decomposed into some sub dictionaries,and column vector of pixel of image block is spread in sub dictionary.Last,heterogeneous data with different resolution in the final dictionary learning and super resolution reconstruction is getting isomorphic,control residual is determined,and process is given.Simulation shows that ODDL algorithm reconstruction result is clear,PSNR is better,and BIQI is lower.