基于Beltrami模型和SURE算法的彩色视频超分辨率重建

Color Video Super-resolution Reconstruction Based on Beltrami Model and SURE Algorithm

本文采用基于Beltrami模型的半隐式格式算法实现受损彩色图像的高效优质重建,并将其应用于彩色视频处理。半隐式格式为基于Beltrami框架模型提出的高效算法,根据LOD/AOS分裂格式形成三对角矩阵,并采用追赶法快速求解方程,以实现算法的高效性。使用SURE算法以确定半隐式格式算法中的五个关键参数,并结合遗传算法动态搜索最佳参数值用于逐帧重建受损图像,从而实现降质视频流的清晰化。实验结果表明本方法的彩色视频重建效果良好。

In this paper, a semi implicit scheme based on Beltrami model is used to reconstruct the damaged color image with high efficiency and high quality. The semi implicit scheme is an efficient algorithm based on the Beltrami framework model, which is based on the LOD/AOS splitting scheme to form the three diagonal matrix, and the method is used to solve the equation quickly to achieve the high efficiency of the algorithm. The SURE algorithm is used to determine the five key parameters of the semi implicit scheme, and the optimal parameters are searched by the genetic algorithm to reconstruct the damaged image frame by frame. The experimental results show that the proposed method is effective for color video reconstruction.