摘要
该文构建了一种视频序列超分辨率重建框架,在此框架下,首先讨论基于正则化处理的单帧图像上行插值和反卷积,然后在详细研究运动补偿矩阵和权值矩阵的构成及其性质的基础上,提出基于加权最小二乘滤波的帧间数据融合算法。实验表明该算法的重建结果相当有效,具有一定的自适应性和鲁棒性。
A framework of reconstructing a super-resolution video sequence from a low resolution video sequence is proposed in this paper. First, an upsampling and deconvolution algorithm for single frame is introduced based on regularization by means of CG algorithm. Then after thoroughly studied the motion compensation matrix and weighted matrix, an adaptive filter is designed for the purpose of interframe data fusion based on weighted least square. Experimental results demonstrate the power of the proposed method with some adaptive and robust.
出处
《电子与信息学报》
EI
CSCD
北大核心
2009年第1期120-123,共4页
Journal of Electronics & Information Technology
基金
江苏省图像处理与图像通信重点实验室开放研究课题(ZK206006)资助
关键词
视频处理
超分辨率重建
运动补偿
TIKHONOV正则化
加权最小二乘
共扼梯度算法
Video processing
Super-resolution reconstruction
Motion compensation
Tikhonov regularization
Weighted Last Square(WLS)
Conjugate Gradient(CG)