节点文献
基于加权最小二乘滤波的视频序列超分辨率重建
Super-resolution Reconstruction of Video Sequence Based on WLS Filtering
【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。
【Abstract】 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.
【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;
【Key words】 Video processing; Super-resolution reconstruction; Motion compensation; Tikhonov regularization; Weighted Last Square(WLS); Conjugate Gradient(CG);
【Key words】 Video processing; Super-resolution reconstruction; Motion compensation; Tikhonov regularization; Weighted Last Square(WLS); Conjugate Gradient(CG);
【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
- 【文献出处】 电子与信息学报 ,Journal of Electronics & Information Technology , 编辑部邮箱 ,2009年01期
- 【分类号】TP391.41
- 【被引频次】12
- 【下载频次】513