杨斌,罗颉.基于压缩传感理论的彩色图像去马赛克[J].南华大学学报(自然科学版),2013,27(2):32~36.[YANG Bin,LUO Jie.Color Image Demosaicing Based on Compressed Sensing Theory[J].Journal of University of South China(Science and Technology),2013,27(2):32~36.] |
基于压缩传感理论的彩色图像去马赛克 |
Color Image Demosaicing Based on Compressed Sensing Theory |
投稿时间:2013-04-10 |
DOI: |
中文关键词: 压缩传感 稀疏表示 去马赛克 |
英文关键词:compressed sensing sparse representation demosaicing |
基金项目:国家自然科学基金资助项目(61102108);湖南省教育厅基金资助项目(11C1101);南华大学博士科研启动基金资助项目(2011XQD29) |
|
摘要点击次数: 1127 |
全文下载次数: 1864 |
中文摘要: |
利用目前最新的压缩传感理论在信号恢复方面的优势,本文提出了一种新型彩色图像去马赛克算法.首先,为了满足压缩传感的约束等距性条件,采用随机马赛克模式得到原始数据;然后把彩色图像去马赛克问题转化为基于压缩传感的信号恢复问题,并充分考虑彩色图像RGB通道的相关性,建立彩色图像的扩展联合稀疏表示模型;最后采用基追踪(BP)算法求解去马赛克问题,得到恢复的真彩色图像.实验证明,本算法能有效减少传统算法恢复图像中出现的锯齿现象,所得到的彩色恢复图像具有更好的视觉效果.客观评价结果显示,本文算法结果的峰值信噪较基于双线性插值和基于梯度等传统算法的结果有明显提高.实验结果验证了本文算法的优越性,显示算法具有重要的应用前景。 |
英文摘要: |
This paper proposes a new color image demosaicing algorithm,taking the advantage of the latest compressed sensing on signal recovery.First,in order to make the demosaicing algorithm satisfy the restricted isometry property of compressed sensing,a new random color filter array (CFA) pattern is used to replace the traditional Bayer Pattern.Then,we transform the demosaicing problem into a signal restoration framework based on compressed sensing.In addition,an extended joint sparse model of the color image is established by considering the inter-channel correlation of different color channels.Finally,the basic pursuit (BP) algorithm is used to reconstruct the color image.The experimental results show that the proposed method can effectively reduce the aliasing appears on the traditional demosaicing algorithm and could achieve better visual qualities.The objective evaluation results show the proposed method is better than those traditional ones such as the bilinear interpolation and the gradient based interpolation in terms of the peak signal-to-noise.The results clearly indicate the superiority and the application prospects of the proposed algorithm. |
查看全文 查看/发表评论 下载PDF阅读器 |
关闭 |