黄金,周先春,吴婷,伍子锴.混合维纳滤波与改进型TV的图像去噪模型[J].电子测量与仪器学报,2017,31(10):1659-1666 |
混合维纳滤波与改进型TV的图像去噪模型 |
Image denoising model based on mixing Wiener filtering and improved total variation |
|
DOI:10.13382/j.jemi.2017.10.020 |
中文关键词: 维纳滤波 全变差 阶梯效应 图像去噪 |
英文关键词:Wiener filtering total variation(TV) staircase effect image denoising |
基金项目:国家自然科学基金(11202106)、东南大学基本科研业务费(CDLS 2016 03)资助项目 |
|
Author | Institution |
Huang Jin | School of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China |
Zhou Xianchun | 1. School of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China; 2.Ministry of Education Key Laboratory of Child Development and Learning Science, Nanjing 210044,China; 3.Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology, Nanjing University of Information Science and Technology, Nanjing 210044, China |
Wu Ting | School of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China |
Wu Zikai | School of Electronic and Information Engineering, Nanjing University of Information Science and Technology, Nanjing 210044, China |
|
摘要点击次数: 3053 |
全文下载次数: 14465 |
中文摘要: |
在图像去噪处理过程中,为了保持图像的边缘及内部纹理信息,提出一种基于全变差改进的加权维纳滤波图像去噪模型。提出的模型利用加权项将维纳滤波与改进后的全变差模型相结合,通过构建新算子建立新的扩散模型使得图像每一个像素点的梯度信息可以自适应地选择去噪的最佳模式来平滑噪声图像,既能够在保护边缘的条件下预先处理高斯噪声,同时可以克服全变差模型的“阶梯效应”。结果表明,新模型不仅能够有效去除噪声,强化边缘还有效地保证了边缘结构的细节信息。在峰值信号噪声比测试中,该模型较之于传统线性滤波法的信噪比提高了20 dB左右,均方差也大幅降低,更具理想性。 |
英文摘要: |
A new image denoising model based on mixing Wiener filtering and improved total variation is proposed for protecting the information of image’s edge and texture in image processing.The proposed model combined Wiener filtering and improved total variation with an optimal weighting value,which was made by new operator to establish diffusion model, and could adaptively select the most appropriate denoising scheme by utilizing the gradient information of each pixel point in the image,remove Gaussian noise while keep edge structures in advance and overcome the “staircase effect” of total variation model.The simulation results indicate that the proposed model can improved enoising performance,protect edge’s structure, and retain more texture details of the image effectively.Compared with other classical linear filtering method,the new model is ideal, the peak signal to noise ratio is improved by 20 dB and the mean square error decreases sharply in the test. |
查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|