一种全变差正则化流场的图像抑噪方法

卢成武; 宋国乡

doi:10.3724/SP.J.1146.2007.01039

一种全变差正则化流场的图像抑噪方法

doi: 10.3724/SP.J.1146.2007.01039

卢成武^{①②;宋国乡},
宋国乡

基金项目:

国家自然科学基金(60473119)资助课题

计量
- 文章访问数: 3166
- HTML全文浏览量: 83
- PDF下载量: 966
- 被引次数: 0
出版历程
- 收稿日期: 2007-06-25
- 修回日期: 2007-11-30
- 刊出日期: 2009-01-19

An Image Denoising Method Using Total Variation Regularization for Flow Field

Lu Cheng-wu^{①②;宋国乡},
Song Guo-xiang

摘要

摘要: 利用Meyer的图像分解理论，提出一种磨光流场的全变差正则化抑噪方法。该方法首先引入负指数Hilbert- Sobolev范数度量逼近项，对图像水平曲线的法向量场进行全变差正则化磨光，然后构造出一个曲面拟合模型，拟合磨光后的流场。最后，利用有限差分法对各模型所导出Euler-Lagrange方程进行数值求解。实验结果表明，该方法在有效去噪的同时，使边缘和纹理信息均得到较好的保持。
- 图像分解;全变差;流场;Hilbert-Sobolev范数
Abstract: Using image decomposition theory proposed by Meyer, a total variation image denoising method based on smoothing flow field is presented. Firstly, through applying Hilbert-Sobolev norm to measure fidelity term, a total variation filter is used to smooth the normal vectors of the level curves of a noise image. And then, a model is constructed to find a surface which fit smoothed normal vectors. Finally, finite difference schemes are used to solve the Euler-Lagrange functions derived from above models. The experiments show that the approach not only can remove noisy efficiently, but also can retain edges and texture.

HTML全文

参考文献(1)

Acar R and Vogel C. Analysis of total variation penaltymethods [J].Inverse Problems.1994, 10(6):1217-1229[2]Chambolle A, DeVore R A, and Lee N Y, et al.. Nonlinearwavelet image processing: variational problems, compression,and noise removal through wavelet shrinkage [J].IEEETrans. on Image Processing.1998, 7(3):319-335[3]Chambolle A and Lions P L. Image recovery via totalvariation minimization and related problems[J]. NumerischeMathematik, 1997, 76(2): 167-188.[4]Vese L A and Osher S. Numerical methods for p-harmonicflows and applications to image processing [J].SIAM J.Numer. Anal.2002, 40(6):2085-2104[5]Osher S, Sole A, and Vese L A. Image decomposition andrestoration using total variation minimization and the e H.1norm [J]. Multiscale Modelling and Simulation, 2003, 1(3):1579-1590.[6]Osher S, Burger M, and Goldfarb D, et al.. An iterativeregularization method for total variation based imagerestoration [J].Multiscale Modelling and Simulation.2005, 4(2):460-489[7]Lysaker M, Osher S, and Tai X C. Noise removal usingsmoothed normal and surface fitting [J].IEEE Trans. onImage Processing.2004, 13(10):1345-1357[8]Rudin L I, Osher S, and Fatemi E. Nonlinear total variationbased noise removal algorithms [J].Phys. D.1992, 60(1-4):259-268[9]Meyer Y. Oscillating pattern in image processing andnonlinear evolution equations [M]. Boston: UniversityLecture Series, American Mathematical Society, 2001: 23-78.[10]Tasdizen T, Whitaker R, and Burchard P. Geometric surfaceprocessing via normal maps [J].ACM Trans. Graph.2003,22(4):1012-1033[11]Chan T, Marquina A, and Mulet P. High-order totalvariation-based image restoration [J].SIAM J. Sci. Compu.2000, 22(2):503-516[12]Besl P and Jain R. Segmentation through variable-ordersurface fitting [J].IEEE Trans. on Pattern Anal. MachineIntell.1988, 10(2):167-192

施引文献

资源附件(0)

访问统计