红外技术, 2018, 40 (7): 647, 网络出版: 2018-08-04
基于非抽样双树复小波变换幅值相位信息的图像去噪算法
Image Denoising Using Magnitude-phase of the Undecimated Dual-tree Complex Wavelet Transform
图像去噪 非抽样双树复小波变换 瑞利分布模型 相位正则化 image denoising undecimated dualtree complex wavelet transform Rayleigh distribution model phase regularization
摘要
提出了一种非抽样双树复小波变换域结合幅值阈值化和相位正则化的自适应图像去噪算法。首先将非抽样双树复小波变换系数进行幅值相位表示,在分析了幅值分布特点后,使用瑞利分布模型作为系数幅值的先验分布,然后在贝叶斯去噪框架下推导出闭式形式的阈值函数,为了更好地抑制噪声,我们亦对相位信息进行平滑处理,最后通过逆非抽样双树复小波变换得到去噪图像。由于同时对幅值和相位信息进行处理,实验显示所提算法抑制噪声效果明显,与一些经典算法相比,本文方法在主、客观上皆获得了有竞争力的结果。
Abstract
: A new image denoising algorithm based on thresholding-based magnitude and phase regularization of the coefficients of undecimated dual-tree complex wavelet transform (UDTCWT) is proposed. First, the magnitude and phase of the UDTCWT coefficients are determined. The magnitude characteristics of UDTCWT coefficients follow the Rayleigh distribution pattern. With this statistical model, a closed-form shrinkage function is derived in the Bayesian theory frame work. To more effectively suppress the noise, the smoothing of phase information is also performed. Finally, the inverse UDTCWT transform is performed to get the denoised image. The simulation results demonstrate that the proposed method provides promising results and is competitive with the classical denoising methods both in terms of peak signal-to-noise ratio and visual quality.
吴建宁, 石满红, 兴志. 基于非抽样双树复小波变换幅值相位信息的图像去噪算法[J]. 红外技术, 2018, 40(7): 647. WU Jianning, SHI Manhong, XING Zhi. Image Denoising Using Magnitude-phase of the Undecimated Dual-tree Complex Wavelet Transform[J]. Infrared Technology, 2018, 40(7): 647.