为有效抑制椒盐噪声对图像信息的影响,根据椒盐噪声随机破坏图像中像素值的显著特征,本文提出一种耦合噪声检测的自适应模糊正则化噪声去除模型.一方面,基于L1范数建立数据保真项,实现对图像统计分布进行有效拟合.另一方面,通过对图像中像素相似性的有效量化实现图像中噪声的检测,并将此耦合至正则项中,使得模型可依据像素点实际受噪声的污染对其施加惩罚程度,最终实现椒盐噪声的自适应模糊去除.本文采用交替方向乘子法(Alternating direction method of multipliers,ADMM)进行模型的数值结果实现,并运用峰值信噪比(Peak signal-to-noise ra-tio,PSNR)及结构相似性(Structural similarity,SSIM)对实验结果进行评定.实验结果表明,本文提出的模型在PSNR及SSIM方面得到显著提升,其中对于灰度图像的去噪实验PSNR最高可提高1.3dB,SSIM最高可提高0.2.
Adaptive fuzzy regularization model of salt and pepper noise removal
In order to effectively suppress the influence of salt and pepper noise on image information,an adaptive fuzzy regularization noise removal model based on salt and pepper noise random destruction of pixel values was proposed in this paper.On the one hand,the data fidelity term is established based on theL1 norm to realize the effective fitting of the statistical distribution of the image.On the other hand,by effectively quantifying the similarity of pixels in the image,the noise detection in the image is realized,and this is coupled to the regular term,so that the model can actually impose a penalty degree on pixels according to the noise pollution,and finally realize the adaptive fuzzy removal of pepper and salt noise.The alternating direction multiplier method(ADMM)was used for the numerical result implementation of the model in this article.The Peak signal-to-noise ratio(PSNR)and Structural similarity(SSIM)were used to evaluate the experimental results.The experimental results show that the proposed model can improve PSNR and SSIM significantly with the denoising experiment for grayscale images,achieving a maximum PSNR improvement of 1.3dB and a maximum SSIM improvement of 0.2.
salt and pepper noisefuzzy regularizationnoise probability matrixadaptive total variation
万方数据
维普
