小波框架方法的彩色图像泊松去噪  

Wavelet frame based color image denoising with Poisson noise

作　　者：高涛[1] 何秀丽[1] 

机构地区：[1]河海大学理学院,南京210098

出　　处：《信息技术》2017年第5期162-165,共4页Information Technology

摘　　要：噪声是影响图像质量的重要因素,噪声的存在导致图像的某些特征细节不能被辨识,图像的信噪比下降,所以图像去噪是图像处理中的一个重要问题。文中提出了基于小波框架的彩色图像的泊松去噪模型。在该模型中,赋权的l^2项作为保真项,包含小波框架的l^1项作为正则项。同时,又提出了解该模型的重新赋权的split Bregman算法。最后,利用仿真实验以及PSNR指标对该模型的去噪效果进行评估,评估结果表明该算法可行、有效。Noise is an important factor to affect the image quality.Noise causes that some feature details of images cannot be identified and signal to noise ratio is declined.Therefore,image denoising is an important issue in image processing.In the following,this paper proposed a wavelet frame based Poisson denoised model in color images.With this model,we acts weighted l1 as fidelity term and l1 item as a regularization item contains wavelet frame.Meanwhile,it also proposed reweighted split Bregman algorithm to solve this model.Finally,it provided supporting numerical experiments and PSNR indicators to assess the model evaluation results.The results show that the algorithm is feasible and effective.

关 键 词：小波框架 SPLIT Bregman算法 彩色图像 泊松噪声 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]