利用偏微分方程的Tetrolet变换图像去噪  

Tetrolet Shrinkage with Partial Differential Equations for Image Denoising

作　　者：李财莲[1,2] 孙即祥[1] 康耀红[2] 李智勇[1] 

机构地区：[1]国防科学技术大学电子科学与工程学院,湖南长沙410073 [2]海南大学信息科学技术学院,海南海口570228

出　　处：《海南大学学报（自然科学版）》2011年第2期166-171,共6页Natural Science Journal of Hainan University

基　　金：国家自然科学基金项目(40901216);国防预研资助项目(513220206)

摘　　要：对图像进行Tetrolet变换后利用偏微分方程对图像进行了质量改善,仿真结果表明,该算法不仅能有效去除噪声,而且可得到更高的峰值信噪比和更好的视觉效果,去噪后图像较光滑,减少了方块效应,更多地保留了图像边缘和细节等局部特征.After tetrolet transform was applied to the noise images,the conventional smooth shrinkage results were further processed by partial differential equations. The simulation results indicated that the method not only remove the noise effectively, but also can obtain higher PSNR （Peak Signal to Noise Ratio） and better visual effects. Compared with typical tetrolet transform, the denoised images by our method were smoother, and the blocking artifacts were reduced, and which preserved more significant information of original images, such as ed- ges and details

关 键 词：Tetrolet变换 方块效应 偏微分方程 图像去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]