基于Tetrolet变换的自适应阈值去噪  

Adaptive Image Thresholding Denoising Based on Tetrolet Transform

作　　者：李财莲[1,2] 孙即祥[1] 康耀红[2] 

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）》2010年第4期348-352,357,共6页Natural Science Journal of Hainan University

摘　　要：提出一种新的阈值函数,在Tetrolet变换域中利用此阈值函数通过最小均方误差进行自适应图像去噪.此方法主要优点是利用图像Tetrolet变换的多尺度框架和方向性来保留原始图像的边缘和纹理等重要信息,同时在Tetrolet变换域中利用新阈值函数通过LMS算法来对图像自适应去噪.实验结果表明:该算法具有更高的峰值信噪比和更好的视觉效果,是有效可行的.In the paper,a new thresholding function was proposed,with which an adaptive method of image denoising was performed through the mean squared error based on the Tetrolet transform domain.This approach took advantage of a multiscale framework and directionality of Tetrolet transform to preserve the significant information of original image such as edges and details and the new thresholding function was performed for adaptive image denoising by LMS algorithm.Numerical results indicated that our proposed method had the higher PSNR values and the better visual quality,and thus it was effective and feasible.

关 键 词：Tetrolet变换 阈值函数 最小均方误差 自适应阈值 图像去噪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]