摘要
方向小波是在多尺度几何分析(MGA)的思想下产生的一种新的图像处理工具。针对含高斯噪声图像的方向小波系数仍服从高斯分布的特点,提出一种方向小波域的维纳滤波去噪方法。修正了多方向框架,给出了维纳滤波处理方向小波系数的方法及过程。并将其分别应用于不同纹理复杂程度和信噪比的图像,利用信噪比进行了验证。实践证明了方向小波域的维纳滤波算法具有计算简单、去噪效果好、保持纹理等特点。
Directionalet is a new image processing technology proposed based on the Multi - scale Geometric Analysis（MGA）. Considering that the directionalet coefficients of image with Gaussian noise obeys the Gaussian distribution, an algorithm is proposed using Wiener filtering in directionalet domain for image denoising. The multidirectional frames are improved. After that, the method and the process of processing directionalet coefficients with Wiener filtering are given. The method is used for the images with different texture levels and SNRs. The experiment shows that this method has several advantages including simple computation, outperformed denoising effect and restoring textures of image.
出处
《电光与控制》
北大核心
2007年第6期26-28,共3页
Electronics Optics & Control
关键词
图像处理
方向小波
维纳滤波
多方向框架
image processing
directionalet
Wiener filtering
muhidirectional frames
