太阳射电爆发中图像网纹消除的小波NeighShrink方法  

Wavelet NeighShrink Method for Grid Texture Removal in Image of Solar Radio Bursts

作　　者：赵瑞珍[1] 胡占义[2] 

机构地区：[1]北京交通大学计算机与信息技术学院,北京100044 [2]中国科学院自动化研究所模式识别国家重点实验室,北京100080

出　　处：《光谱学与光谱分析》2007年第1期198-201,共4页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(60402041);北京交通大学科技基金项目(2005SM011)资助项目

摘　　要：绝大多数观测得到的太阳爆发数据中均含有大量的噪声,这给数据的后续处理带来极大的困难。文章分析和研究了NeighShrink阈值函数的特点以及邻域窗口大小的选择依据,在此基础上提出了一种新的小波NeighShrink平方根阈值方法用于图像去噪。首先对太阳爆发灰度图中的每一通道作了规范化处理,在一定程度上去除由于通道间的差异造成的横条纹,然后将预处理后的图像进行小波分解,采用小波Neigh-Shrink平方根阈值函数对其小波系数作阈值处理,最后利用小波反变换恢复图像。实验结果表明该方法可以有效地实现去除干扰、增强有用信息的目的。The data received from solar bursts contain a lot of noise, which makes further processing more difficult. To remove the noise and enhance the image, we studied the properties of the NeighShrink threshold function and analyzed the influence of neighborhood window size on the denoising result, on the basis of which a new wavelet NeighShrink square root method for image denoising is presented. Firstly, each channel of the solar burst image is normalized, which can, to some extent, remove the horizontal grid texture in the image. Secondly, the preprocessed image is decomposed by wavelet transform, and the obtained wavelet coefficients are thresholded by NeighShrink square root method. Finally, the denoised image is reconstructed by inverse wavelet transform. The experimental results show that the presented method is effective in noise removal and image enhancement.

关 键 词：小波变换 NeighShrink 噪声消除 图像增强 

分 类 号：TN911.7[电子电信—通信与信息系统]