基于相关性的自适应阈值去噪方法研究  

Research on adaptive threshold denoising method based on correlation

作　　者：袁丽英[1] 张蕊[1] 徐鹏[1] 刘仲杰[1] 荆宇[1] 

机构地区：[1]哈尔滨理工大学自动化学院,哈尔滨150080

出　　处：《哈尔滨商业大学学报（自然科学版）》2013年第5期571-574,共4页Journal of Harbin University of Commerce：Natural Sciences Edition

基　　金：黑龙江省教育厅基金项目(12521113)

摘　　要：为了更有效的去除数字图像中的高斯噪声,提出一种基于相关性的自适应阈值去噪方法,此方法对小波阈值去噪法以及基于小波系数相关性的去噪算法的不足进行改进.提出一种新的阈值函数,并对小波系数相关量的计算方法进行改进,结合尺度间相关性判断图像中少量弱相关及不相关噪声并进行去噪处理.通过与传统阈值去噪算法、改进阈值去噪算法以及基于小波系数相关性的去噪算法进行对比实验,结果显示该算法更有效的去除图像中的高斯噪声,获得更高的峰值信噪比.In order to remove gaussian noise in images more effectively,an adaptive threshold denoising method based on correlation was proposed,this method improves defects of wavelet threshold denoising method and the denoising method based on wavelet coefficients correlation.In this paper,a new threshold function and improving the calculation method of amount of wavelet coefficients correlation was put forward,combined with the multi-dimension relevance judgment to deal with a few weak and not related noise in the image.Through compared with the traditional threshold denoising method,the improved threshold denoising method and the denoising method based on wavelet coefficients correlation,results showed that this method was more effective to remove the gaussian noise,and got higher peak signal -to-noise ratio.

关 键 词：小波变换 阈值去噪 相关性 自适应 

分 类 号：TN911[电子电信—通信与信息系统]