摘要
针对图像处理中存在椒盐噪声和高斯噪声的问题,提出了一种新的小波阈值函数与改进中值滤波融合的噪声抑制算法。依据椒盐噪声与其周围邻域像素灰度值存在的明显差异,估计椒盐噪声,采取3×3模板进行反复迭代滤除椒盐噪声。对高斯噪声,利用极限思想提出了一种新的小波阈值函数,并引入三个控制变量,通过调节控制变量使所构成的小波系数在一定阈值范围内无限接近原小波系数。实验结果表明,提高了峰值信噪比,减小了均方误差,较好的保留了图像细节。
For the presence of salt-and-pepper noise and Gaussian noise in image processing, an algorithm of sup- pressing image noise is proposed based on the integration of wavelet threshold function and improved median filter. Based on the significant differences of pixel grayscale values between the salt-and-pepper noise and its surroundings, the salt- and-pepper noise and take the 3 × 3 template iterative filter out salt-and-pepper noise are estimated. To processing Gaussi- an noise a new threshold function of using the thought of limit is proposed. Three control variables are used. By used con- trol variables the new composition of the wavelet coefficients infinitely is closed to the original wavelet coefficients within a certain threshold range. Experimental results show that the peak signal to noise ratio (PSNR) is increased and the mean square error (MSE) is decreased, image detail preserved is better.
作者
马良慧
李东兴
张华强
杜钦君
常晓刚
MA Lianghui LI Dongxing ZHANG Huaqiang DU Qinjun CHANG Xiaogang(Shandong University of Technology, Zibo 255049, Chin)
出处
《光学技术》
CAS
CSCD
北大核心
2017年第1期38-42,共5页
Optical Technique
基金
国家自然科学基金(61175090
61302162)
山东省自然科学基金(ZR2014EL024
ZR2015FL012
ZR2014JL027)
山东省高等学校科技计划项目(J14LN76)
关键词
混合噪声
改进中值滤波
小波阈值
峰值信噪比
mixed noise
improved median filter
wavelet threshold
peak signal to noise ratio (PSNR)