摘要
本文提出了一种基于小波系数相关性的图像去噪方法。针对自适应去噪方法平滑掉实际信号的不足,利用小波分解系数近指数的衰减特性及同一尺度上各信号分量的相关性分析,对每个小波系数设置不同的阈值,区别对待信号的低频和高频系数,最后重构小波。实验结果表明,从视觉和峰值信噪比(PSNR)上,本文使用的方法优越于全局阈值法。
An image denoising method based on the wavelet coefficients was proposed in the paper.According to exponent attenuation of the wavelet coefficients in image decomposition and the correlation analysis of coefficients at the same scale,each coefficient of wavelet component was set different threshold and different algorithms for the coefficients of low and high frequencies were adopted respectively.Finally,the denoised image was obtained by wavelet reconstruction.The results showed that the proposed method was more efficient compared with traditional approaches in terms of visual quality and PSNR（Peak Signal Noise Ratio）.
出处
《测绘科学》
CSCD
北大核心
2012年第1期94-95,共2页
Science of Surveying and Mapping
基金
国土环境与灾害监测国家测绘局重点实验室开放基金资助项目(LEDM2009A01)
关键词
小波变换
相关性
图像去噪
峰值信噪比
wavelet transform
correlation
image denoising
peak signal noise ratio
