混合滤波图像去噪方法

Mixed Filtering Method for Removing Images Noise

针对传统的小波去噪方法会产生马赛克的问题,根据小波去噪理论以及阈值去噪理论,设计了一种基于小波阈值的混合滤波图像去噪方法.通过计算发现,新方法的峰值信噪比比传统的几种去噪方法更加优异.仿真实验证明,改进的去噪方法能够消除传统小波在阈值去噪时所产生的马赛克现象.最后分析了分解层数以及小波基对该方法去噪效果的影响,发现当分解层数为三层、采用Bior4.4小波基时去噪效果最为理想,此时峰值信噪比为35.670 2.

For the problem of mosaic phenomenon produced during the traditional de-noising methods, a novel wavelet de-noising method was introduced in this paper based on the wavelet de-noising theory and the threshold denoising theory. Compared with the traditional de-noising method, the peak value Signal-to-Noise Ratio （PSNR） of the new method showed excellent characteristics saic phenomenon in experiment while keep the layers and wavelet radix on de-noising was inves through calculation. The improved method could eliminate the mo- Signal-to-Noise Ratio （SNR） equal. The effect of decomposition tigated. When treated with three decomposition layers and Bior4. 4 wavelet radix, the best experimental result was acquired with SNR of 35.670 2.