基于改进中值滤波和提升小波变换的阈值去噪方法研究

Threshold de-noising method for mixed noise based on improved median filter and lifting wavelet transform

摘要: 对于实际拍摄的一些图像信噪比低，噪声密度大，且含有混合噪声，而现有算法大多只能去除单一噪声的问题。针对混合噪声中含有的脉冲噪声和高斯噪声，提出基于改进中值滤波和提升小波变换去噪相结合的方法。去噪过程中，使用中值滤波器提取脉冲噪声并采用中值滤波算法滤波后，构造提升小波，采用改进阈值函数提升小波阈值去噪方法去除高斯噪声。实验结果表明，当噪声值(，)=(0.4, 20)时，采用本文去噪方法，峰值信噪比（PSNR）为34.002 1，平均绝对误差（MAE）为2.365 3。

Abstract: Images taken actually always have low signal-noise ratio(SNR),large noise density and mixed noise，however, most existing algorithms can only remove a single noise. Aiming at the mixed noise containing impulse noise and Gaussian noise, an efficient algorithm for mixed noise removal in image was proposed, including space impulse noise removal and wavelet Gaussian noise removal. In the process of de-noising, an impulse noise detection algorithm based on median filter was given to filter impulse, and at the same time a lifting wavelet transform for image was applied to remove Gaussian noise. Simulation result shows when (，)=(0.4, 20), the peak SNR(PSNR) is 34.002 1 and the mean absolute error( MAE) is 2.365 3.