一种去除图像混合噪声的新方法

某一种去噪方法通常只对某一类噪声的滤除较为有效,为了抑制混合噪声,提出一种结合中值滤波与小波去噪的图像去噪算法。该算法首先检测出脉冲噪声点,并采用自适应窗口对脉冲噪声点进行中值滤波处理,然后用基于高斯混合模型的小波去噪法滤除图像中的高斯白噪声。仿真实验表明,对于被高斯、脉冲混合噪声污染的图像,该算法的去噪效果显然比单一的中值滤波和小波去噪法好得多。

An Effective Method for Removing Mixed Noise in Images

One denoising method is only effective to one kind of noise.In order to suppress mixed noise,an image denoising algorithm combining median filter with wavelet denoising is proposed.The algorithm detects impulse noise points first,and median filter with adaptive window is applied to the points,then wavelet denoising based on Gaussian Mixture Model is used to filter the Gaussian white noise in images.Simulation experiments show that the algorithm is obviously much better than single median filter or wavelet denoising method when filtering mixed noise of Gaussian and impulse in images.