一种基于椒盐噪声新型组合滤波算法仿真研究

提出了一种新型组合滤波算法。该算法首先在噪声方差估计、滤波模板类型和尺寸大小等方面对自适应维纳滤波进行改进,对图像噪声进行预处理;其次将预处理后的图像进行二维多尺度小波分解,由于低频子图像基本不受噪声污染,故不作处理;然后对开关中值滤波分别从噪声检测、噪声分类、噪声滤波等方面进行改进,并给出具体实现步骤,用于小波域高频子图像滤波;最后将滤波后高频子图像和低频子图像进行小波系数重构。实验结果表明,两类改进滤波算法在滤波性能上均优于原始算法,在抗噪性和细节保持等方面具有一定优势。

A new combination filtering algorithm based on the salt & pepper noise in digital image

A new combination filtering algorithm is put forward.Firstly,according to the adaptive wiener filtering existing "excessive filtering"defect,the noise variance estimation, the type and size of filtering template ,etc have been improved, to preprocess the image noise. Secondly the image of preprocessed is conducted two-dimensional wavelet transform, keeping low-frequency sub images.unchanged.Then, the switch median filter is improved by noise detection, noise classification, noise filtering process. Specific implementation steps is given to filtering the noise in high-frequency wavelet domain high-frequency sub images.Finally,the low-frequency sub image and processed high-frequency sub images are reconstructed. The experiment results show that the performance of the two kings of improved filtering algorithm is superior to the original filtering algorithm. The filtering algorithm in this paper has the advantage in noise suppression.