基于方向信息自适应2次噪点检测的降噪方法

A Denoising Algorithm Based on Adaptive Two-step Noise Detection According to Directional Information

椒盐噪声的处理是图像噪声处理的重要环节,针对传统中值滤波方法存在的不足,提出1种基于方向信息自适应2次噪声点检测去椒盐噪声的方法.该方法在噪声密度不同时进行分类处理,小于某一阈值时通过方向信息区别边缘点和噪声点,降低噪声点误判的概率,在噪声密度大于某一阈值时,将所有可疑噪声点全部确定为噪声点.之后对确定的噪声点进行改进的自适应中值滤波,根据窗口中非噪声点的密度不同采用不同的滤波方法.在处理加入噪声密度为80%的Lena图时,PSNR达到28.97 d B,SSIM达到0.981 2,实验结果表明,能够去除图像中的椒盐噪声,对不同密度噪声降噪鲁棒性较强.

Denoising of salt and pepper noise is a very important step in image processing. Aiming at the disadvantage of the traditional median filter, a new denoising algorithm of salt and pepper noise based on adaptive two-step noise detection according to directional information is proposed. The algorithm adopts different methods to denoise salt and pepper noise when the noise density is not same. When noise density is less than a certain threshold, edge points and noise points are differentiated by direction information in this method which effectively reduces the probability of noise point misjudgment. When noise density is more than a certain threshold, all suspicious noise points are considered as noise points. An improved adaptive median filter is used to remove the noise points. After treated Lena added 80% noise density with this algorithm, PSNR and SSIM is 28.97 dB and 0.981 2. The experiment results shows that this algorithm can effectively remove the salt and pepper noise and has a strong robustness for different noise density.