基于12个方向的方向加权滤波法去除随机值脉冲噪声

Directional weighted filter with 12 directions to remove random valued impulse noise

为了解决方向加权中值滤波方法的不足,提出了一种新的基于12个方向的方向加权滤波法滤除图像中的随机值脉冲噪声。这种方法在脉冲检测阶段是利用当前像素与邻域12个方向像素的不同识别噪声,然后在噪声滤除阶段选取合适的加权值去除噪声。实验结果表明,该方法从主客观两方面验证了在滤除噪声率较高或较低的噪声时对保护图像细节方面的效果有所提高,并且提高了程序运行效率。

In order to remove the lack of directional weighted median filter method, this paper proposed a new directional weigh- ted filter with 12 directions. It could remove the random value impulse noise in digital images. In this method, it identified the noise pixels by the differences between the current pixel and neighbors aligned with 12 directions in the impulse detection phase, and then it selected the appropriate weighted value for noise removal in the noise filtering phase. The extensive experimental re- suits show that this method has improved the effect of protecting image details on subjective and objective aspects when the noise rate is higher or lower, and has also improved the effieieney of program.