Removal of high density salt-and-pepper noise by recursive enhanced median filtering

In many applications the existence of impulsive noise in the acquired images is one of the most common problems. In present scenario median filters are very functional solution for the removal of impulse noise in the images. In general procedure firstly the noise is detected and then removed secondly. This paper proposes a novel method with two thresholds (Minimum threshold and maximum threshold) which detects noise more efficiently. Further the enhanced median value is use to replace the noisy pixels recursively. It furnishes a better Peak Signal to Noise Ratio (PSNR) as compared to many existing median filter based image de-noising methods.