An Efficient Denoising Algorithm for Detection and Removal of High-Density Impulse Noise from Digital Images

Background/Objectives: Images that are affected by noise should be denoised in order to obtain information from it. In this paper an efficient denoising algorithm based on adaptive and unsymmetric trimmed median filter is proposed which effectively gets rid off the high-density salt and pepper noise. Methods/Stastical Analysis: The proposed algorithm restores the noisy pixel with the trimmed median value when, 0’s and 255’s are found in the neighborhood and also adaptively controls the neighborhood size based on the relative amount of noisy pixels in the selected neighborhood. Many of the existing algorithms do not preserve fine details of the image while removing noise at high noise densities. Results/Findings: The Proposed Algorithm (PA) is tested against grayscale and color images and it gives an exceptional Peak Signal to Noise Ratio (PSNR) and Image Enhancement Factor (IEF) especially at higher noise densities. Conclusion/Application: This algorithm effectively removes high density noise caused by bit error transmission or introduced during the image acquisition stage.