Impulsive Noise Removal With An Adaptive Weighted Arithmetic Mean Operator For Any Noise Density

Many computer vision algorithms which are not robust to noise incorporate a noise removal stage in their workflow to avoid distortions in the final result. In the last decade, many filters for salt-and-pepper noise removal have been proposed. In this paper, a novel filter based on the weighted arithmetic mean aggregation function and the fuzzy mathematical morphology is proposed. The performance of the proposed filter is highly competitive when compared with other state-of-the-art filters regardless of the amount of salt-and-pepper noise present in the image, achieving notable results for any noise density from 5% to 98%. A statistical analysis based on some objective restoration measures supports that this filter surpasses several state-of-the-art filters for most of the noise levels considered in the comparison experiments.