Mixed Noise Reduction in Gray Scale Images Using Hybrid Filters Scheme

Abstract

An interactive algorithm to restoration of noisy image is presented in this paper. This algorithm composites of two filters working together with scalar recursive Kalman filter at same time. The first one is the arithmetic median filter with a suitable window length, while the second one is the mean filter with only five neighboring point, the output of these two filter are entered to a scalar Kalman filter, which process the image with raster scan. The suggested scheme is able to reduce the mixed noise(Additive Gaussian and Impulse) from the gray scale images. The output of the scalar Kalman filter is used to adjust (modify) the median and mean filters. Different examples are given, which illustrate the performance of the proposed algorithm comparing with the single mean, single Kalman, single median filter, and median- rational hybrid filter (MRHF) using two bidirectional median sub-filters.