A Comparison some Filters to Remove Noise for Digital Images Employed some Statistical Techniques

Abstract

In this research, two images (saturn, Lena) (original image) were selected on the basis that they are digital images in black and white and colored size (256 × 256 pixels) and added Gaussian noise and different noise ratios (0.25,0.50, 0.80) to study the process of improving the digital image with mean a zero to the original image, a distorted image with different noise ratios is produced, and then the noise is removed from the image employment a number of digital noise removal filters (Average Filter), the Median filter (Median Filter), (Weiner Filter), And finally the LPG-PCA, Then employment the quality criterion (PSNR) criterion, the MSE criterion, and the AIC criterion to measure the improvement made on the image after using the filters to remove noise and compare the performance of digital filters and choose the best, finally, some statistical measures have been calculated to describe the digital image (Mean, Median, Variance), and it was concluded that the LPG-PCA Filter method was the best method for removing caustic noise from other methods and for different noise ratios of 25%, 50%, and 80% as it achieved the lowest value for the (MSE) , the (AIC ) and (Variance), and The highest value is for (PSNR), (Mean) and (Median) And for the two pictures (saturn, Lena).