Pixel density based trimmed median filter for removal of noise from surface image

Noise may occur in an image during the image acquisition process. In most of the cases, during acquisition process in industrial automation, the SPN may affects surface image. For elimination of noise present in the surface/work-piece image, various filtering methods have been used by many researchers. Here, a pixel density-based trimmed median filter (PDBTMF) was proposed, which works in the two different stages . In the first stage, the test pixel is diagnosed whether it is corrupted by the salt and pepper noise (SPN) or not (value of the test pixel is 0 or 255). In this paper, a 3 × 3 window is taken as the processing pixel as a center element and check for the similar pixels in the window. This algorithm works well for low-to-high-density impulse noise levels. The proposed algorithm, PDBTMF for reduction of noise in images shows good results in the elimination of SPN in grayscale, color images. The results obtained from the proposed PDBTMF shows better results as compared with the present existing methods such as adaptive weighted mean filter (AWMF), Decision Based Unsymmetrical Trimmed Variants Filter (DBUTVF), modified decision-based unsymmetric trimmed median filter (MDBUTMF), and Noise adaptive fuzzy switching median filter (NAFSMF). The proposed method, PDBTMF is experienced alongside various grayscale images and the color images and it exhibits a high Peak Signal-to-Noise Ratio, low mean square error and better Structural Similarity Index, image enhancement factor and Correlation Index.