Periodic noise removal using local thresholding

Periodic noise has a global impact in spatial domain and can only be removed using frequency domain analysis. In this paper, a novel frequency domain technique is proposed for the removal of semi-periodic and quasi-periodic noise for both gray scale images and color images. This method includes auto detection using local thresholding. Weighted Gaussian Notch Filter (WGNF) is used to suppress noisy spikes to get a noise free image at the later stage. This method also includes a preprocessing step to preserve image components near origin. Proposed method reduces intensive periodic noise while preserving image components. Qualitative and quantitative analysis shows the effectiveness of the proposed model as compared to other methods.