Two-dimensional gray scale image denoising via morphological operations in NSST domain & bitonic filtering

Abstract

The denoising of an image is one of the most classical and basic step in image processing. The most challenging task is to design a feature preserving denoising algorithm. This article presents an efficient denoising method derived from morphological filtering in NSST domain and Bitonic filtering. In the first stage the noisy components are processed by morphological circular disc operators i.e. Top Hat/ Bottom Hat filtering in NSST domain, as Shearlet is a powerful multi-scale and multi-directional image representation tool. The resultant image is then decomposed into 8 bit planes and each bit plane is passed through bitonic filter separately. These filtered images are assembled to obtain the final denoised image. Experimental results on standard test images substantiate that the proposed method achieves reasonable and consistent denoising performance, especially in preserving fine structure information as compared with existing algorithms specifically at high noise levels.