Authors: Ajay Kumar Nain; Shailender Gupta; Bharat Bhushan
Addresses: YMCA University of Science and Technology, Faridabad, India ' YMCA University of Science and Technology, Faridabad, India ' YMCA University of Science and Technology, Faridabad, India
Abstract: Noise elimination is a fundamental operation of image processing in order to enhance, analyse and interpret the important information in an image. Several types of noise have been studied but it is found that more than one noise type contaminate the images. Gaussian and impulse noises are the ones that are more often introduced as mixed noise. Despite the existence of many filtering solutions to reduce the different noise types separately, only a few methods to process mixed noise have been proposed. In addition these methods work for grey scale images. This paper proposes a technique for removing mixed noise from colour image while preserving edges at same time by modifying the Switching Bilateral Filter (SBF). The implementation is done in MATLAB-7 and compared with previous ones. It is found that the proposed technique provides better results for densely corrupted images. Moreover, the proposed technique provides less colour blurriness.
Keywords: Gaussian noise; impulse noise; mixed noise; PSNR; MSE; MAE; histograms; nonlinear filters; colour image de-noising; vector filtering; switching bilateral filters; colour images; peak SNR; signal to noise ratio; noise elimination; image processing; corrupted images; blurred images; mean square error; mean absolute error.
International Journal of Signal and Imaging Systems Engineering, 2016 Vol.9 No.1, pp.1 - 19
Available online: 08 Feb 2016Full-text access for editors Access for subscribers Purchase this article Comment on this article