Abstract:

Digital images are often corrupted by impulse noise during the acquisition or transmission. In order to improve the image quality, it is necessary to remove impulse noise. Aiming at the shortcoming of traditional median filter and the other medianbased nonlinear filters in denoising image, an improved denoising algorithm is proposed in this paper. Impulse noise checking is made before filtering so as to confirm the pixels polluted by noise, which are recorded and marked. Then, according to the nose detection results, an improved median filtering is adopted to remove the impulse noise. This proposed algorithm only handle the pixels polluted by noise. It not only make use of the standard median, but also use the last value before the standard median and the first value after the standard median. Experiments proves that the improved method can not only provide better performance of suppressing impulse noise than other medianbased nonlinear filters, but also preserve more detail features.

Key words: image denoising;impulse noise;noise detection;adaptive threshold;median filtering