Dark Channel Applied for Reduction of the Effects of Non-uniform Illumination in Image Binarization

Abstract

Non-uniform illumination is a common issue in images acquired in uncontrolled environments. Elimination or reduction of the non-uniform illumination problem is required in order to get an accurate image binarization. This paper introduces the combination of the dark channel and the atmospheric scattering model along with the k-means segmentation to reduce the effects of non-uniform illumination conditions in image binarization. The results show the effectiveness and robustness of this approach.