Contrast Enhancement Based on a Morphological Rational Multiscale Algorithm

Abstract

Abstract. Contrast enhancement is an important task in image processing and it is commonly used as a preprocessing step in order to improve the results for other tasks such as segmentation. However, not only do some methods for contrast improvement have good performance working on low contrast regions, but they also affect good contrast regions; owing to the fact that some elements could be vanished, representing a loss of information. A method focused on images with different luminance conditions is introduced in the present work. The proposed method is based on morphological filters by reconstruction and rational operations, which together, allow a uniform contrast enhancement. Furthermore, due to the properties of these morphological transformations, the creation of new elements on image is avoided. The processing was made on luminance values in the u’v’Y’ color space, which permits to keep the chrominance and to avoid the creation of new colors. As a result of the previous considerations, the proposed method keeps the natural color appearance of the image.