Underwater Image Enhancement Using SIFT and HOG Transformation

An image fusion method combines the data from a sample consisting of source images using the features of information/pixel or interference level approaches. The image obtained by this method is a single image. The purpose of an image fusion algorithm is to integrate the redundant and matching data collected from the source representative to produce a novel picture. This picture gives better interpretation of vision for person or system point of view. It is possible to refine the quality of undersea pictures using white balance and image blending methodologies. To perform image blending, the feature matching approach can be applied. To do feature matching, the HOG and SIFT methods can be used. After being implemented in MATLAB software, the methodologies presented in this work are evaluated by means of specific metrics.