Underwater Image Enhancement using Image Fusion Technique

Abstract

Underwater photography has been around long enough for all of the problems to be recognized and their respective solutions be discovered. Water being multiple times heavier and viscous than air, it has the ability to retain particles in suspension through which the light required for the purpose of imaging needs to travel. This causes the light levels to drop drastically the deeper we go, a phenomenon called as Backscatter. Other problems such as loss of color, loss of contrast, Refraction, Magnification, flooding etc., creep up in underwater imaging. The Proposed method is a single-image approach that doesn’t require special hardware and knowledge of underwater conditions or scene structure. It is practically a blend of multiple images which are themselves derived from the color-compensated and white balanced version of the original input image. To avoid the sharp weight map transitions that results in creation of artifacts in the lower frequency components of the reconstructed image, we adapt a Multiple-scale fusion approach. The method enhances the contrast and sharpness of the image. The enhanced result is further improved by color-correction, lightened dark regions, naturalness preservation, and well enhanced edges and details.