Fast Under Water Image Enhancement for Real Time Applications

Abstract

Ocean exploration is a major challenge that we are facing today. With advancements to fields of marine engineering and aquatic robotics, we are capable of performing autonomous and complex decision making deep underwater. Significance of online Underwater Computer Vision Algorithms is ever increasing. Underwater images, however, suffer from inaccurate colors, hazing, colour cast and degradations because of unequal absorption of light by water. Algorithms designed for detection/enhancement in the air are of no use underwater. Although a lot of underwater image enhancement algorithms have come up in recent times, most of them are not suitable for real-time applications like AUV, due to their high computational times. These algorithms are more suitable for offline analysis. In this paper, we propose an algorithm which is fast enough for real-time systems such as AUVs/ROVs and is comparable to the offline state of the art image enhancement algorithms. We will be exploring histogram equalization techniques for dehazing and automatic white balancing algorithms for color correction. UIEB (Underwater Image Enhancement Benchmark) is used for evaluation of our algorithm. The codes and results are available at https://github.com/opgp/underwater-image-processing.