Underwater Image Quality: Enhancement and Evaluation

Abstract

The underwater optical images are commonly captured by camera but with different statistical features to natural images. Due to the refraction and scattering of light in different water types, the colors and shapes of objects can be twisted that are incapable of providing acceptable visual qualities. Thus, it is imperative to develop algorithms to enhance underwater images. Besides, the quality evaluation of underwater images is also exploited as a criteria of underwater image enhancement. In the past decade, the related issues have attracted considerable attention. This paper presents a comprehensive review of the related techniques and their most recent achievements. In particular, we observe a significant trend of applying deep learning in underwater image processing in a small volume of data. We hope our review could benefit both the beginners and the experts of this area for discovering appealing and challenging research topics.