关键词: 水下图像增强, 深度学习, 图像增强方法, 评价体系

Abstract: Underwater image enhancement technology has become a research hotspot due to its extensive applications in fields such as marine exploration, underwater robotics, and underwater monitoring systems. However, the complexity of underwater environments, such as uneven lighting, light scattering and absorption, often leads to images with low contrast, blurred details, and color distortion, which significantly reduces the visibility and interpretability of images. To address these challenges, researchers have proposed a variety of enhancement strategies aimed at improving the quality of underwater images to make them more suitable for visual analysis and automated processing. This article systematically reviews the recent research progress in the field of underwater image enhancement, categorizing these methods into three types: physical model methods, non-physical model methods, and deep learning methods. Among them, physical model methods construct a physical model of underwater imaging process, such as the Jaffe-McGlamery model and the dark channel prior (DCP) model, to accurately estimate scene parameters. Non-physical model methods do not rely on an in-depth understan ding of physical processes. Instead, they employ techniques such as histogram equalization and wavelet transform to enhance images through operations in the spatial domain or transform domain. Deep learning methods utilize frameworks such as convolutional neural networks (CNN) and generative adversarial networks (GAN) to significantly improve image quality by learning image features from a large amount of data. In addition, this paper comprehensively reviews and summarizes the currently commonly used subjective and objective evaluation systems as well as specific indicators for underwater image quality. Finally, this paper provides an in-depth analysis of the current challenges faced in the field of underwater image enhancement technology, and looks forward to the development trends of future research directions. It is believed that with the continuous advancement of marine exploration and underwater technology, underwater image enhancement technology is expected to achieve more breakthroughs in the future, providing a stronger technical support for humanity to understand and develop marine resources.

Key words: underwater image enhancement, deep learning, image enhancement methods, evaluation system