Quality Enhancement for Underwater Images using Various Image Processing Techniques: A Survey

Abstract

Underwater images are essential to identify the activity of underwater objects. It played a vital role to explore and utilizing aquatic resources. The underwater images have features such as low contrast, different noises, and object imbalance due to lack of light intensity. CNN-based in-deep learning approaches have improved underwater low-resolution photos during the last decade. Nevertheless, still, those techniques have some problems, such as high MSE, PSNT and high SSIM error rate. They solve the problem using different experimental analyses; various methods are studied that effectively treat different underwater image distorted scenes and improve contrast and color deviation compared to other algorithms. In terms of the color richness of the resulting images and the execution time, there are still deficiencies with the latest algorithm. In future work, the structure of our algorithm will be further adjusted to shorten the execution time, and optimization of the color compensation method under different color deviations will also be the focus of future research. With the wide application of underwater vision in different scientific research fields, underwater image enhancement can play an increasingly significant role in the process of image processing in underwater research and underwater archaeology. Most of the target images of the current algorithms are shallow water images. When the artificial light source is added to deep water images, the raw images will face more diverse noises, and image enhancement will face more challenges. As a result, this study investigates the numerous existing systems used for quality enhancement of underwater mages using various image processing techniques. We find various gaps and challenges of current systems and build the enhancement of this research for future improvement. Aa a result of this overview is to define the future problem statement to enhance this research and overcome the challenges faced by previous researchers. On other hand also improve the accuracy in terms of reducing MSE and enhancing PSNR etc.