Research on the method of underwater image enhancement
Tiansong Li, Jian Yang, Haiyan Zhou, Tao Lian · Journal of Physics Conference Series · 2021
Abstract A multi-angle image enhancement method based on underwater imaging was proposed to solve the problems of color bias, geometric distortion and low contrast of the observed object in the complex and changeable underwater environment. Firstly, the gray world algorithm is used to preprocess the underwater image. Then an image optimization denoising algorithm based on K-SVD dictionary learning is used to keep the structure and texture information of the original image. Finally, combining with the underwater optical imaging model, the blue-green channel enhancement algorithm is used to realize the underwater imaging defog processing, which makes the restored image closer to the reality. The results show that this method can solve the problems of noise interference, fuzzy underwater image and low contrast.