A fast algorithm for underwater image restoration based on CUDA

Due to the severe underwater environment condition and the effects of water scattering, the images acquired by underwater optical vision system often have many problems such as low resolution, contrast and serious color shifts. Many scholars have proposed many image restoration algorithms and proved to be effective, but there is still a problem of excessive time consumption. In this regard, this article comprehensively considers the principle of dark channel underwater image imaging, combined with parallel computing related theories, and modified the serial model of the traditional underwater image restoration algorithm to be suitable for CUDA (Compute Unified Device Architecture) Parallel model. Through experimental tests, while ensuring a restoration effect similar to similar algorithms, a speed increase of about 14 times has been achieved, and the real-time restoration of underwater images is realized.