Abstract:

In order to realize real-time surface vessels’ video image sea-fog removal, a fast video image sea-fog removal method based on guided filtering is proposed. For a single image, according to the classical atmospheric scattering physical model, the dark channel prior is used to estimate atmospheric light and initial transmittance. Then the transmittance is optimized by guided filtering, and fast defogging is obtained by solving the atmospheric scattering physical model. For video images, the frame difference method and guided filtering are combined, and fast video defogging is realized by using guided filtering to estimate fog mask of video frames. Compared with other single image and frame by frame video image defogging algorithm from subjective and objective image quality evaluation, the simulation experimental results show that the proposed algorithm can realize fast single sea-fog defogging, and improve the efficiency of video defogging.