Coordinated Underwater Dark Channel Prior for Alleviating Halos and Patch Artifacts of Challenging Image Enhancement
Jiaokuan Zhang, Hao Liu · 2022 IEEE 24th International Workshop on Multimedia Signal Processing (MMSP) · 2022
For the challenging underwater images, there are always some problems such as hazy object, low contrast, and serious noises, which often are caused by various changes in the light transmission pathway. The complex underwater environments possibly lead to scattered and refracted light, and thus some original images become blurry like a veil. It is a popular solution by applying the dark channel prior (DCP) to the dehazing process of underwater images. However, there still are some side effects such as halos and patch artifacts after DCP-based dehazing. According to the environmental dissimilarity between underwater imaging and land imaging, this paper proposes a coordinated underwater dark channel prior method so as to alleviate the side effects of the traditional DCP-based enhancement methods. By considering the differences of the light absorption in the deep sea or the shallow sea, the attenuation rates of the RGB channels are calculated and coordinated in different regions. Then, the color restoration is further implemented by a compensated gray world theory that coordinates the intensity of different RGB channels. Finally, the proposed method adopts a subpatch-wise mean filtering according to sub-patch classification. The experimental results show that compared with other enhancement methods, the proposed method can significantly alleviate the halos and patch artifacts of challenging image enhancement, and also effectively improve the objective quality of challenging underwater images.