Robust polarization-based underwater image enhancement method using anchor brightness adaptation

Most of the polarization-based image enhancement methods do not take into account the inaccurate parameter estimation caused by the amplified camera noise at low illumination conditions. To solve this problem, this paper proposes a robust polarization-based underwater image enhancement method using anchor brightness adaptation (ABA). Relying on the relationship between the Stokes Vector and the angle of polarization (AoP), the proposed neighborhood high fidelity constraint (NHFC) can robustly select the region of the background light that is most suitable for parameter estimation, greatly reducing the interference of camera random noise under low illumination conditions and making full use of the characteristics of polarized optical imaging. The background light intensity at infinity and the scene transmission map are then estimated based on the polarization characteristics of the scene, thus effectively enhancing the image. Finally, ABA is introduced with the help of greyscale information of the original image to ensure the best exposure of the enhanced image. The experiments simulate underwater imaging at night, and the results show that the enhancement performance of this method is stable and can effectively solve the problem of being susceptible to random noise interference at low light, which has better applicability for real-time underwater image enhancement.