Robust back-scattered light estimation for underwater image enhancement with polarization

The scattering of particles in the turbid underwater environment reduces the contrast of the image, which can be alleviated by underwater image enhancement. However, the noise in the imaging process often interferes with the enhancement. In this paper, we propose a novel polarization-based approach that robustly estimates the background light to reduce the impact of noise. Specifically, we present a polarization parameter estimation method that adopts both spatial and frequency-domain operations. With these polarization parameters, the background light is estimated by combining the image information before and after filtering in the frequency-domain. For accurately estimating the back-scattered light at the infinity point, we exploit the characteristics of the polarization difference image and use a quad-tree method to determine it. The proposed approach can reduce the influence of the biased estimated parameters, boosting the image quality effectively. Experiments demonstrate that the our method produces better visual effect and robustness than the classic polarization-based and traditional underwater enhancement algorithms for different concentrations.