Underwater image recovery based on wavelet decomposition and saturation estimation

In underwater imaging, the underwater image is seriously degraded due to the influence of light absorption, forward scattering and backward scattering. This paper proposes an underwater image restoration method based on wavelet decomposition and saturation estimation of transmittance. Firstly, based on the spectral characteristics of underwater images, the scattering light is suppressed by estimating transmittance with saturation in the lowest frequency subband. The noise is suppressed and the details are enhanced in the high frequency subband with the help of wavelet decomposition.Secondly, automatic white balance and saturation enhancement histogram matching are used to enhance the underwater image, so as to improve the color deviation and contrast deviation of the underwater image. The average values of NIQE and UCIQE are 3.2177 and 0.4895 respectively, which are better than other algorithms. Experimental results show that this method can get more natural color and more balanced saturation of the restored image.