基于水下成像模型的图像清晰化算法

摘要

受水下场景中有机物和悬浮颗粒的影响,水下图像存在对比度低、颜色失真和细节丢失等问题.同时,水下场景中通常有人工光源存在,造成图像光照不均.传统基于图像去雾的方法用于水下图像复原时效果欠佳,为充分考虑水对光的吸收和散射作用,近期提出了新的水下成像模型和图像复原方法.但是这些方法未考虑红通道影响,导致估计的散射比偏大;另外,也未考虑人工光源的影响,导致估计的背景光过大.针对这些问题,该文提出一套有效的水下图像清晰化方案.首先,通过设置阈值确定是否将红通道信息用于暗通道计算,并将反映人工光源影响的饱和度指标用于散射比估计,以减小人工光源的影响.由此,提出了基于红通道预判和饱和度指标的暗通道计算方法.然后,根据三通道衰减系数比估计每个通道的透射率,可弥补目前很多方法假设蓝绿通道透射率一致的缺陷.最后,利用Shades of Gray算法估计环境光,并结合新的水下成像模型得到复原图像.实验结果表明,该文算法可显著提升图像的对比度,得到颜色自然、细节清晰的复原图像.%As a result of the existence of organisms and suspended particles under underwater conditions, images captured under water usually have low contrast, color distortion and loss of visibility. At the same time, due to the existence of the artificial light source, the underwater image usually has the non-uniform illumination. Traditional hazy-removal methods perform poorly under water. In order to take both absorption and scattering into consideration, a new underwater image formation model and restoration methods are proposed recently. However, these methods ignore the great impact of the red channel information and artificial light source. To solve this problem, a new approach is proposed for underwater image visibility restoration. Firstly, a threshold is set to determine whether to use the red channel information to estimate the dark channel, and a saturation indicator which is used to indicate the impact of artificial light source is utilized to calculate the scattering rate. Based on the red channel information anticipation and the saturation indicator, a new method is proposed to estimate the dark channel. Then, the transmission of each channel is estimated according to the attenuation coefficient ratio, which makes the proposed method more robust. Finally, the ambient light is obtained using the Shades of Gray algorithm, and the visibility restoration result is achieved based on a new underwater image formation model. Experimental results demonstrate that the proposed algorithm can significantly improve the contrast of the underwater image with more natural color and better visibility.