基于亮通道的单幅海域浓雾图像去雾算法

Defogging algorithm of single sea fog image based on bright channel

为解决浓雾天气条件下近岸海域监控系统获取的图像清晰度低的问题,提出一种基于改进亮通道及引导滤波的海域图像去雾算法。首先,将Lab颜色空间L通道作为引导图,通过对原图像进行快速引导滤波来估计大气光值;其次,基于大气光值并利用亮通道先验理论,获得亮通道透射率函数;最后,通过改进的加权引导滤波优化亮通道透射率函数,利用大气散射模型,实现模糊图像的去雾,并利用伽马校正增强去雾结果。实验结果表明,改进方法图像去雾结果视觉效果明显,可以有效减少细节信息损失,提高图像清晰度。与经典算法最优结果对比,NIQE、对比度及信息熵分别提升10.92%、4.12%和4.37%,对完善近岸海域监控系统具有一定的理论价值和现实意义。

In order to solve the problem of low image definition obtained by coastal sea area monitoring system under dense fog weather,a sea area image defogging algorithm based on improved bright channel and guided filtering was proposed.Firstly,the L channel of lab color space was used as the guide map,and the atmospheric light value was estimated by fast guide filtering of the original image.Secondly,based on the atmospheric light value and using the prior theory of bright channel,the transmittance function of bright channel was obtained.Finally,the transmittance function of the bright channel was optimized by the improved weighted guidance filter,the fog removal of the blurred image was realized by using the atmospheric scattering model,and the fog removal result was enhanced by gamma correction.The experimental results show that this method has obvious visual effect,which can effectively reduce the loss of detail information and improve the image definition.Compared with the optimal results of classical algorithms,NIQE,contrast and information entropy are increased by 10.92%,4.12%and 4.37%respectively,which has certain theoretical value and practical significance for improving the coastal area monitoring system.