The estimation of the degree of polarization of backscattered light and target light is the key to realize scattering removal in underwater polarimetric imaging. In this paper, a new polarization scattering removal method is proposed by realizing polarimetric calculation of the target light with each pixel of the scene automatically, which is conducive to the restoration of the details of underwater scene. The global model of pixel level is established by utilizing the gradient prior of the polarization information of the total radiance image. Different from the previous study that the degree of polarization of the target light is a constant, the global calculation method can solve the global variable image of the degree of polarization of the target light and it is used for reconstruction experiments of underwater real-world scene to verify the effectiveness of the algorithm. The experimental results demonstrate that the global pixel calculation method has obvious effect on image detail recovery of the underwater scene to realize the clear underwater vision. The influence of the backscattered light on underwater imaging is well suppressed and image contrast has been significantly improved, which provides a new method for enhancing performance of underwater polarimetric imaging with better vision.

中文翻译：

---

用于水下图像复原的全局像素极化计算方法

背向散射光和目标光偏振度的估计是水下偏振成像中实现散射去除的关键。本文提出了一种新的偏振散射去除方法，通过自动实现目标光与场景每个像素的偏振计算，有利于水下场景细节的还原。利用总辐射图像偏振信息的梯度先验建立像素级全局模型。与之前研究目标光的偏振度为常数不同，全局计算方法可以求解目标光偏振度的全局可变图像，用于水下真实世界场景的重建实验，验证算法的有效性。实验结果表明，全局像素计算方法对水下场景的图像细节恢复有明显的效果，可以实现清晰的水下视觉。后向散射光对水下成像的影响得到很好的抑制，图像对比度得到显着提高，为提高水下偏振成像的性能提供了一种新的方法，具有更好的视觉效果。实验结果表明，全局像素计算方法对水下场景的图像细节恢复有明显的效果，可以实现清晰的水下视觉。后向散射光对水下成像的影响得到很好的抑制，图像对比度得到显着提高，为提高水下偏振成像的性能提供了一种新的方法，具有更好的视觉效果。实验结果表明，全局像素计算方法对水下场景的图像细节恢复有明显的效果，可以实现清晰的水下视觉。后向散射光对水下成像的影响得到很好的抑制，图像对比度得到显着提高，为提高水下偏振成像的性能提供了一种新的方法，具有更好的视觉效果。