Deep learning is a powerful technique based on neural networks, which has distinct advantages in terms of finding the relation between the input and the label, and it has shown the superior performance for image recovery in complex and strong noise environment than other methods. We employ the deep learning method for polarimetric imaging in turbid media and in strong noise environment. For underwater imaging, the proposed learning-based method can effectively remove the veiling light and outperforms other existing methods, even in dense turbid water. For image denosing, the experimental results show that the proposed learning-based method has an evident performance on the noise suppression and outperforms other existing methods. Especially for the images of the degree of polarization and the angle of polarization, which are quite sensitive to the noise, the proposed learning-based method can well reconstruct the details flooded in strong noise.
|