Color polarization imaging demosaicing based on Stokes vector information complementation and fusion

The division of focal plane polarimeter (DoFP) imaging sacrifices imaging resolution in order to achieve real-time polarization imaging. Especially for color DoFP imaging, the sampling rate is twice as low as monochrome polarized DoFP imaging. The existing demosaicing methods often fail to integrate color and polarization information, resulting in significant color differences and artifacts in the generated high-resolution images. We propose what we believe to be a novel Stokes Complementary Network (SCN) using Stokes vector mutual guidance, which effectively integrates color and polarization information to reduce color and polarization artifacts. We also believe we have a novel Polarization Fusion Complements Network (PFCN) employing DoLP and AoP that is proposed, which uses high-frequency information in polarization fusion features to aid in restoring more details of images. Experimental results demonstrate that the proposed method can sufficiently suppress color and polarization artifacts. Compared with state-of-the-art methods, the proposed method achieves superior performance in both quantitative metrics and visualization results.