A Joint Multi-Gradient Algorithm for Demosaicing Bayer Images

In image sensors such as CMOS and CCD, de-mosaicing is usually required to replace Bayer-format images with RGB-format images. Poor de-mosaicing algorithms can produce zipper effects and pseudo-colors, etc. In view of this situation and the hardware requirements of image devices, the original de-mosaicing technology gradually fails to meet the practical needs. On the other hand, although the deep learning-based de-mosaicing technology is more effective, the computational volume is too large, occupies too much hardware resources, and is difficult to be realized on the hardware platform. Therefore, this paper proposes a de-mosaicing algorithm based on multi-gradient union, which improves the traditional gradient-based first-order differentiation method, designs a new gradient operator, and ultimately judges the image edges by means of four gradient direction decisions to interpolate along the image edges. Experiments show that the algorithm in this paper has better recovery in image edges as well as texture complex regions with higher PSNR and SSIM values and better subjective visual perception compared to the traditional gradient algorithms such as BI, Cok, Hibbard, Laroche, Hamiton, while the algorithm involves only the add-subtract and shift operations, which is suitable to be implemented on the hardware platform.