A novel automatic white balance for image stitching on mobile devices

When several images are stitched into a panorama, the illumination difference across input images remains as a major problem. To maintain the color constancy between input images, automatic white balance (AWB) can play an important role in reducing the artifact. In this paper, we propose a new AWB algorithm to alleviate the undesirable color cast in the stitched result. The proposed algorithm estimates the global neutral color considering all the characteristics of input images to obtain the natural panorama. Experimental results show that the proposed algorithm successfully produces a stitched panorama with relieved color difference when compared with the results from the conventional AWB algorithms.