Automatic white balance algorithm through the average equalization and threshold

Automatic White Balance (AWB) is one of the important functions for digital cameras. The purpose is removing the color cast caused by the un-canonical illuminants where the image captured at. Images with color cast always lose the color contrast and variations, especially those with heavy color cast. For traditional white balance algorithms, the lost color contrast and variations would probably bring the incorrect color temperature predictions and the lack of contrast in the output images. In this paper, a new automatic white balance algorithm through the average equalization and threshold (AWBAAET) is proposed. The proposed approach keeps the assumption of gray world method (GW) and the wide color range of color histogram stretching algorithm (CHS) at the same time. In the simulation results, the proposed algorithm can meet the theory of GW, with similar averages on the R, G, and B channels. Also, it can enhance the color contrast and variances as CHS algorithm efficiently, without over stretching the color range and changing the character of the original image.