Automatic White Balance Based on Gray World Method and Retinex

Abstract

In different illumination, the same scene has different images. These images have different color. While adjust the image color, white balance is a very critical step. In the color adjustment, effective and automatic white balance algorithm can realize the intelligent adjustment of image color and save time. There are two kinds balance methods are Gray world method and Retinex theory. Two methods are commonly used. These have different treatment goals and have different characteristics. And their applications are relate to image attribute. Therefore, the two methods are combined and improve the white balance effect. In this paper, image was processed with gray world method firstly. Using the color constancy of Retinex, we maintain the green channel. We use nonlinear methods to adjust the pixel intensity. Through different parameters, we change the R, G, B channel value and realize the non linear adjustment of image intensity. The experimental results show that the proposed automatic white balance method can effectively avoid image distortion and keep the white balance is better.