Research of illumination compensation for driver color video image

In this paper, we propose a robust method to compensation illumination for driver video color image for enhancement image quality. Our algorithm is divided into four parts: image decomposition, illumination compensation, image composition, global adaptation. The original image is decomposed into high-frequency image and low-frequency image based on spatial image decomposition. The information of illumination contained in the low-frequency image absolutely. Low-frequency image is compensated used Retinex method which can brighten the darker area. We can get the new image after global adaptation of Synthesis image which is composed compensated low-frequency image and high-frequency image. Results show, the algorithm has better results compared with the traditional algorithm.