Image restoration using prior information physics model
Le Xing, Lianhe Yang · 2011
Images captured in haze weather suffer from serious degradation of color and contrast due to incident light scattering and absorption. However, most of the existing methods which are based on depth information could not get satisfactory haze removal effect. In this paper, a simple but effective method is presented - image restoration using improved dark channel prior. The dark channel prior is based on a large number of statistical information. Most local patches in haze-free outdoor image contain some pixels which have very low intensities in at least one color channel. Using this prior with the haze image model, we can recover a high quality haze-free image. Image or video taken in haze day is restored using guided filtering finally. The simulation experiments based on Matlab demonstrate that the method is easy to use and able to improve quality of the images in haze efficiently. Meantime, the method can meet certain practical requirements.