Abstract:

An improved algorithm based on dark channel prior was proposed which compenstaes the transmission estimated via dark channel prior, redefines atmospheric light via analyzing the fog images fail to recover by original method based on physical model. Experiments prove our algorithm works better than original method, especially on sunlight lacking or sky including images. Meanwhile, mean square error(MSE) was used in our method to verify the images we recovered are close to true images based on fog database.

Key words: haze removal, dark channel, mean squared error(MSE)