Fast Defogging and Restoration Assessment Approach to Road Scene Images

Abstract

Single image defogging has received increased attention recently. However, most existing methods are generic restoration methods which may cause partial over-enhancing to the image at a short distance. This certainly makes these defogging methods may be not suitable for road images. In this paper, we propose a new image defogging algorithm that divides the original foggy image into three regions and different enhancement processes are performed for the different region. As the results, the far-away region image that is most concerned by a car driver can be moderately enhanced, and over-enhancement of the bottom region image can be also limited effectively. To effectively measure the defogging effects of various algorithms, the assessment methods based on image feature and visibility distance are proposed for road scene image. The comparative study and quantitative evaluation show that the proposed method can obtain very good defogging image with relatively fast speed.