Image haze removal: Status, challenges and prospects

Abstract

Image haze removal has become an important researching direction in computer vision because of the development of computer vision system and its increasing demand. Images acquired in bad weather, such as haze and fog, are seriously degraded by the scatting of the atmosphere, which makes the image color gray, reduces the contrast and make the object features difficult to identify. The bad weather not only lead to the variation of the visual effect of the image, but also to the disadvantage of the post processing to the image, as well as the inconvenience of all kinds of instruments which rely on optical imaging, such as satellite remote sensing system, aerial photo system, outdoor monitoring system and object identification system. That's the reason why they need enhancement and restoration for the improvement of the visual effects and convenience of post processing. This paper sums up the status of image haze removal methods. After that, quantitative and qualitative evaluations of these techniques are presented and the challenges of the existing haze removal method are summarized. Finally, we proposed some expectations to the future research in the field of image haze removal.