Sharpness-aware real-time haze removal algorithm for advanced driver assistance systems

Abstract

This paper proposes a new haze removal technique for automotive applications. Unlike to the previous studies, which focus only on the qualitative elements of the recovered images, the proposed haze removal algorithm uses the weighted filter to sharpen the edges of lanes so that the clarity of lanes is remarkably enhanced rather than the general quality of images. In addition, the entire processing time is reduced by eliminating unnecessary steps that are not related to the sharpness quality. As a result, the proposed haze removal algorithm provides the real-time image processing for advanced driver assistance systems, while significantly improving lane sharpness.