A real-time fast defogging system to clear the vision of driver in foggy highway using minimum filter and gamma correction

Abstract

Fog is the most hindrance and unavoidable problem faced by drivers while driving. Due to foggy condition and poor visibility, especially in early morning and late-night, drivers are unable to see distant object on the road. As a result, possibility of road accident increases. In this article, a fast-real-time vision-based defogging system is proposed to clear the vision of highway during driving in the foggy environmental condition. The proposed system can remove the haziness of the driver’s vision and can present a clear view of the road within a very short span of time. Processing of each frame is comprised of four steps: calculation of atmospheric light using minimum filter, transmission map, scene radiance and finally gamma correction is applied for removing the haziness with perfect contrast adjustment. In order to reduce time complexity, instead of estimating atmospheric light for each frame, it is calculated at an interval of 5000 frames. Many real-time heuristic tests have been conducted during day as well at night on the highway and test analysis reveals that, after defogging, the distance of visibility increases by more than 65% during heavy fog. Besides, there is a massive increase in visibility during low foggy condition also.