Real time image haze removal using bilateral filter

A real time image dehazing method using bilateral filter is proposed to eliminate ill effects of haze or smoke for outdoor machine vision system. Airlight vector was estimated based on statistics of dark channel prior at first, and then the atmospheric veil was inferred by the estimation of fast bilateral filter. The ideal scene radiance could be recovered by solve imaging physical equation of hazing scene. Experiment results indicate that this method performs well both for color and gray images with haze. Its processing speed is faster than 20 frame/s dealing with images of 320×240 pixels and could satisfies the demands of real time image applications.