Image Dehazing Algorithm And Its Application In License Plate Detection

Images of outdoor scenes are usually degraded by the foggy weather, which reduced the performance of intelligent transportation, surveillance, and target identification systems. Therefore, haze removal from a single image of a weather-degraded scene has important value in improving image quality and robustness of the visual system.This paper introduced the image haze removal algorithm based on atmospheric scattering model and dark channel prior law. Recovered haze-free image from a single foggy image on the VC++6.0software platform. A guided image filtering is used to improve the transmission rate diagram. Besides, In order to reduce the complexity of this algorithm, we used a fast algorithm for local minimum filters on rectangular when we approach the dark channel mapped to a color foggy image. Experiment results demonstrate the improved method has more natural dehazing result image. Compare the dark channel prior algorithm with the two other classic methods, the experiment results show that the dehazing algorithm this paper introduced is better than the other two methods in thick foggy weather, and it has advantages of more natural results and less information loss.Applying the haze removal algorithm to the license plate detection system, the goal of license plate detection in foggy weather is reached, accordingly, the robustness of the license plate detection system is improved.