Improving the performance of vehicle detection system in bad weathers

Abstract

Vehicle detection has been applied in many fields, such as intelligent transportation, video surveillance, driving assistance system and so on. In the case of fine weather, the state-of-the-art vehicle-detection systems may achieve good performance. However, the performance has a substantial decline in bad weathers, such as fog, night and so on. Therefore, improving the performance of vehicle-detection systems in different weather conditions becomes an important issue in vehicle-detection system. In the fog or night, the quality of the images is reduced. In this paper, we propose some algorithms of image defogging and color enhancement in order to improve the performance of vehicle detection. The result of vehicle detection get much better after image processing in bad weathers.