Improvement of the Vehicle License Plate Recognition System in the Environment of Rain and Fog

License Plate Recognition (LPR) is one of the key technologies of the intelligence of communication management. A certain amount of difficulties to license plate recognition are caused by the environment of rain and fog. License plate recognition system for this kind of environment is studied in this paper, based on the theory of digital image processing, computer vision and pattern recognition technology. In order to improve the ability to identify the license plate, Gamma correction algorithm and the denoising algorithm of color image are added to license plate locating method based on color. In the pre-processing of the segmentation method based on connected area detection, a kind of double color space binarization method is proposed by the article. Finally a kind of adaptive fusion algorithm based on BP, RBF, GRNN neural network is proposed to finish the recognition for license plate character. Experiment shows that the method adopted by this system applied in the bad environment of rain and fog achieves good recognition effect.