A Method of Vehicle License Plate De-noising and Location in Low Light Level

Abstract

License plate recognition system (LPRS) is one of the most important parts of the intelligent transportation system (ITS), and the license plate location is the key part of the LPRS, it affects the latter phase of the character segmentation and recognition directly. According to the characteristic of the vehicle license plate which contains the slat and pepper noise in low light level, we take wavelet transform to vehicle license plate in low light level, then reconstruct the high frequency coefficient based on characteristic of noise, and use the soft-threshold method to do threshold processing of the reconstructed image, and obtain the effect of image de-noising. Finally, we use the method of edge detection and projection to locate the vehicle license plate area for de-noising vehicle license plate image. The simulation results show that the wavelet soft-threshold method can de-noise the low light level image greatly and the method of edge detection with projection can locate the vehicle license plate area fast and accurately.