A new method of vehicle license plate location under complex scenes

Vehicle license plate location is an important part of a license plate recognition system. This paper proposed a new location method for plate under complex scenes based on horizontal gradient and morphological operation. Firstly, horizontal gradient image was obtained by horizontal difference on the grayscale image. Secondly, an adaptive threshold proposed in this paper was used to segment the gradient image. Then, the candidate vehicle license plates were detected by morphological and region operations. Finally, the result of the license plate location was decided by candidate plates searching according to the features of the plate region. The experiments on 191 vehicle images that were taken under different illumination, weather and distance conditions show that the efficiency of the proposed algorithm is 96.34%.