Research on Chinese Vehicle License Plate Detection and Recognition Technology Based on BP Neural Network

License plate recognition has gotten significant consideration over the previous decade as applications under the booming development of motor vehicles. However, the previous studies mainly identified the information by template matching method and often face the problems of low accuracy and speed in identification. In this paper, the License Plate Recognition Method (LPRM) is applied to acquire accurate license plates in various environments. The first stage is to remove the areas that do not correspond to the characteristics of the license plate by edge detection. Then, the position of the license plate is captured by erosion expansion technology while the feature image is obtained by the width-based cutting method. In the second stage, the BP neural network is adopted to identify license plates and arrange characters in their original order. Finally, the identified complete license plate is transferred to the background control for management. An example is also given to validate the applicability of this two-phase recognition method. The results show the proposed method is capable of better accuracy and feasibility in realistic environment and therefore can assist managers to improve the efficiency of vehicle management in multiple scenarios.