Algorithm And Implementation, Of The Motor Vehicle License Plate Recognition System

While monitoring system is an important part of Intelligent Transportation System (ITS), recognizing vehicle license plate is the key technology of monitoring system. It's applied in various field, including: street traffic monitoring, traffic violation recording, residential area entrance guarding, tollbooth, etc.Vehicle license plate recognition, LPR, is an emerging application field of information technology. In an LPR system, digital video camera is used to capture image of traffic scene and image recognition technology is used for recognition of the target license plate number.A practical LPR system should be adaptive for the various scenes, considering different monitoring places, different weather conditions, and different lighting conditions and be tolerant of image quality. On the other hand, a practical LPR system should promise of reliability, which refers to control of recognition ratio. For cost control, it should be deliberately balance between hardware condition and algorithm efficiency in design stage, and power consumption is another important aspect.A complete LPR system is composed of hardware and softer ware. in this paper , hardware design is briefly described, softer ware design is emphasized in two aspects: algorithm design and implementation.The key of this system is the image recognition algorithm, which would be separated in three parts:1. License plate location. Locate and crop out the license plate region in a captured Image.2. Character segmentation. Since a license plate contains several characters, segment them out is a preparation for further recognition.3. Character recognition. The final key step gives out result that what the separated characters are by Support Vector Machine (SVM) .The design of software is implemented by C/C++ programming language, the target platform is embedded system: 1. Using object oriented programming method, manage the complexity, and obtain a clear architecture.2. Reference the design patterns, manage the mutable parts and consider the immigration requirement and refactor ability.3. Use unit testing technology to guarantee the quality of the code.This thesis focuses on Nation Standard: License plate of motor vehicle of the People's Republic of China (GA36-2007).The discussing covers GA36-1992 also, considering the fact that most license plates in China still adhere to it.