The Research On Vehicle License Plate Recognition System

Vehicles' License Plate Recognition (LPR) system is an important application of computer vision and pattern recognition in the intelligent transportation management. The rough set can analyze and deal with vagueness, disaccord and uncertainty information, discover the connotative rule. Neural network is widely used in pattern recognition for the better generalization ability. So the research on LPR system based on the combination of rough set and neural network is very meaningful and practical. The following work is finished in this paper.1. A novel segmentation method based on texture and color is presented, which has fewer limits against the complex background.2. The research on the plate knowledge expression & acquisition method based on rough set is carried out. It includes construct decision table, attribute discretization, attribute reduction and rule abstract. A discretization algorithm based on the attribute importance is presented. The experiment results show that the algorithm is effective. Knowledge acquisition based on the Rosetta and Matlab 6.1 software is accomplished.3. Two effective character recognition methods are presented. One is based on the improved rule matching algorithm; another is based on the combination of rough set and (RBF) neural network.4. Implement the above algorithms in Visual C++ 6.0 based on the theory research and develop the software of Vehicles' License Plate Recognition System. The software has some functions include reading and memory, pretreatment, vehicles' license plate location, characters segmentation and recognition, etc. The software can solo complete every above function with multi-algorithms. It realizes automatic recognition to the vehicle license plates.