Rapid development of informatization and intelligentization of the traffic management system requires higher speed and better performance in recognition of the license plate. For this reason, the original algorithm is improved in this paper with voice announcement function added, and the license plate recognition system with voice announcement is designed based on MATLAB simulation. Through the improved Otsu pseudo binarization algorithm, the processing speed of the system is increased. The location method of the license plate based on line-scan gray level jump analysis integrates characteristics of textural features analysis method and edge detection analysis method, which can achieve a high speed and high precision. The experimental result shows that, 300 random images of vehicles are chosen for test and accuracy rate of recognition is higher than 98%. The system realizes accurate location correction, segmentation and recognition of the license plate under natural light, and the recognition result can be announced by means of voice as well. So the system has excellent foresight and man-machine interaction‥
展开▼
