摘要
车牌汉字识别效率一直是制约车牌自动识别系统推广应用的瓶颈。采用多级分类方法,充分利用车牌汉字图像的空间结构特征(像素空间分布投影,字符的笔段类型、数量及拓扑关系等)可实现车牌汉字的快速辨识。实践表明,该方法能快速从车牌汉字图像中辨识出汉字字符,识别准确度可达98%。
The application of license plate recognition system is restricted by the efficiency of license plate character recognition. A recognition method of license plate character is presented, which extracts the space structure feature of Chinese character and then recognizesthem using multilevel classifier. The experiments and their relevant results show that the recognition accuracy reaches 98%.
出处
《计算机测量与控制》
CSCD
2005年第6期592-594,共3页
Computer Measurement &Control
基金
国防科技预研项目(30020490523)
中国工程物理研究院电子工程研究所创新基金(S20041111)。
