摘要
设计了一种基于图像理解的智能车牌识别系统,利用水平和垂直投影进行车牌区域定位.通过基于垂直投影和模板匹配的字符分割方法有效地解决了背景变化、车牌尺寸变化、字迹模糊等问题,在对120幅472×332大小的车辆图像采用基于模板匹配的ORC识别方法,在Hausdorff距离分类器上进行了相关匹配测试实验.识别率达96.7％,识别时间稳定在0.3～1.2 s之间.
This paper designs an intelligent license plate recognition system based on image understanding by using horizontal and vertical projection to license plate localization. The character segmentation method based on vertical projection and template matching can effectively solve the plate＇s background changes, size changes and the character blur. In 120 images with the size of 472 ×332 are recognized using ORC recognition method based on template matching method, and correlation matching test experiment is carried out with the Hausdorff distance classifier. The recognition rate reaches 96.7% with a stable recognition time between 0.3 to 1. 2 s.
出处
《电子科技》
2014年第1期147-149,165,共4页
Electronic Science and Technology
关键词
图像理解
LPR
特征提取
模板匹配
image perception
license plate location
feature extraction
character match
