摘要
针对车牌检测中关键的3个环节分别进行了改进和优化。利用数学形态学结合Canny算子实现车牌定位。在传统扫描字符算法中加入边界限定,提高字符分割的准确度。在识别环节,采用计算效率较高的KNN算法进行字符的识别。最终,在OpenCV平台上实现车牌识别。
The three links are improved and optimized respectively in the fieldof license plate detection.Vehicle license plate location is realized by using mathematical morphology and Canny operator.Adding boundary restriction in traditional scanning character algorithm also improves accuracy ofcharacter segmentation.In the stage of character recognition,KNN is adopted due to high computational efficiency.All these processes are implemented on the OpenCV platform.
作者
毕波
邵永谦
孙冬军
贾思超
BI Bo;SHAO Yong-qian;SUN Dong-jun;JIA Si-chao(China Shanghai Earthquake Agency,Shanghai 200062,China)
出处
《电子设计工程》
2019年第1期37-41,共5页
Electronic Design Engineering
基金
国家地震局测震台网青年骨干培养专项(20150409)
上海市科委项目(14231202602)
关键词
OPENCV
车牌定位
边缘检测
字符分割
图像识别
K阶最邻近
OpenCV
vehicle license plate location
edge detection
character segmentation
image recognition
K-Nearest Neighbor
