视频图像中的实时车牌识别

车牌识别在智能交通系统中起着重要作用。本文提出一种对视频图像进行实时操作的车牌识别方法。利用边缘特征，使用快速的边缘点连接算法生成连通区域，进行车牌的检测定位，结合字符小波特征实现准确的识别。算法平均消耗时间在40ms以内，定位准确率达99．5％，完全识别率在91％以上。实际采集数据实验结果证明，本方法速度快、准确率高、实用性好。

Real-time Vehicle License Plate Recognition in Videos

Automatic License Plate Recognition (LPR) plays an important role in numerous applications in Intelligent TransportSystem (ITS). This paper presents a novel algorithm for real-time license plate recognition in video images. The algorithmdetects LP region by the connective components which are generated by a fast neighboring edge points connecting method andrealizes accurate recognition with character wavelet features. The whole process takes less than 40 minutes, with the successfulrate of locating LP reaching 99. 5% and the total successful recognition rate above 91%. The experiments have proved that ouralgorithm has high speed, accuracy and wide availability.