摘要
以改进智能交通控制中的车牌定位性能为目的,提出一种基于改进的边缘特性提取和改进BP神经网络的车牌定位新方法。该方法通过改进的图像的边缘特征提取方法,将边缘特征送入改进的BP神经网络进行训练学习,实现车牌的粗定位;再利用车牌本身的几何特征实现车牌的准确定位。实验结果表明该方法定位精度高,并提高了信息的实时处理性。
The article aims at improving license plate positioning performance in smart traffic control.A new method of vehicle license plate positioning is presented based on the improved edge characteristics extraction and the improved BP neural network.In this method,the edge characteristic is inputted into an improved BP neural network to train through the improved image edge characteristics extraction method to achieve rough positioning of a vehicle license plate.Then the geometric characteristics of the license plate are taken into account to get accurate positioning.It is shown by the experimental result that this method has high positioning accuracy,and the real time information processing performance is improved as well.
出处
《计算机应用与软件》
CSCD
2011年第12期86-87,131,共3页
Computer Applications and Software
基金
国家高技术研究发展计划(2008AA11A134)
四川省教育厅基础应用研究课题(2009ZX002)
关键词
车牌定位
BP神经网络
动量项
模式识别
License plate positioning BP neural network Momentum Pattern recognition