基于贝叶斯网络的车辆身份识别系统的研究

The Research of Vehicle Identification System Based on Bayesian Networks

在高速公路收费站实现电子不停车收费(Electronic Toll Collection,简称ETC)是发展智能交通(Intelligent Transportation)的关键,而车辆身份判别的准确度又是实现电子不停车收费的决定性因素。为此,本文采用贝叶斯网络的推理算法,对采集到的多通道车辆信息进行融合,根据融合结果识别车辆身份。当某一通道数据信息缺失时,用EM算法修补出完整数据信息。仿真结果表明,该方法极大的提高了车辆身份识别的精确度,在智能交通领域有较强的实用性。

Electronic Toll Collection（ETC） in highway toll stations is the key to the development of Intelligent Transportation,and the accuracy of the vehicle identification is the decisive factor of the ETC. To this end,we fuse the acquisition of multichannel via Bayesian network inference algorithm,and know the vehicle identity by the integration of the vehicle information in this paper. When a lack of channel information,We adopt the EM algorithm to complete data. The simulation results show that the method greatly improves the vehicle identification accuracy,and it is more practical in the field of intelligent transportation.