一种基于K-Means分类的状态机车辆检测算法

地磁车辆检测器是一种基于车辆对地球磁场的扰动效应实现的交通信息采集设备。目前基于地磁传感器的车辆检测算法主要有状态机检测算法及自适应阑值算法,但是基线漂移、阅值选取等因素对算法的检测精度有很大影响。结合车辆地磁响应信号的具体特征,提出一种基于K—Means分类的状态机车辆检测算法,将车辆地磁响应信号先进行K—Means分类,解决了现有算法阈值选取困难的问题,将分类后的信号输入状态机判别,解决了慢速车和长型车易被误判的问题。实验结果表明,该算法具有很高的检测准确率,且鲁棒性好。

A State Machine Vehicle Detection Algorithm based on Clustering Algorithm of K-Means

Geomagnetism vehicle detector is a traffic information collecting device,which is based on the perturbing effect caused by the vehicle passing the earth magnetic field.At present,there are several algorithms based on this geomagnetism vehicle detector,such as State Machine Algorithm and Adaptive Threshold Detection Algorithm,but baseline drifting and threshold choosing make great influence to the detecting accuracy.According to the special characteristic of vehicle geomagnetic signal （VGM）,a State Machine Algorithm based on the online Clustering Algorithm of K-Means is presented in this paper.Classifying the VGM by using K-means algorithm,the difficulty in choosing threshold is solved,then using the state machine,the problem that slow cars or trailers would be misjudged is solved also.