Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (10): 210-213.

• 工程与应用 • Previous Articles     Next Articles

An Online Vehicle Detection Algorithm Based on PLA Time Series Representation

Kov Chai   

  • Received:2006-04-20 Revised:1900-01-01 Online:2007-04-01 Published:2007-04-01
  • Contact: Kov Chai

基于PLA时间序列重现的在线车辆检测算法

柴可夫 霍宏 方涛   

  1. 上海交通大学电子信息与电气工程学院图像处理与模式识别研究所 上海交通大学自动化系模式识别专业 上海交通大学自动化系模式识别专业
  • 通讯作者: 柴可夫

Abstract: A new online vehicle detection algorithm designed for magnetic sensor node is presented. The algorithm smoothes magnetic time series using a variant length sliding-window averaging filter, then a PLA representation scheme is employed to extract the feature, which in turn is passed to a threshold detection module. Offline simulations demonstrate that this approach can greatly reduce the false alarm rate mainly caused by slow long vehicles.

Key words: vehicle detection, time series, piecewise linear approximation, sliding-window filtering

摘要: 文章提出了一种新的用于磁敏传感器的车辆检测算法。算法首先将磁敏数据时间序列经过变长滑动窗口滤波预处理,由PLA抽取平滑后的时间序列特征,用于车辆检测,从而获得相关的交通信息。仿真实验表明,算法有效地减少了慢速行驶的大型车辆对检测结果的影响,保持了较高的准确率。

关键词: 车辆检测, 时间序列, 分段线性拟合, 滑动窗口滤波