摘要
乘客上下车自动人数统计系统是一个亟待解决的难题,而传统的统计方法存在诸多不合理因素。本文介绍了使用红外传感器的自动乘客计数方法的原理和结构,同时引入动态时间规整算法对信号进行识别,但传统的DTW(Dynamic Time Warping)算法存在缓存空间大、对端点检测依赖性较强等缺点。作者对其进行改善并提出了具有良好的端点检测方法。实验统计结果表明,其准确率达到96%。
The automatic passenger counting system is an unsolvable urgent project now. There are many unsuitable factors in traditional solutions. This article presents the principles and system structure of the automatic passenger counting using infrared sensors. And dynamic time warping (DTW) is introduced to conduct the pattern recognition. But traditional DTW algorithm has many shortcomings, such as huge buffer-space and strong dependence on point-detection. The writers improved it and brought forward a new point-detected method. Experimental results show that the accuracy reaches 96%.
出处
《微型电脑应用》
2008年第4期16-19,2,共4页
Microcomputer Applications
关键词
自动人数统计
红外传感器
端点检测
动态时间规整算法
Automatic passenger counting
Infrared sensor
Point-detection
Dynamic time warping (DTW)
