【摘要】 针对现有交通流数据采集设备存在造价高、采集精度不够及难于后期分析等实际问题,提出了一种利用车辆音频时域信号特征进行交通量检测的方法。该方法根据车辆噪音信号的波形特点,通过提取车辆噪音时域特征参数短时能量和过零率,进行端点、特征跳变点检测及车型辨别、分类,统计出交通量数据。试验结果表明,结合音频信号时域能量分布和过零率能够实现交通量信息的有效检测,同时为保证音频信息采集的有效性,数据采集设备应安装在车辆加速行驶路段或凸形竖曲线顶部附近。该检测方法具有检测方便、设备投入少的特点,统计精度能够满足工程实际需要,适用于一般公路交通量的调查。更多还原

【Abstract】 A novel survey method of traffic volume was put forward and experimented using the feature extraction of vehicle’s time-domain noise,considering the existing problems like costly equipments for traffic flow data collection,lower efficiency in operation and obstacle of data analysis,etc.Traffic volume survey is one of the main works in traffic investigation and implementation.According to vehicle’ noise waveform characteristics,it is shown that,the changed point can be obtained by analyzing the volume and zero cross rate in time-domain,and then the number and type of vehicles can be got.The research result states that this method is simple and economical.The statistic precision can correspond with need of the engineering.It can be used to investigate traffic volume in common highways.更多还原