基于视频的车辆检测在交通噪声预测中的应用

目前交通噪声预测使用的FHWA模型在实际应用以人工方式来统计通过的车辆数和车型，人工成本高，统计值容易产生偏差。通过引入检测线上发生改变像素的连接程度，提出基于中介度量理论和车辆长度信息的车辆检测算法，利用先进的图像处理和模式识别等技术，自动的获得道路通过的车辆数和车型，车辆检测结果可与FHWA预测模型有效结合。从实验结果可以看出，算法达到了很好的效果，为FHWA模型提供了准确的车流量数据。

Application of Video-Based Vehicle Detection in Traffic Noise Prediction

Now the FHWA model used in traffic noise prediction get the number and type of vehicles by artificial way, the labor cost is high and the statistical value often not very accuracy. By introduced the connection level of changed pixels on the virtual detection line, a new vehicle detection algorithm proposed was based on the theory of Medium measurement logic and the information of the length of vehicle. Using the technologies of Image Processing and pattern recognition, the algorithm can get the number and type of vehicles automatic, the result can combine with FHWA model effectively. The experiment results proved the algorithm achieves good performance, provided accurate traffic volume data for the FHWA model.