【摘要】 在交通数据实时采集的诸多技术中,基于视频图像处理的方法不但能够克服人工采集方法精度低、可靠性差等缺陷,还能够解决传统的感应线圈检测、雷达检测和微波检测等方法无法检测非机动车和行人的问题.本文针对该问题设计和提出在混合交通情况下的视频交通数据采集系统,通过将视频图像处理技术和神经网络相结合,采集交通数据并识别和区分不同种类的交通对象.经过测试实验,效果比较理想.更多还原

【Abstract】 Among many techniques for real-time traffic data collection,vision-based image processing method has many advantages.It could not only overcome the limitations of precision and robustness in manual collection,but also solve the problem of automatic identification and classification of detected objects which cannot be realized in inductive loop,sonar and microwave sensors.Due to the lack of traffic data collecting system for pedestrians and cyclists,a vision-based system for mixed traffic data which combines image processing technique with neural network was proposed in this paper.The system is designed to collect real-time traffic data with high fidelity,as well as automatically distinguish and classify different objects.And most of the designated aims have been achieved in later test experiments.更多还原