基于云平台的交通视频检测数据优化研究

Data Optimization of Traffic Video Vehicle Detector Based on Cloud Platform

交通视频检测技术在智能交通领域应用广泛,成为当前交通信息采集的主要手段.但在恶劣天气条件下,视频车辆检测器采集的交通数据误差较大,难以准确反映路面实际状况.为解决此问题,本文提出基于不同气象能见度等级构建k型BP网络,对气象能见度低于10 km时的原始交通数据作预处理优化.分析了云计算在交通信息处理方面的优势,基于云计算平台实现了该模型的构建与推广.最后以成都绕城高速七里沟大桥定点观测得到的数据做样本进行实例分析,对比了该模型方法与传统处理方法的数据处理效果,得出了本文方法较传统方法先进的结论.

Traffic video detection technology is widely used in ITS and becomes the main form of transportation information collection. But in bad weather condition, the traffic data deviation is so large that it’s difficult to accurately reflect the traffic situation. To solve this problem,k -type BP neural network model is established based on different meteorological visibility level, then preprocess the data which collected in the meteorological visibility level below 10 km. It also analyzes the advantages of cloud computing in traffic information processing, realizes the construction and promotion of the model based on cloud computing platform. The superiority of the new model compared with the traditional method is validated by an example. Finally, an example is analyzed based on observatory date of Chengdu beltway Qiligou Bridge, compare the effectiveness of the proposed method and the traditional processing methods. The conclude shows that the proposed method is more advanced than the conventional methods.