基于结构化视频数据的交叉口评估及问题自动化诊断

Intersection Evaluation and Automatic Problem Diagnosis Based on Structured Video Data

摘要:

为提高视频数据用于交叉口运行评估的处理效率并实现交叉口问题实时诊断，提出了一种使用结构化视频数据、以车头时距为主要分析对象的方法。首先利用流量曲线可复现性快速查找高峰时段；然后仅基于车头时距曲线实现视频数据与信控方案的同步比对，并提出了从信号控制方案和车道两方面进行交叉口效率综合评估方法；最后以动态时间规整和灰色聚类方法，由实际数据识别出车道车头时距曲线的5种类型，并提出了相应的问题及优化方案，可用于低效车道及相位的问题诊断。

Abstract:

To improve the processing efficiency of video data used in intersection operation evaluation and to realize real-time diagnosis of intersection problems, a structured video-data based method is proposed, with headway as the primary analysis object. First, an algorithm based on the reproducibility of the flow curve is presented to search for the peak period. Then, the headway curve is used to realize the synchronous comparison of video data and signal control schemes. A comprehensive evaluation method of intersection efficiency is proposed from two aspects of signal control scheme and lane. Finally, the dynamic time warping and the grey clustering method are utilized to identify five types of lane headway curves based on actual data, with the corresponding problems and optimization schemes listed. The method provides a solution for problem diagnosing of inefficient lane and phase.