陈真

       随着我国经济的迅猛发展，城市快速路也进入了快速发展的阶段。我国的城市快速路虽然起步较晚，但发展较快。然而，交通拥挤和各类交通事件的频繁发生，不仅使城市快速路效率较低，而且给你们的生命、财产以及社会和经济均造成了不同程度的损失，有的甚至导致了严重后果。

       如果有效地利用现代化的交通监控手段来确保城市快速路的畅通和高效和如何研制和运用先进的自动检测技术快速准确地检测、判断并及时排除因交通事件造成的交通拥挤和堵塞、尽可能将事件的影响和损失降低到最低限度是当今城市快速路管理所面临的重大课题。

       本文的研究是基于传统的时间序列模型与人工智能模型建立上海城市快速路交通事件检测模型，主要内容包括：

       （1）论文对上海城市快速路的交通事件主要特征及形成机理进行了详细的研究分析。分析了上海城市快速路的交通事件的时空特征，形成消散过程，主要类型等。

       （2）论文基于ARIMA与神经网络算法建立交通事件检测算法研究。通过上海城市快速路空间几何与线圈数据，对上海中环路的一段路程建立交通事件检测算法，算法为ARIMA模型与BP神经网络的混合模型，并根据结果进行分析，证明混合模型较一般模型要好。

       （3）论文建立了可视化模型。结合已建立的交通事件检测算法，本文以Visual Studio.Net 2010为开发平台，对GIS进行了二次开发，构建了 交通事件检测可视化系统界面及相应GIS功能部分。

关键词：线圈数据，交通事件检测，可视化

       With the booming economy, the urban expressway in our country enters the phase of rapid development. Although the urban expressway in our country started late, it developed rapidly. However, with traffic congestion and traffic incidents occurring frequently, not only make the city expressways’ efficiency decrease, but also cause varying degrees of damage to people's life and property, society and economy, and some even have serious consequences.  

       It is an important topic in urban expressways management that how to effectively use modern means of traffic monitoring to ensure the city expressway smooth and efficient and how to develop and use the advanced automatic detection technology to quickly and accurately detect and reduce traffic incidents caused by traffic congestion to a minimum impact as much as possible.  

The research of this paper is based on the traditional time series model and artificial intelligence model to establish the traffic incident detection model of Shanghai urban expressways:  

       1, In this paper, the main characteristics and formation mechanism of traffic accidents in Shanghai urban expressway are analyzed in detail

       2, Based on ARIMA and neural network algorithm, the traffic incident detection algorithm is developed. Based on the spatial geometry data and the coil data of Shanghai urban expressway, a traffic incident detection algorithm is established for a section of the middle ring road in Shanghai, and the results are analyzed after.  

       3, The visualization model is established in this paper. Combined with the establishment of the traffic incident detection algorithm, based on the ArcGIS platform to achieve visual display, adding more direct image display to algorithm results.

Key Words: coil data, traffic incident detection, visualization.