基于分形理论与GIS的交通事故黑点预判方法研究

发布时间：2019-03-01 08:02

【摘要】：交通事故黑点判别是交通安全研究的重要内容之一。目前的事故黑点判别方法主要应用于发生事故后的路段的鉴别和研究，这类事后型控制的方法是以发生一定的交通事故数量作为研究基础，已然造成了生命财产的重大损失。因此，，研究切实有效的事故黑点预判方法，为设计、研究及管理部门在道路设计、建设阶段分析道路安全性，和及时发现潜在危险路段提供技术支持，对促进我国道路交通安全管理水平提升有重大的实际意义。本文以高速公路为例，在交通事故黑点特征分析、事故黑点影响因素及成因分析基础上，提出了分形理论应用于交通事故黑点预判及区域宏观评价的新方法。主要进行了下述研究：提出了本文对事故黑点的新定义；从定性分析与案例结合的角度论述了事故黑点的自相似性；分析了分形理论应用于交通事故黑点鉴别的理论支撑与可行性；分析了交通事故黑点影响因素；通过主成分分析法提炼出致因关键；将关键致因融入分形关联维数的计算模型；利用案例分析予以验证、说明该分形模型的有效性和可行性；将事故黑点分形维数预判模型置入Arcgis平台中，实现自动计算区域内所有路段的关联维数值；应用K均值聚类法对维数计算结果进行分类和事故黑点危险程度划分；最后利用帕累托维数对整个研究区域的事故黑点分布情况进行宏观评价。
[Abstract]:Traffic accident black spot discrimination is one of the important contents of traffic safety research. At present, the identification method of accident black spots is mainly used in the identification and research of road sections after accidents. This kind of post-incident control method is based on the number of traffic accidents, which has caused a great loss of life and property. Therefore, we should study practical and effective methods for predicting accident black spots, and provide technical support for the design, research and management departments to analyze road safety during the road design and construction stages, and to identify potentially dangerous sections in a timely manner. It is of great practical significance to promote the level of road traffic safety management in China. Taking expressway as an example, this paper puts forward a new method of fractal theory applied to traffic accident black-spot prediction and regional macro-evaluation on the basis of analysis of traffic accident black-spot characteristics, influencing factors and cause of formation of traffic accident black-spot. The main contents are as follows: the new definition of accident black spot is put forward, and the self-similarity of accident black spot is discussed from the point of view of qualitative analysis and case analysis. This paper analyzes the theoretical support and feasibility of the application of fractal theory to traffic accident black spot identification, analyzes the influencing factors of traffic accident black spot, extracts the causative key by principal component analysis, integrates the key causative factor into the calculation model of fractal correlation dimension, and puts the key cause into the calculation model of fractal correlation dimension. The validity and feasibility of the fractal model are proved by case analysis, and the fractal dimension prediction model of accident black point is put into Arcgis platform to automatically calculate the correlation dimension value of all sections in the region. The K-means clustering method is used to classify the calculated results of dimension and the hazard degree of accident black spots. Finally, the distribution of accident black spots in the whole study area is evaluated macroscopically by using the Pareto dimension.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208;U491.31

【相似文献】