考虑系统聚类的公路交通事故多发点鉴别研究

Research on identification of black spots for highway based on system clustering method

为改善高原地区公路交通环境,提高公路事故多发点鉴别的准确性,结合云南省2010—2015年道路交通事故数据统计资料,采用统计分析方法分析高原地区公路交通事故形态及时空分布特征;以百万车公里事故率和当量事故次数为指标,运用系统聚类法鉴别高原地区公路交通事故;并以昆石高速公路为实例进行分析,鉴别出11个事故多发点段,且与实际情况基本相符。结果表明:在时间分布上,高原地区公路交通事故多发在冬春季,一天中下午及傍晚时分事故率较高;在空间分布上,高原地区县乡道路事故较为高发;从事故形态看,发生正面碰撞、刮撞行人的事故数量约占事故总数的57%,为高原地区的主要事故主要形态。

In order to improve the road traffic environment in the plateau area,and increase the identification accuracy of black spots,based on statistical data on road traffic accidents in Yunnan province from 2010 to 2015,the statistical analysis method was used to analyze the characteristics of highway traffic accidents and temporal and spatial distribution in the plateau area. Regarding the number of accidents and the number of equivalent accidents in millions of kilometers as indicators,the systematic clustering method was used to identify the type of highway traffic accidents in the plateau area. Then,it was analyzed with Kunshi expressway as an example,a total of 11 accident-prone points were identified,which match the actual situation effectively. The results show that,in terms of time distribution,highway traffic accidents in the plateau area happen frequently in winter and spring,that the accident rate is highest in the afternoon and the evening,that in terms of space,road accidents in the counties and townships in the plateau region occur more frequently,that from the point of view of accidents,highway accidents in the plateau area aremostly frontal collisions and scratching pedestrians,which occupied 57% of the total number of accidents.