摘要
基于车牌识别数据可以得到较为准确的行程时间数据,但是由于识别系统自身原因、驾驶员路径选择行为、停车行为等因素的影响,行程时间数据中存在不少的异常值,剔除异常值才能将所得数据应用于实际研究和服务.在详细分析了行程时间异常值产生原因的基础上,提出了基于异常值数据表现以及行程时间分布特征的异常值剔除方法,最后以北京市的车牌识别系统数据为例,验证了方法的适用性.
The accurate travel time can be derived from data of license plate recognition system.Outliers can be expected due to the system′s identification error,drivers' route choices and stopping activities.It is essential to filter the outliers before travel time data are used in research and services.After analyzing the causes of outliers,a filtering method was proposed based on the behaviors of outliers and distribution of travel time.To verify the application of the new method,the license plate recognition data of Beijing were used.
出处
《武汉理工大学学报(交通科学与工程版)》
2012年第1期116-119,共4页
Journal of Wuhan University of Technology(Transportation Science & Engineering)
基金
智能交通技术交通行业重点实验室开放基金
关键词
车牌识别数据
行程时间
异常值处理
标准差
绝对偏差
data of license plate recognition system
travel time
outlier filtering
standard deviation
absolute deviation