伴随车检测技术应用研究.pdf

第12卷 第3期 交通运输系统工程与信息 Vol郾12 No郾3 2012 年6 月 Journal of Transportation Systems Engineering and Information Technology June 2012 文章编号: 1009鄄6744 (2012)03鄄0036鄄05 伴随车检测技术应用研究 赵新勇,安摇 实* (哈尔滨工业大学 交通科学与工程学院,哈尔滨 150001) 摘要:摇 海量动态交通流中,经常出现结伴而行的车辆.特定区域内,当结伴车辆出现的 概率较大时,即可将其视为伴随车辆,这类车辆具有相互掩护和团伙作案的重大嫌疑. 及早检测和识别伴随车辆,能有效降低道路交通安全系统中的危险因素,对预防和减少 与道路有关的治安和刑事案件,也具有十分重要的意义.本文在车牌自动识别数据库基 础上,应用数据挖掘技术,提出伴随车辆检测和识别算法,并进行了实地测试.实验结果 表明:应用数据挖掘技术对伴随车辆进行分析检测,具有检测效率高、检测误差小、应用 范围广的特点,完全可以满足刑侦等部门对伴随嫌疑车辆进一步排查的需要. 关键词:摇 交通工程;车辆识别;数据挖掘;伴随车辆 中图分类号:摇 U268.6 文献标识码:摇 A Research on Accompanying Cars Recognition in Practical Application ZHAO Xin鄄yong,AN Shi (School of Transportation Science and Engineering,Harbin Institute of Technology,Harbin 150001,China) Abstract:摇 In massive dynamic trafficflows,it isvery commonto seethecarsmovinginaqueue. In some scenarios,these cars are regarded as accompanying cars and suspected of gang crime support each other when that condition occurs in a high rate. It isvery important to identify the accompanyingvehiclesasearly as possibleandtoreducepotentialrisksofroadtrafficsystemandtoreduceroad鄄relatedpublicsecuritycases and criminal cases. Based on the automatic license plate recognition database and data mining technology, this paper proposes a set of algorithms in identifying accompanying cars and a field test is conducted. The results demonstrate theperformanceof thealgorithmwitheffectiveness,low detectionerror,wideapplication and capability for further investigation. Key words:摇 traffic engineering;ve