基于车辆轨迹的高速公路异常事件自动检测算法

李斌; 马静; 徐学才; 马昌喜

doi:10.3963/j.jssn.1674-4861.2023.03.003

基于车辆轨迹的高速公路异常事件自动检测算法

doi: 10.3963/j.jssn.1674-4861.2023.03.003

李斌^1,,
马静¹,
徐学才^2, ,,
马昌喜³

1.
兰州朗青交通科技有限公司兰州 730030
2.
华中科技大学土木与水利工程学院武汉 430074
3.
兰州交通大学交通运输学院兰州 730070

基金项目:

国家自然科学基金项目 52062027

国家自然科学基金项目 72131008

甘肃省科技重大专项计划项目 22ZD6GA010

详细信息

作者简介:
李斌（1982—），本科，高级工程师. 研究方向：交通新基建、智能交通. E-mail: 54810022@qq.com

通讯作者:
徐学才（1979—），博士，副研究员. 研究方向：智能交通系统、机器学习.E-mail: xuecai_xu@hust.edu.cn

中图分类号: U491.5
计量
- 文章访问数: 838
- HTML全文浏览量: 311
- PDF下载量: 94
- 被引次数: 0
出版历程
- 收稿日期: 2021-10-20
- 网络出版日期: 2023-09-16

An Automatic Freeway Incident Detection Algorithm using Vehicle Trajectories

LI Bin^1
,,
MA Jing¹,
XU Xuecai^{2
, ,},
MA Changxi³

1.
Lanzhou LongKing Transportation Science & Technology Co. Ltd., Lanzhou 730030, China
2.
School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology, Wuhan 430074, China
3.
School of Traffic and Transportation Engineering, Lanzhou Jiaotong University, Lanzhou 730070, China

摘要

摘要: 高速公路异常事件自动检测是有效保障道路交通安全和运输效率的重要手段，由于监控视频数据量巨大，现有自动检测算法存在实时性、准确性低的问题。为此本文提出了基于轨迹分类的对比性悲观似然（comparative pessimistic likelihood estimation，CPLE）算法。构建了包含车辆检测、车辆跟踪和轨迹分类3种功能的异常事件自动检测模型框架，采用YOLO v3对车辆进行目标检测，获得4类不同车辆类型的相关信息，采用简单在线和实时跟踪算法对车辆进行多目标跟踪，获得不同场景的异常事件车辆轨迹；基于半监督学习，采用极大似然法对车辆轨迹分类进行改进，引入对比性悲观似然估计，围绕其对比和悲观原则进行参数设置和标定，进行异常事件轨迹分类和确认，提出基于车辆轨迹的异常事件自动检测算法。以甘肃省G312线公路智能化检测系统为测试对象，共收集1 300段视频，形成530条测试集轨迹和630条验证集轨迹，测试结果表明：通过对不同场景异常事件进行检测和预警，基于对比性悲观似然估计的轨迹分类算法性能准确率达到89.7%，比自学习和监督学习方法的准确率分别高出23.6%和41.3%，尽管对散落货物和超速事件的检测正确性稍低，平均为77.0%, 但突发性停车、拥堵和事故的检测平均正确率达98.2%，在严重影响交通的事件检测方面的平均正确率达到94%。本方法丰富了高速公路异常事件自动检测算法，可作为异常事件自动检测提供备选方法。
- 交通安全 /
- 高速公路 /
- 车辆轨迹 /
- YOLO v3 /
- SORT /
- 对比性悲观似然估计
Abstract: An automatic freeway incident detection method is important for maintaining a safe, efficient traffic operation. Due to the fact that a large number of surveillance videos may hinder the real-time and accurate response of current automatic incident detection algorithms, a comparative pessimistic likelihood estimation (CPLE) algorithm based on trajectory classification is proposed. A framework for automatic detection of anomalous events, which contains vehicle detection, vehicle tracking and trajectory classification, is developed. YOLO v3 is employed to detect the vehicles, and related information about four different types of vehicles is obtained. Online real-time tracking algorithms are used for multi-target tracking of vehicles. Anomalous event vehicle trajectories are obtained for different scenarios. Based on semi-supervised learning, the maximum likelihood method is employed to improve the classification of vehicle trajectories. CPLE is introduced and parameter setting and labeling are centered on comparison and pessimistic rules in order to classify and determine the incident trajectories, consequently, the automatic incident detection algorithm based on vehicle trajectories is proposed. The intelligent inspection system of Gansu Province G312 highway is used as a test object. A total of 1 300 videos were collected. Among them, 530 and 630 tracks are employed as test set and validation set, respectively. By testing difference scenarios of incidents and prewarning, the algorithm accuracy of trajectory classification based on CPLE reaches 89.7%, which is 23.6% higher than that of self-learning and 41.3% higher than that of supervised learning, respectively. Although the accuracy of scattered goods and speeding is averaged about 77.0%, the accuracy of sudden stopping, congestion, and accidents reaches 98.2%, and as for the incident detection influencing traffic seriously, the average accuracy reaches 94%. The proposed method enriches automatic incident detection algorithms and can be considered an alternative for freeway incident detection.
- traffic safety /
- freeway /
- vehicle trajectory /
- YOLO v3 /
- SORT /
- comparative pessimistic likelihood estimation

HTML全文

图 1 YOLO的检测过程

Figure 1. Detection process of YOLO

下载: 全尺寸图片幻灯片

图 2 车辆检测边框图

Figure 2. Bounding box of vehicle detection

下载: 全尺寸图片幻灯片

图 3 车辆轨迹样本图

Figure 3. Sample of vehicle trajectory

下载: 全尺寸图片幻灯片

图 4 有标签样本算法的准确率对比

Figure 4. Accuracy comparison of labeled sample algorithm

下载: 全尺寸图片幻灯片

表 1 数据分类

Table 1. Data classification

类型	正常	异常	测试集	验证集
有标签数据	15	15	530	630
无标签数据	1 000		530	630

下载: 导出CSV

表 2 检测算法性能对比

Table 2. Accuracy comparison of detection algorithm

算法	CP	CN	AC	改善效果/%
CPLE	84	92	0.89
自学习	56	88	0.72	23.6
监督学习	59	96	0.63	41.3

下载: 导出CSV

表 3 异常事件报警正确率统计

Table 3. Statistics of incidents alarming accuracy rate

事件类型	报警总数	正确数	误报数	正确率/%
散落货物	50	36	14	72.0
突发性停车	5 315	5 097	218	95.9
超速	128	105	23	82.0
拥堵	141	139	2	98.6
事故	4	4	0	100.0

下载: 导出CSV

参考文献(19)

[1]	李浩澜. 基于视频图形的高速公路异常事件实时检测系统[D]. 重庆: 重庆理工大学, 2020. LI H L. A real-time detection system for abnormal events of highway based on image[D]. Chongqing: Chongqing University of Technology, 2020. (in Chinese)
[2]	常丹丹. 基于深度学习的公路货物运输量预测方法研究[D]. 西安: 长安大学, 2018. CHANG D D. Research on volume of road cargo transportation forecasting method based on deep learning[D]. Xi'an: Chang'an University, 2018. (in Chinese)
[3]	EVANS J, WATERSON B, HAMILTON A. Evolution and future of urban road incident detection algorithms[J]. Journal of Transportation Engineering Part A: Systems, 2020, 146(6): 03120001. doi: 10.1061/JTEPBS.0000362
[4]	MASTERS P H, LAM J, WONG K. Incident detection algorithms for COMPASS-An advanced traffic management system[C]. Vehicle Navigation and Information Systems Conference, New York, U. S. A. : IEEE, 1991.
[5]	LEVEN M, Krause G M. Incident detection: A Bayesian approach[J]. Transportation Research Record, 1978(682): 52-58.
[6]	LYALL B B. Performance evaluation of the McMaster incident detection algorithm[D]. Ontario: McMaster University, 1991.
[7]	AHMED S A, COOK A R. Application of time-series analysis techniques to freeway incident detection[J]. Transportation Research Record, 1982(841), 19-21.
[8]	CREMER M, SCHUTT H. A comprehensive concept for simultaneous state observation, parameter estimation and incident detection[C]. The 11th International Symposium on Transportation and Traffic Theory, Amsterdam, Netherlands: Elsevier, 1990.
[9]	TENG H, QI Y. Application of wavelet technique to freeway incident detection[J]. Transportation Research Part C: Emerging Technologies, 2003, 11(3/4): 289-308.
[10]	朱金凤. 基于YOLO-v3的高速公路交通事件检测系统研究[D]. 徐州: 中国矿业大学, 2020. ZHU J F. Research on expressway traffic incident detection system based on YOLO-v3[D]. Xuzhou: China University of Mining and Technology, 2020. (in Chinese)
[11]	MICHALOPOULOS P G. Vehicle detection video through image processing: the Autoscope system[J]. Transactions on Vehicular Technology, 1991, 40(1): 21-29. doi: 10.1109/25.69968
[12]	CHARKRABORTY P. Freeway traffic incident detection using large scale traffic data and cameras[D]. Iowa: Iowa State University, 2019.
[13]	冯汝怡, 李志斌, 吴启范, 范昌彦. 航拍视频车辆检测目标关联与时空轨迹匹配[J]. 交通信息与安全, 2021, 9(2): 61-69. doi: 10.3963/j.jssn.1674-4861.2021.02.008 FENG R Y, LI Z B, WU Q F, et al. Association of vehicle object detection and the time-space trajectory matching from aerial videos[J]. Journal of Transport Information and Safety, 2021, 9(2): 61-69. (in Chinese) doi: 10.3963/j.jssn.1674-4861.2021.02.008
[14]	LIN Y, LI L, JING H, et al. Automated traffic incident detection with a smaller dataset based on generative adversarial networks[J]. Accident Analysis & Prevention, 2020(4): 105628.
[15]	乔鹏. 基于深度学习和边缘任务卸载的交通流量检测研究[D]. 西安: 西安电子科技大, 2019. QIAO P. Research on traffic flow detection based on deep learning and edge task offloading[D]. Xi'an: Xi Dian University, 2019. (in Chinese)
[16]	高新闻, 沈卓, 许国耀, 等. 基于多目标跟踪的交通异常事件检测[J]. 计算机应用研究, 2021, 38(6): 1879-1883 https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202106054.htm GAO X W, SHEN Z, XU G Y, et al. Traffic anomaly detection based on multi-target tracking[J]. Application Research of Computers, 2021, 38(6): 1879-1883. (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202106054.htm
[17]	孙士杰. 复杂交通场景下多目标检测跟踪问题研究[D]. 西安: 长安大学, 2019. SUN S J. Research on multiple object detection and tracking problems in complex traffic scenes[D]. Xi'an: Chang'an University, 2019. (in Chinese)
[18]	高铭. 基于深度学习的复杂交通环境下目标跟踪与轨迹预测研究[D]. 长春: 吉林大学, 2020. GAO M. Research on object tracing and trajectory prediction in complex environment based on deep learning[D]. Changchun: Jilin University, 2020. (in Chinese)
[19]	杨明辉. 基于循环神经网络的运动目标轨迹预测[D]. 武汉: 武汉大学, 2019. YANG M H. Trajectory prediction of moving target based on recurrent neural network[D]. Wuhan: Wuhan University, 2019. (in Chinese)