YOLOv5监测算法下轨道交通应急预警感知系统构建

doi:10.16265/j.cnki.issn1003-3033.2024.09.0120

摘要/Abstract

摘要：

为减少城市轨道交通站点中因大客流状态下应急预警系统的滞后性而发生的行人安全问题,选取YOLOv5算法预测客流信息,并利用人工神经网络(ANN)模型来构建城市轨道交通应急预警感知系统。首先,通过模型训练超参数优化和先验框参数优化改进YOLOv5算法;然后,通过预警指标选取、权重分析和阈值界定设计应急预警感知系统;最后,采用Matlab软件构建基于ANN的自组织竞争网络应急预警模型,将优化后的YOLOv5算法采集的数据通过计算代入应急预警感知系统中,通过试验验证应急预警感知系统。结果表明:优化后的YOLOv5算法相较原算法,城市轨道交通大客流状态下行人目标监测精确度提高7.04%;由优化后的YOLOv5算法所采集到的行人数据代入构建的应急预警感知系统后得到的判断结果与实际预警等级一致,证明了该系统的可实施性和有效性,有助于提高城市轨道交通应急预警水平。

关键词: YOLOv5算法, 监测算法, 轨道交通, 应急预警感知系统, 目标监测, 超参数

Abstract:

In order to reduce the pedestrian safety problems caused by the lag of the emergency warning system in urban rail transit stations under large passenger flow conditions, the YOLOv5 algorithm was selected to predict passenger flow information. The artificial neural network (ANN) model was used to construct the urban rail transit emergency warning perception system. Firstly, the YOLOv5 algorithm was improved by optimizing the model training hyperparameters and prior frame parameters. Then, the emergency warning perception system was designed by selecting warning indicators, weight analysis and threshold definition. Finally, the self-organizing competitive network emergency warning model based on ANN was constructed by using Matlab software. The data collected by the optimized YOLOv5 algorithm were substituted into the emergency warning perception system through calculation, and the emergency warning perception system was verified by experiments. The results show that the optimized YOLOv5 algorithm can improve the accuracy of pedestrian target monitoring under large passenger flow conditions of urban rail transit by 7.04%. The judgment results obtained by substituting the pedestrian data collected by the optimized YOLOv5 algorithm into the constructed emergency warning perception system are consistent with the actual warning level, which proves the feasibility and effectiveness of the system and helps to improve the emergency warning level of urban rail transit.

Key words: YOLOv5 algorithm, detection algorithm, rail transit, emergency warning perception system, target monitoring, hyperparameter

中图分类号:

X924.4安全控制技术

雷斌, 闫浪浪, 余华, 温岩, 张亮, 李哲旭. YOLOv5监测算法下轨道交通应急预警感知系统构建[J]. 中国安全科学学报, 2024, 34(9): 225-233.

LEI Bin, YAN Langlang, YU Hua, WEN Yan, ZHANG Liang, LI Zhexu. Construction of rail transit emergency early warning perception system under YOLOv5 detection algorithm[J]. China Safety Science Journal, 2024, 34(9): 225-233.

图/表 16

表1

图1

表2

图2

图3

图4

表3

表4

表5

表6

表7

表8

表9

表10

表11

表12

参考文献 22

[1]	SHI Jungang, YANG Lixing, YANG Jing, et al. Service-oriented train timetabling with collaborative passenger flow control on an oversaturated metro line: an integer linear optimization approach[J]. Transportation Research Part B Methodological, 2018, 110(4):26-59.
[2]	DU Peng, HAMDULLA A. Infrared moving small-target detection using spatial-temporal local difference measure[J]. IEEE Geoscience and Remote Sensing Letters, 2020, 17(10): 1817-1821. doi: 10.1109/LGRS.2019.2954715
[3]	LI Zhilei, LI Jun, MA Yunqing, et al. Spatio-temporal adaptive network with bidirectional temporal difference for action recognition[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023, 33(9): 5174-5185.
[4]	XIE Zhao, CHEN Jiansong, WU Kewei, et al. Global temporal difference network for action recognition[J]. IEEE Transactions on Multimedia, 2023, 25: 7594-7606.
[5]	何文韬, 徐永能, 谭忠磷. 轨道交通动态异物轨迹追踪与预测研究[J]. 兵器装备工程学报, 2022, 43(9):109-114.
	HE Wentao, XU Yongneng, TAN Zhonglin. Research on trajectory tracking and prediction of dynamic foreign bodies in rail transit[J]. Journal of Ordnance Equipment Engineering, 2022, 43(9): 109-114.
[6]	盛蕊. 基于视频图像的地铁车站客流安全状态识别[D]. 南京: 南京理工大学, 2020.
	SHENG Rui. Safety status identification of passenger flow in metro station based on video image[D]. Nanjing: Nanjing University of Science & Technology, 2020.
[7]	闫香玲. 基于视频监控的地铁车站客流安全状态智能感知关键技术研究[D]. 北京: 北京交通大学, 2020.
	YAN Xiangling. Research on key technologies for intelligent perception of passenger flow safety status in metro station based on video surveillance[D]. Beijing: Beijing Jiaotong University, 2020.
[8]	常捷, 张国维, 陈文江, 等. 基于YOLO-V3算法的加油站不安全行为检测[J]. 中国安全科学学报, 2023, 33(2):31-37. doi: 10.16265/j.cnki.issn1003-3033.2023.02.0298
	CHANG Jie, ZHANG Guowei, CHEN Wenjiang, et al. Gas station unsafe behavior detection based on YOLO-V3 algorithm[J]. China Safety Science Journal, 2023, 33(2): 31-37. doi: 10.16265/j.cnki.issn1003-3033.2023.02.0298
[9]	LAI Yeonglin, CHOU Yunghua, CHANG Lichih. An intelligent IoT emergency vehicle warning system using RFID and Wi-Fi technologies for emergency medical services[J]. Technology and Health Care, 2018, 26(1): 43-55. doi: 10.3233/THC-171405 pmid: 29060952
[10]	YANG Yongjun, ZHAO Qian, ZHANG Jianli, et al. Research and application of marine environmental emergencies prediction and early warning system in China[J]. Environmental Pollution & Control, 2023, 45(11): 1597-1602.
[11]	LI Jing, JI Zheng. Research on planning technology of key elements for emergency warning and notification system[J]. Journal of Safety Science and Technology, 2021, 17(8): 183-188.
[12]	YU Xi, SUN Fuquan. Intelligent urban emergency early warning system based on dynamic rough set and cloud computing[C]. Proceedings of 2013 IEEE 4^th International Conference on Software Engineering and Service Science, 2013: 701-704.
[13]	李美洁. 基于视频的地铁客流安全状态智能感知技术研究[D]. 北京: 北京交通大学, 2021.
	LI Meijie. Research on intelligent perception technology of metro passenger flow security status based on video[D]. Beijing: Beijing Jiaotong University, 2021.
[14]	LI Haoran, XU Li, ZHANG Yin, et al. YOLOv5 vehicle detection model in fog based on channel attention enhancement[J]. Journal of Physics: Conference Series, 2023, 2589(1): DOI:10.1088/1742-6596/2589/1/012004.
[15]	牛为华, 殷苗苗. 基于改进的YOLO v5的道路小目标检测算法[J]. 传感技术学报, 2023, 36(1):36-44.
	NIU Weihua, YIN Miaomiao. Road small target detection algorithm based on improved YOLO v5[J]. Chinese Journal of Sensors and Actuators, 2023, 36(1): 36-44.
[16]	XUE Jinlin, CHENG Feng, et al. Detection of farmland obstacles based on an improved YOLOv5s algorithm by using CIoU and anchor box scale clustering[J]. Sensors, 2022, 22(5):DOI:10.3390/s22051790.
[17]	吕跃进. 指数标度判断矩阵的一致性检验方法[J]. 统计与决策, 2006(18):31-32.
[18]	张晗. 城市轨道交通运营安全综合评估预警平台设计研究[D]. 北京: 北京交通大学, 2012.
	ZHANG Han. Study on the design of urban rail transit operational safety assessment early warning platform[D]. Beijing: Beijing Jiaotong University, 2012.
[19]	李冰玉, 秦孝敏. 城市轨道交通网络大客流拥堵传播机理研究[J]. 中国安全科学学报, 2016, 26(1):162-168.
	LI Bingyu, QIN Xiaomin. Study on propagation mechanism of mass passenger flow congestion in urban rail transit network[J]. China Safety Science Journal, 2016, 26(1): 162-168.
[20]	GB50157—2013,地铁设计规范[S].
	GB50157-2013, Subway design standards[S].
[21]	GAJERA H, DAWDA N, DAVE S, et al. Evaluating operational performance of city bus transit service using tcqsm guidelines: a case study of Surat city[J]. Journal of the Eastern Asia Society for Transportation Studies, 2019,13:1361-1381.
[22]	谭一帆. 综合客运枢纽客流预警研究[D]. 成都: 西南交通大学, 2019.
	TAN Yifan. Research on passenger traffic warning of the comprehensive passenger transport hub[D]. Chengdu: Southwest Jiaotong University, 2019.

方法	实际人数 N₁	监测人数 N₂	漏检人数 N₃	监测精度 A₀/%
Faster R-CNN	1 321	1 140	181	86.30
YOLOv3	1 321	1 110	211	84.02
YOLO v5	1 321	1 123	198	85.01

方法	N₁	N₂	N₃	A₀/%
原YOLOv5	1 321	1 123	198	85.01
优化后YOLOv5	1 321	1 216	105	92.05

B_i与B_j相比较的重要程度	标度
同等重要	1
稍微重要	3
明显重要	5
强烈重要	7
绝对重要	9
B_i与B_j重要程度介于各等级之间,分别取2、4、6、8

预警指标	客流量	区域面积密度	排队滞留度	权重
客流量	1	1.482	1.539	0.34
区域面积密度	1.816	1	2.137	0.4
排队滞留度	0.356	1.300	1	0.26

s/(m²·s·人^-1)	预警等级
≥1.2	Ⅳ
[0.9,1.2)	Ⅲ
[0.3,0.9)	Ⅱ
<0.3	Ⅰ