分割可跨越车道分界线的多视角视频车速提取方法

doi:10.13229/j.cnki.jdxbgxb.20230818

摘要/Abstract

摘要：

针对现有视频交通参数提取方法过度依赖人工标注、单一视角无法有效修正现场车辆动态行驶偏差的问题，本文提出一种分割可跨越车道分界线的多视角视频车速提取方法。该方法由标注点自动生成模块、多视角纠偏模块构成：标注点自动生成模块通过构建基于等长车道分界线的参照块，实现自动生成标注点的过程；多视角纠偏模块通过车辆与车道分界线的多种映射方式和基于平均速度概率密度函数的纠偏测速方法，校正车辆在动态行驶时产生的两类偏差。在公开数据集和实测数据集上的实验结果表明，本文所提方法的车速提取精度优于其他测速方法，且具有一定的普适性。

关键词: 计算机应用, 智能交通, 车道分界线, 多视角, 车速检测, 计算机视觉

Abstract:

Aiming at the problem that the existing video traffic parameter extraction method relies too much on manual labeling and the single perspective cannot effectively correct the dynamic driving deviation of on-site vehicles， a multi-view video traffic parameter extraction method that can split the lane demarcation line is proposed. This method consists of an automatic generation module for labeling points and a module for multi-view correction. The automatic generation of label points module realizes the process of automatically generating label points by constructing a reference block based on the dividing line of equal-length lanes. The multi-view deviation correction module proposes a variety of mapping methods for the boundary between vehicles and lanes and a correction speed measurement method based on the average speed probability density function to correct two types of deviations generated by vehicles when driving dynamically. The experimental results on the public dataset and the measured dataset show that the speed extraction accuracy of the proposed method is better than that of other speed measurement methods， and has certain universality.

Key words: computer application, intelligent transportation, lane demarcation line, multiple perspectives, vehicle speed detection, computer vision

中图分类号:

TP391

侯越,郭劲松,林伟,张迪,武月,张鑫. 分割可跨越车道分界线的多视角视频车速提取方法[J]. 吉林大学学报(工学版), 2025, 55(5): 1692-1704.

Yue HOU,Jin-song GUO,Wei LIN,Di ZHANG,Yue WU,Xin ZHANG. Multi-view video speed extraction method that can be segmented across lane demarcation lines[J]. Journal of Jilin University(Engineering and Technology Edition), 2025, 55(5): 1692-1704.

图/表 22

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

图11

表1

图12

表2

图13

表3

图14

表4

图15

图16

图17

表5

参考文献 16

[1]	姚俊峰, 何瑞, 史童童, 等. 基于机器学习的交通流测方法综述[J]. 交通运输工程学报, 2023, 23(3):44-67.
	Yao Jun-feng, He Rui, Shi Tong-tong, et al. Revi-ew on machine learning-based traffic flow prediction methods[J]. Journal of Traffic and Transportation Engineering, 2023, 23(3): 44-67.
[2]	Luvizon D C, Nassu B T, Minetto R. A video-based system for vehicle speed measurement inurban roadways[J]. IEEE Transactions on Intelligent Transportation Systems, 2016, 18(6): 1393-1404.
[3]	姜桂艳, 郭海锋, 吴超腾. 基于感应线圈数据的城市道路交通状态判别方法[J]. 吉林大学学报: 工学版, 2008(): 37-42.
	Jiang Gui-yan, Guo Hai-feng, Wu Chao-teng. Urban road traffic status discrimination method based on induction coil data[J]. Journal of Jilin University (Engineering and Technology Edition), 2008,38(Sup. l): 37-42.
[4]	刘丽, 李勇. 一种基于自聚焦的交通雷达测速算法[J].南京航空航天大学学报, 2013, 45(6): 843-848.
	Liu Li, Li Yong. A traffic radar speed measurement algorithm based on self-focusing[J]. Journal of Nanjing University of Aeronautics and Astronautics, 2013, 45(6): 843-848.
[5]	吴杭彬, 刘豆, 刘启远, 等.基于断面激光扫描的移动车辆交通流参数提取[J]. 同济大学学报: 自然科学版, 2017, 45(7): 1069-1074.
	Wu Hang-bin, Liu Dou, Liu Qi-yuan, et al. Parameter extraction of traffic flow of moble vehicles based on cross-sectional laser scanning[J]. Journal of Tongji University (Natural Science Edition), 2017, 45(7): 1069-1074.
[6]	王成钢, 刘海, 谭中慧. 一种基于视频图像的车辆测速算法[J].上海船舶运输科学研究所学报, 2019, 42(1): 65-69.
	Wang Cheng-gang, Liu Hai, Tan Zhong-hui. A vehicle speed measurement algorithm based on video image[J]. Journal of Shanghai Institute of Shipping Transportation Science, 2019, 42(1): 65-69.
[7]	Rahim H A, Ahmad R B, Zain A S M, et al. An adapted point based tracking for vehicle speed estimation in linear spacing[C]∥International Conference on Computer and Communication Engineering, Zurich, Switzerland, 2010: 1-4.
[8]	Peruničić A, Djukanović S, Cvijetić A. Vision-based vehicle speed estimation using the YOLO detector and RNN[C]∥27th International Conference on Information Technology, Wuhan, China, 2023: 1-4.
[9]	田会娟, 刘嘉伟, 翟佳豪,等. 基于多入侵线的视频车速检测方法[J]. 交通运输系统工程与信息, 2022, 22(1): 49-56.
	Tian Hui-juan, Liu Jia-wei, Zhai Jia-hao, et al. Video-based vehicle speed measurement method using multiple intrusion lines[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(1): 49-56.
[10]	余越, 柏兴, 狄岚. 基于YOLO网络和小波降噪的视频车辆跟踪测速优化[J]. 应用科技,2023, 50(1):14-20.
	Yu Yue, Bai Xing, Di Lan. Video vehicle tracking and speed measurement optimization based on YOLO network and wavelet noise reduction[J]. Applied Science and Technology, 2023, 50(1): 14-20.
[11]	中华人民共和国国家质量检验检疫总局,中国国家标准化委员会. 道路交通标志和标线第3部分: 道路交通标线: [S].
[12]	于艳玲, 王韬, 袁彬, 等. 基于视频的车速检测算法研究[J]. 现代电子技术, 2013, 36(3): 158-161.
	Yu Yan-ling, Wang Tao, Yuan Bin, et al. Research on vehicle speed detection algorithm based on video[J]. Modern Electronics Technique, 2013, 36(3): 158-161.
[13]	Sochor J, Juránek R, Špaňhel J, et al. Comprehensive data set for automatic single camera visual speed measurement[J]. IEEE Transactions on Intelligent Transportation Systems, 2019, 20(5): 1633-1643.
[14]	Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection[C]∥2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016: 779-788.
[15]	Bewley A, Ge Z, Ott L, et al. Simple online and realtime tracking[C]∥2016 IEEE International Conference on Image Processing (ICIP), Phoenix, USA, 2016: 3464-3468.
[16]	Javadi S, Dahl M, Pettersson M I, et al. Design of a video-based vehicle speed measurement system: an uncertainty approach[C]∥2018 Joint 7th International Conference on Informatics, Electronics & Vision (ICIEV) and 2018 2nd International Conference on Imaging, Vision & Pattern Recognition, Kitakyushu, Japan, 2018: 44-49.

相关文章 15

[1]	赵宏伟,周明珠,刘萍萍,周求湛. 基于置信学习和协同训练的医学图像分割方法[J]. 吉林大学学报(工学版), 2025, 55(5): 1675-1681.
[2]	申自浩,高永生,王辉,刘沛骞,刘琨. 面向车联网隐私保护的深度确定性策略梯度缓存方法[J]. 吉林大学学报(工学版), 2025, 55(5): 1638-1647.
[3]	王友卫,刘奥,凤丽洲. 基于知识蒸馏和评论时间的文本情感分类新方法[J]. 吉林大学学报(工学版), 2025, 55(5): 1664-1674.
[4]	王军,司昌馥,王凯鹏,付强. 融合集成学习技术和PSO-GA算法的特征提取技术的入侵检测方法[J]. 吉林大学学报(工学版), 2025, 55(4): 1396-1405.
[5]	刘广文,赵绮莹,王超,高连宇,才华,付强. 基于渐进递归的生成对抗单幅图像去雨算法[J]. 吉林大学学报(工学版), 2025, 55(4): 1363-1373.
[6]	程德强,刘规,寇旗旗,张剑英,江鹤. 基于自适应大核注意力的轻量级图像超分辨率网络[J]. 吉林大学学报(工学版), 2025, 55(3): 1015-1027.
[7]	刘广文,谢欣月,付强,才华,王伟刚,马智勇. 基于时空模板焦点注意的Transformer目标跟踪算法[J]. 吉林大学学报(工学版), 2025, 55(3): 1037-1049.
[8]	徐涛,孔帅迪,刘才华,李时. 异构机密计算综述[J]. 吉林大学学报(工学版), 2025, 55(3): 755-770.
[9]	江晟,王祎笛,谢睿麟,夏淼磊. 基于门控循环网络时空关联的交通事故预测模型[J]. 吉林大学学报(工学版), 2025, 55(3): 954-962.
[10]	李立,鲍宇健,杨文臣,楚庆玲,汪贵平. 路侧多源感知数据集规范化构建方法[J]. 吉林大学学报(工学版), 2025, 55(2): 529-536.
[11]	赵孟雪,车翔玖,徐欢,刘全乐. 基于先验知识优化的医学图像候选区域生成方法[J]. 吉林大学学报(工学版), 2025, 55(2): 722-730.
[12]	车翔玖,武宇宁,刘全乐. 基于因果特征学习的有权同构图分类算法[J]. 吉林大学学报(工学版), 2025, 55(2): 681-686.
[13]	才华,郑延阳,付强,王晟宇,王伟刚,马智勇. 基于多尺度候选融合与优化的三维目标检测算法[J]. 吉林大学学报(工学版), 2025, 55(2): 709-721.
[14]	蔡晓东,周青松,张言言,雪韵. 基于动静态和关系特征全局捕获的社交推荐模型[J]. 吉林大学学报(工学版), 2025, 55(2): 700-708.
[15]	郭晓然,王铁君,闫悦. 基于局部注意力和本地远程监督的实体关系抽取方法[J]. 吉林大学学报(工学版), 2025, 55(1): 307-315.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

实验环境	配置
操作系统	Windows 11
处理器	AMD Ryzen 7 5800H with Radeon Graphics
内存	16.0 GB
编程语言	Python

分割次数	第一次抽样		第二次抽样		第三次抽样
分割次数	平均误差率	最大误差率	平均误差率	最大误差率	平均误差率	最大误差率
r=0	1.64	2.75	1.75	2.87	1.34	2.23
r=1	1.53	2.56	1.45	2.42	1.27	1.94
r=2	1.35	2.38	1.33	2.46	0.97	1.94
r=3	1.18	2.34	1.06	2.28	0.91	1.71

映射方式	正视角		侧视角		模糊视角
映射方式	平均误差率	最大误差率	平均误差率	最大误差率	平均误差率	最大误差率
LL-TR	1.06	2.28	1.71	2.84	1.23	2.47
LL-PA	2.17	3.37	1.18	2.34	2.14	3.30
LL-MX	1.75	2.47	1.48	2.34	0.91	1.71
LL-N	2.71	3.53	2.73	3.28	3.45	4.53

组别	正视角		侧视角		模糊视角
组别	平均误差率	最大误差率	平均误差率	最大误差率	平均误差率	最大误差率
对照组	2.43	3.96	2.19	3.37	1.82	2.52
实验组	1.18	2.34	1.06	2.28	0.91	1.71

方法	平均误差率
方法	BrnoCompSpeed	LzVedioSpeed
GPS	2.18	—
RADAR	1.33	—
T-LL	1.71	1.78
M-LL	1.40	1.42
Vcm	3.06	3.02
video-speed	1.84	4.21
Vcm+centroid	5.43	2.89
本文	1.12	1.16