都市快轨交通

基于面部特征与Bi-LSTM的智轨驾驶员疲劳检测算法

PDF下载

闵思远 ¹^,² , 孟新宇 ¹^,² , 张锐骥 ¹^,² , 辛开元 ¹^,² , 殷勇 ¹^,²

都市快轨交通 | 人工智能 2025,38(6): 116-124

收起

都市快轨交通 |人工智能 2025 , 38 (6) : 116 -124

基于面部特征与Bi-LSTM的智轨驾驶员疲劳检测算法

全屏

闵思远¹^,², 孟新宇¹^,², 张锐骥¹^,², 辛开元¹^,², 殷勇¹^,²

作者信息

^1.西南交通大学交通运输与物流学院，成都　611756

^2.西南交通大学宜宾研究院，成都　644000

通讯作者:

殷勇，男，博士，副教授，从事交通运输组织优化研究，yinyong@home.swjtu.edu.cn

作者简介:

闵思远，男，硕士研究生，从事驾驶疲劳检测研究，minsiyuan@my.swjtu.edu.cn

Fatigue Detection Algorithm for Intelligent Rail Drivers Based on Facial Features and Bi-LSTM

Siyuan MIN¹^,², Xinyu MENG¹^,², Ruiji ZHANG¹^,², Kaiyuan XIN¹^,², Yong YIN¹^,²

Affiliations

^1.School of Transportation and Logistics, Southwest Jiaotong University, Chengdu 611756

^2.Yibin Research Institute, Southwest Jiaotong University, Yibin 644000

出版时间: 2025-12-10 doi: 10.3969/j.issn.1672-6073.2025.06.015

文章导航

摘要

收起

针对现有视觉疲劳检测方法普遍存在场景适应性不足以及时间特征依赖关系建模不充分的问题，提出一种基于面部多特征与双向长短期记忆网络的智轨驾驶员疲劳检测算法。首先，算法通过开源机器学习与计算机视觉库（Dlib）定位驾驶员面部关键点，计算出眼睛纵横比（eye aspect radio，EAR）、嘴部纵横比（mouth aspect ratio，MAR）、头部姿态角等5个疲劳特征参数，并提出纵向头部巡视行为频率用于分心检测；其次，通过双向长短期记忆网络（bidirectional long short-term memory，Bi-LSTM）网络建立时序依赖关系，完成驾驶状态的判定；再次，引入时间注意力来增强关键特征的表征能力，提高算法整体性能；最后，与基于眼部状态检测的模型相比，本文模型的准确率为94.24%，F1-Score值由0.8324提升至0.9027，验证了本文方法的有效性与可行性。研究结果可为公共交通安全保障提供参考。

关键词

深度学习 / 面部多特征 / 疲劳检测 / Bi-LSTM / 人脸关键点

Abstract

收起

Aiming at the common issues of insufficient scene adaptability and inadequate modeling of temporal feature dependencies in existing visual fatigue detection methods, this paper proposes an intelligent rail driver fatigue detection algorithm based on facial multi-features and Bi-LSTM. First, the algorithm uses Dlib to locate the driver's facial keypoints and calculates five fatigue feature parameters, including the Eye Aspect Ratio (EAR), Mouth Aspect Ratio (MAR), and head pose angles. It also proposes the frequency of longitudinal head scanning behavior for distraction detection. Second, the Bi-LSTM network is employed to establish temporal dependencies and complete the determination of the driving state. Third, temporal attention is introduced to enhance the representation capability of key features, improving the overall performance of the algorithm. Finally, the research results show that, compared to traditional models, the proposed method achieves an accuracy of 94.24% and improves the F1-Score from 0.8324 to 0.9027, validating the effectiveness and feasibility of the proposed approach. The findings can provide a reference for public transportation safety assurance.

Key words

deep learning / multiple facial features / fatigue detection / Bi-LSTM / facial landmarks

引用本文

闵思远, 孟新宇, 张锐骥, 辛开元, 殷勇. 基于面部特征与Bi-LSTM的智轨驾驶员疲劳检测算法. 都市快轨交通, 2025 , 38 (6) : 116 -124 . DOI: 10.3969/j.issn.1672-6073.2025.06.015

Siyuan MIN, Xinyu MENG, Ruiji ZHANG, Kaiyuan XIN, Yong YIN. Fatigue Detection Algorithm for Intelligent Rail Drivers Based on Facial Features and Bi-LSTM[J]. Urban Rapid Rail Transit, 2025 , 38 (6) : 116 -124 . DOI: 10.3969/j.issn.1672-6073.2025.06.015

基金

收起

宜宾双城市校协议专项科研经费科技项目管理委员会(YBSCXY2023020011)
国家重点研发计划(2022YFB4300503)

参考文献

收起

文献

收起

参考文献引证文献

排序方式：

[1]

LIU Z J, GU C H, XIE Y, et al. Realistic sketch face generation via sketch-guided incomplete restoration[C]//2023 IEEE 29th International Conference on Parallel and Distributed Systems (ICPADS). Ocean Flower Island, 2023: 32-37.

[2]

CHANG R C, WANG C Y, CHEN W T, et al. Drowsiness detection system based on PERCLOS and facial physiological signal[J]. Sensors, 2022, 22(14): 5380.

[3]

WIERWILLE W, WREGGIT S, KIRN C, et al. Research on vehicle-based driver status/performance monitoring; development, validation, and refinement of algorithms for detection of driver drowsiness. final report[R]. 2019.

[4]

FRIEDRICHS F, YANG B. Camera-based drowsiness reference for driver state classification under real driving conditions[C]//2010 IEEE Intelligent Vehicles Symposium. La Jolla, 2010: 101-106.

[5]

LI K N, GONG Y B, REN Z L. A fatigue driving detection algorithm based on facial multi-feature fusion[J]. IEEE Access, 2020, 8: 101244-101259.

[6]

ZHANG F, SU J J, GENG L, et al. Driver fatigue detection based on eye state recognition[C]//2017 International Conference on Machine Vision and Information Technology (CMVIT). Singapore. 2017: 105-110.

[7]

DWIVEDI K, BISWARANJAN K, SETHI A. Drowsy driver detection using representation learning[C]//2014 IEEE International Advance Computing Conference (IACC). Gurgaon, 2014: 995-999.

[8]

LIU M Z, XU X, HU J, et al. Real time detection of driver fatigue based on CNN-LSTM[J]. IET image processing, 2022, 16(2): 576-595.

[9]

杨艳艳, 李雷孝, 林浩. 提取驾驶员面部特征的疲劳驾驶检测研究综述[J]. 计算机科学与探索, 2023, 17(6): 1249-1267. YANG Yanyan, LI Leixiao, LIN Hao. Review of research on fatigue driving detection based on driver facial features[J]. Journal of frontiers of computer science and technology, 2023, 17(6): 1249-1267.

[10]

CAO Z W, CHU Z C, LIU D F, et al. A vector-based representation to enhance head pose estimation[C]//2021 IEEE Winter Conference on Applications of Computer Vision (WACV). Waikoloa, 2021: 1187-1196.

[11]

YE M, ZHANG W W, CAO P C, et al. Driver fatigue detection based on residual channel attention network and head pose estimation[J]. Applied sciences, 2021, 11(19): 9195.

[12]

GERS F A, SCHMIDHUBER J, CUMMINS F. Learning to forget: continual prediction with LSTM[J]. Neural computation, 2000, 12(10): 2451-2471.

[13]

HU X M, YUAN S S, XU F Z, et al. Scalp EEG classification using deep Bi-LSTM network for seizure detection[J]. Computers in biology and medicine, 2020, 124: 103919.

[14]

ABTAHI S, OMIDYEGANEH M, SHIRMOHAMMADI S, et al. YawDD: a yawning detection dataset[C]//Proceedings of the 5th ACM Multimedia Systems Conference. Singapore, 2014: 24-28.

[15]

雷富强, 张博雅, 张一帆, 等. 基于RetinaFace和ERT的眼部疲劳检测方法[J]. 计算机应用与软件, 2024, 41(10): 227-232. LEI Fuqiang, ZHANG Boya, ZHANG Yifan, et al. Eye fatigue detection based on retinaface and ert[J]. Computer applications and software, 2024, 41(10): 227-232.

[16]

周先春, 邹清宇, 陆滇. 基于卷积神经网络的疲劳检测改进算法[J]. 计算机应用与软件, 2024, 41(6): 156-160, 168. ZHOU Xianchun, ZOU Qingyu, LU Dian. An improved algorithm for fatigue detection based on cnn[J]. Computer applications and software, 2024, 41(6): 156-160.

2025年第38卷第6期

PDF下载

引用本文

BibTeX

文章信息

doi: 10.3969/j.issn.1672-6073.2025.06.015

接收时间：2025-05-14
首发时间：2026-10-19
出版时间：2025-12-10

补充材料

相关文章

文章信息

作者

出版历史

收稿日期：2025-05-14
修回日期：2025-07-10

基金

宜宾双城市校协议专项科研经费科技项目管理委员会(YBSCXY2023020011)

国家重点研发计划(2022YFB4300503)

作者信息

^1.西南交通大学交通运输与物流学院，成都　611756

^2.西南交通大学宜宾研究院，成都　644000

通讯作者:

殷勇，男，博士，副教授，从事交通运输组织优化研究，yinyong@home.swjtu.edu.cn

参考文献

分享链接

https://castjournals.cast.org.cn/joweb/dskgjt/CN/10.3969/j.issn.1672-6073.2025.06.015

分享至

全文二维码

扫描看全文

引用本文

BibTeX

本文的引用情况

2种不同金属材料的力学参数

科 Family	属数 Number of genus	种数 Number of species	占总种数比例 Percentage of total species (%)	属 Genus	种数 Number of species	占总种数比例 Percentage of total species (%)
鹅膏菌科Amanitaceae	2	11	5.26	鹅膏菌属 Amanita	10	4.78
小菇科 Mycenaceae	2	12	5.74	丝盖伞属 Inocybe	5	2.39
多孔菌科 Polyporaceae	8	14	6.70	蜡蘑属 Laccaria	5	2.39
红菇科 Russulaceae	3	23	11.00	小皮伞属 Marasmius	6	2.87
				小菇属 Mycena	11	5.26
				光柄菇属 Pluteus	5	2.39
				红菇属 Russula	17	8.13
				栓菌属 Trametes	5	2.39

关闭全屏

BibTeX
EndNote
RefWorks
TxT