摘 要：据资料统计，有近 90% 的交通事故源于驾驶员的危险驾驶行为。人们对于驾驶员在行驶过程中低头、看手机、双手离开方向盘、视线未朝向前方、闭眼等危险驾驶行为的关注度仍未提高。文章采用 Jetson Xavier NX、YOLOv5 算法、百度AI 平台实现对驾驶员行车过程中玩手机、抽烟、未系安全带、双手离开方向盘、视线未朝向前方、闭眼、打哈欠、低头等 8 种危险驾驶行为进行检测，并结合当下疫情加入口罩检测功能。利用语音技术合成危险行为提醒语音，实时播报。通过实验验证了该检测系统的可行性与实用性。
关键词：Jetson Xavier NX；驾驶行为分析；YOLO v5；百度 AI 平台
DOI:10.19850/j.cnki.2096-4706.2022.09.025
基金项目： 中原科技创新领军人才（214200510013）； 国家自然科学基金项目（62171318）； 河南省高校重点科研项目（21A510016，21A520052）；留学人员科研资助和创业启动项目（HRSS2021[36]）；校内重大项目成果培育计划（K2020ZDPY02）
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2022）09-0100-04
Design of Intelligent Driver Behavior Detection System Based on Jetson Xavier Nx
GUO Xiaoyu, YANG Lei, LU Deyuan
(School of Electronic and Information, Zhongyuan University of Technology, Zhengzhou 450007, China)
Abstract: According to statistics, nearly 90% of traffic accidents come from the dangerous driving behavior of drivers. People’s attention to dangerous driving behaviors of drivers when driving have not been improved, such as lowering head, watching mobile phone, their both hands leaving the steering wheel, not looking forward and closing eyes. This paper adopts Jetson Xavier NX, YOLOv5 algorithm and Baidu AI platform to detect 8 dangerous driving behaviors from driver when driving, such as playing with mobile phones, smoking, not wearing seat belts, their both hands leaving the steering wheel, not looking forward, closing eyes, yawning and lowering head, and adds mask detection function in combination with the current epidemic situation. Use voice technology to synthesize dangerous behavior warning voice and broadcast it in real time. The feasibility and practicability of the detection system are verified by experiments.
Keywords: Jetson Xavier NX; driving behavior analysis; YOLO v5; Baidu AI platform
参考文献：
[1] 胡林，郭广涛，黄晶，等 . 驾驶人熟悉程度对其安全驾驶行为的影响研究综述 [J/OL]. 中国公路学报：1-21[2022-02-09].http://kns.cnki.net/kcms/detail/61.1313.U.20210616.1506.004.html.
[2] ZHU Z Q，CHAN K W，BU S Q，et al. RealTime interaction of active distribution network and virtual microgrids: Market paradigm and data-driven stakeholder behavior analysis [J].Applied Energy，2021，297：117107.[3] GERSHON P，SEAMAN S，MEHLER B，et al.
Driver behavior and the use of automation in real-world driving [J]. Accident Analysis and Prevention，2021，158：106217.
[4] MARAFIE Z，LIN K J，WANG D B，et al. AutoCoach: An Intelligent Driver Behavior Feedback Agent with Personality-Based Driver Models [J].Electronics，2021，（10）：1361.
[5] ZHANG Z W，SU Tao. Behavioral Analysis and Immunity Design of the RO-Based TRNG under Electromagnetic Interference [J]. Electronics，2021，10（11）：1347.
[6] GERSHON P，SEAMAN S，MEHLER B，et al. Driver behavior and the use of automation in real-world driving [J].Accident analysis and prevention，2021，158：106217.
作者简介：郭小玉（1995—），女，汉族，河南开封人，硕士研究生在读，主要研究方向：计算机视觉、深度学习；杨蕾（1979—），女，回族，河南洛阳人，教授，博士，主要研究方向：图像处理、计算机视觉；鹿德源（1998—），男，汉族，河南信阳人，硕士研究生在读，主要研究方向：计算机视觉、深度学习。