摘 要:针对高速公路路段长,无法密集布置用于检测车祸、车流量的传感器这一问题,文章设计一种可用于在高速公路上监测交通情况的无人机,同时设计一种车流量预测算法,可以较为准确地预测某一路段一定时间内的车流量。无人机搭载摄像头并通过图传系统将高速公路的实时运行情况传回地面站,采用 YOLOv3 算法计算视野内的车辆数目,使用改进后的灰色预测算法预测之后到来的车辆数目。实验结果表明,该无人机可以实时采集并预测公路环境,具有良好的可行性和实用性。
关键词:无人机;灰色预测;YOLOv3;交通预测
DOI:10.19850/j.cnki.2096-4706.2022.013.025
基金项目:河北省省属高等学校基本科研业务费研究项目(50199990400); 省级大学生创新创业训练计划(S202010107059);河北省大中学生科技创新能力培育专项项目(22E50109D)
中图分类号: TP391.4 文献标识码: A 文章编号:2096-4706(2022)13-0101-05
Design of UAV for Dynamic Traffic Information Transmission Based on YOLOv3
ZHANG Xuefeng, YU Yang, REN Bin
(School of Mechanical Engineering, Shijiazhuang Tiedao University, Shijiazhuang 050043, China)
Abstract: Aiming at the problem that the highway section is long and cannot be densely arranged with sensors for detecting traffic accidents and traffic flow, this paper designs a UAV that can be used to monitor traffic conditions on the highway, and designs a traffic flow prediction algorithm, which can accurately predict the traffic flow of a certain section in a certain period of time. The UAV is equipped with a camera and sends the real-time operation of the highway back to the ground station through the image transmission system. The number of vehicles in the field of vision is calculated by using the YOLOv3 algorithm, and the number of vehicles coming later is predicted by using the improved gray prediction algorithm. The experimental results show that the UAV can collect and predict the road environment in real time, and has good feasibility and practicability.
Keywords: UAV; grey prediction; YOLOv3; traffic forecast
参考文献:
[1] 王永杰 . 无人机倾斜摄影测量在海外公路勘测中的应用[J]. 科学技术创新,2021(5):128-129.
[2] 石秀 . 基于高速公路违章检测的无人机地面站的设计实现[D]. 南京:南京邮电大学,2020.
[3] 宋宇,陶柳 . 无人机测绘技术在山区公路选线测量中的应用 [J]. 交通世界,2020(35):65-66.
[4] 樊宝安 . 基于高速公路违章检测的四旋翼无人机平台的设计与实现 [D]. 南京:南京邮电大学,2020.
[5] 曾荻清 . 视觉无人机高速公路违章识别技术的研究及实现[D]. 南京:南京邮电大学,2020.
[6] TANG YANFENG,LIN SONGBIN. Design of Patrol Unmanned Aerial Vehicle Based on MPU9250 [J].IOP Conference Series: Earth and Environmental Science,2021,621(1):012151.
[7] 张雁鹏,高建勇,周志杰 . 基于 VR 设备中 IMU 的头部姿态感知算法研究 [J]. 计算机测量与控制,2020,28(12):139-143.
[8] REDMON J, DIVVALA S, GIRSHICK,et al. You Only Look Once: Unified, Real-Time Object Detection [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas:IEEE,2016:779-788.
[9] HAN BG,LEE J G,LIM KT,et al. Design of a Scalable and Fast YOLO for Edge-Computing Devices [EB/OL].[2022-05-20].https://www.researchgate.net/publication/347209483_Design_of_a_Scalable_and_Fast_YOLO_for_Edge-Computing_Devices.
[10] 毕永 . 高级计算器在灰色预测 GM(1,1) 模型精度检验中的使用 [J]. 世界最新医学信息文摘,2019,19(30):270-271.
[11] 孔雪,王丽,冯益华 . 灰色预测 GM(1,1) 模型应用现状与展望 [J]. 齐鲁工业大学学报,2018,32(6):49-53.
[12] 吴志荣 . 灰色预测 GM(1,1) 模型的应用及改进 [J]. 西部皮革,2016,38(4):235-236.
作者简介:张学峰(2001—),男,汉族,河北张家口人,本科在读,研究方向:嵌入式开发、机器人控制;郁洋(1989—),男,汉族,山东泰安人,讲师,博士,研究方向:机器人智能控制、预测控制理论与应用;任彬(1982—),女,汉族,河北石家庄人,副教授,博士,研究方向:故障诊断及人工智能。