针对可再生能源施工现场（风电场/太阳能电站）行人、车辆与重型设备碰撞事故频发的问题,该研究通过融合机器视觉与深度学习技术,构建多模态监控框架,开发了一种高精度实时交通监控系统,以提升动态风险感知能力。该文基于改进YOLO v8模型实现目标检测,结合双目立体视觉与卡尔曼滤波算法完成三维定位与轨迹预测,最后在灵璧县风电项目部署系统原型,通过实时视频流验证其性能。实测数据表明,系统在施工场景下的行人检测准确率达96.2%,车辆识别精度为98.5%,可同步追踪运输车动态参数（距离误差±3.0 m,速度误差±1.2 km/h）,相较传统监控方式响应速度提升40%。该研究为可再生能源施工现场的智能化安全管控提供了通用解决方案,助力智慧工地与智能交通系统的深度融合。
Abstract
To address the frequent collision accidents among pedestrians, vehicles, and heavy equipment at renewable energy construction sites (wind farms/solar power stations）, this study develops a high-precision real-time traffic monitoring system by integrating machine vision and deep learning technologies within a multi-modal monitoring framework to enhance dynamic risk perception. An improved YOLO v8 model is employed for object detection, combined with binocular stereo vision and Kalman filtering algorithms for three-dimensional positioning and trajectory prediction. A prototype system is deployed at the Lingbi County wind power project, with its performance validated through real-time video streams. Experimental results demonstrate that the system achieves pedestrian detection accuracy of 96.2% and vehicle recognition accuracy of 98.5% under construction scenarios, while simultaneously tracking transport vehicle dynamic parameters (distance error ±3.0 m, speed error ±1.2 km/h）, and improves response speed by 40% compared to conventional monitoring approaches. This study provides a universal solution for intelligent safety management at renewable energy construction sites, facilitating the deep integration of smart construction sites and intelligent transportation systems.
关键词
机器视觉 /
可再生能源 /
叶片 /
动态风险感知 /
三维空间定位
Key words
machine vision /
renewable energy /
blades /
dynamic risk perception /
three-dimensional spatial positioning
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 石腾, 许波峰, 陈鹏, 等. 基于机器视觉的风电机组叶片多类型损伤检测方法研究[J]. 太阳能学报, 2024, 45(6): 487-494.
Shi T, Xu B F, Chen P, et al.Study on multi-type damage detection method for wind turbine blades based on machine vision technology[J]. Acta Energiae Solaris Sinica, 2024, 45(6): 487-494.
[2] 邬伟骏, 吴江波, 周强, 等. 基于贝叶斯推理的风电机组风轮偏航协同智能控制方法[J]. 可再生能源, 2024, 42(9): 1205-1210.
Wu W J, Wu J B, Zhou Q, et al.Collaborative intelligent control method for wind turbine rotor yaw based on Bayesian inference[J]. Renewable Energy Resources, 2024, 42(9): 1205-1210.
[3] 贾晓红, 石岚, 郝玉珠. 基于风电场片区风速的人工智能插补方法对比[J]. 太阳能学报, 2025, 46(1): 168-175.
Jia X H, Shi L, Hao Y Z.Comparison of artificial intelligence interpolation methods based on wind speed of wind farm divided into wind turbine groups[J]. Acta Energiae Solaris Sinica, 2025, 46(1): 168-175.
[4] Musial W, Spitsen P, Duffy P, et al.Offshore wind market report: 2021 edition[R]. U.S. Department of Energy, 2021.
[5] International Renewable Energy Agency}. Renewable capacity statistics 2023[R]. Abu Dhabi: International Renewable Energy Agency, 2023.
[6] 戴丽萍, 张泽能, 丛龙福, 等. 基于神经网络的风力机叶片三维失速模型研究[J]. 太阳能学报, 2025, 46(1): 53-59.
Dai L P, Zhang Z N, Cong L F, et al.Study on three-dimensional stall model of wind turbine blades based on neural networks[J]. Acta Energiae Solaris Sinica, 2025, 46(1): 53-59.
[7] 全球能源互联网发展合作组织. 全球清洁能源开发与投资研究[M]. 北京: 中国电力出版社, 2020.
GEIDCO. Research on global clean energy development and investment[M]. Beijing: China Electric Power Press, 2020.
[8] 邹龙洲, 王文韫, 郭迎福, 等. 基于改进YOLOv4的风力机叶片损伤检测方法[J]. 太阳能学报, 2024, 45(7): 718-723.Zou L Z, Wang W Y, Guo Y F, et al. Wind turbine blade damage detection method based on improved YOLOv4[J]. Acta Energiae Solaris Sinica, 2024, 45(7): 718-723.
[9] 徐一鸣, 张娟, 刘成成, 等. 迁移学习模式下基于GoogLeNet网络的风电机组视觉检测[J]. 计算机科学, 2019, 46(5): 260-265.
Xu Y M, Zhang J, Liu C C, et al.Wind turbine visual inspection based on GoogLeNet network in transfer learning mode[J]. Computer Science, 2019, 46(5): 260-265.
[10] 李奕炜, 骆立实, 赵波, 等. 基于YOLO v8算法的变电站视频监控多目标智能跟踪方法[J]. 电信科学, 2025, 41(3): 179-189.
Li Y W, Luo L S, Zhao B, et al.Multi-target intelligent tracking method for substation video surveillance based on YOLO v8 algorithm[J]. Telecommunications Science, 2025, 41(3): 179-189.
[11] 王文亮, 解锡伟, 徐玉龙, 等. 基于物联网的风力发电机塔筒倾斜监测系统设计[J]. 工业仪表与自动化装置, 2024(5): 40-43, 64.
Wang W L, Xie X W, Xu Y L, et al.Design of a wind turbine tower tilt monitoring system based on the Internet of Things[J]. Industrial Instrumentation & Automation, 2024(5): 40-43, 64.
[12] Liu X T, Hu Y T, Ji H T, et al.A deep learning method for ship detection and traffic monitoring in an offshore wind farm area[J]. Journal of Marine Science and Engineering, 2023, 11(7): 1259.
[13] 王文举, 唐邦, 顾泽骅, 等. 深度学习的多视角三维重建技术综述[J]. 计算机工程与应用, 2025, 61(6): 22-35.
Wang W J, Tang B, Gu Z H, et al.Overview of multi-view 3D reconstruction techniques in deep learning[J]. Computer Engineering and Applications, 2025, 61(6): 22-35.
基金
国家重点研发计划（2021YFC3101800）; 国家自然科学基金面上项目（42476213）; 华东院科技项目（KY2023-JD-03-06）
PDF(3395 KB)
