[1]吴攀超,王佳硕,王婷婷.基于 FACV-YOLOv10 模型的油田作业人员异常行为检测[J].机械与电子,2025,(10):18-25.
　WU Panchao,WANG Jiashuo,WANG Tingting.Abnormal Behavior Detection of Oilfield Operators Based on FACV-YOLOv10 Model[J].Machinery & Electronics,2025,(10):18-25.
点击复制

基于 FACV-YOLOv10 模型的油田作业人员异常行为检测()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2025年10期
页码:
18-25
栏目:
研究与设计
出版日期:
2025-10-25

文章信息/Info

Title:
Abnormal Behavior Detection of Oilfield Operators Based on FACV-YOLOv10 Model
文章编号:
1001-2257 ( 2025 ) 10-0018-08
作者:
吴攀超; 王佳硕; 王婷婷
东北石油大学电气信息工程学院,黑龙江 大庆 163318
Author(s):
WU Panchao ;  WANG Jiashuo ;  WANG Tingting
( School of Electrical & Information Engineering , Northeast Petroleum University , Daqing 163318 , China )
关键词:
YOLOv10 ; 小目标检测; 损失函数; 模型轻量化; AdditiveBlock-CGLU
Keywords:
YOLOv10 ;  small target detection ;  loss function ;  model lightweighting ;  AdditiveBlock-CGLU
分类号:
TP391.4
文献标志码:
A
摘要:
为解决油田作业现场因操作不规范导致的安全事故频发问题,提出一种改进的 YOLOv10 模型( FACV-YOLOv10 )。针对油田监控中目标尺度变化大、背景复杂等问题,采用 Focaler-CIoU 替代传统损失函数,提升了对小目标的检测能力;结合 AdditiveBlock 和 ConvGLU 优化骨干网络,实现轻量化的同时增强特征提取效果;此外,使用 VoV-GSCSP 和 GSConv 替换 Neck 中的 C2f 和 Conv ,提升了网络对多尺度目标的检测精度,并增强了其在复杂场景中的鲁棒性。在自建油田数据集上进行模型验证, mAP@50 和 mAP@50∶95 分别达到 92.7% 和 67.2% ,在夜间巡检和恶劣天气等复杂工况下仍保持稳定性能,展现出良好的应用价值。
Abstract:
To address the issue of frequent safety accidents caused by improper operations at oilfield sites , this paper proposes an improved YOLOv10 model ( FACV YOLOv10 ) .To address the issues of large target scale changes and complex backgrounds in oilfield monitoring , Focaler CIoU is used instead of traditional loss functions to enhance the detection capability for small targets ; AdditiveBlock and ConvGLU are combined to optimize the backbone network , achieving lightweight while enhancing feature extraction performance ; in addition , replacing C2f and Conv in Neck with VoV-GSCSP and GSConv improves the detection accuracy of the network for multi-scale targets and enhances its robustness in complex scenes.The model is validated on a self-built oilfield dataset , mAP@50 and mAP@50∶95 achieved 92.7% and 67.2% respectively , maintaining stable performance under complex conditions such as night inspections and severe weather , demonstrating good application value.

参考文献/References:

[ 1 ] 张千,梁鸿,童彦淇,等 . 基于深度学习的油田在线视频目标检测[ J ] . 计算机与数字工程, 2024 , 52 ( 3 ): 864-872.

[ 2 ] 张硕羲,任佳亮,陈峰,等 . 基于机器学习方法的员工安全帽佩戴检测[ J ] . 信息技术与信息化, 2025 ( 1 ): 98-101.
[ 3 ] 宋春宁,李寅中 . 面向复杂环境的改进 YOLOv5 安全帽检测算法[ J ] . 电子测量技术, 2025 , 48 ( 7 ): 163-170.
[ 4 ] 雷源毅,朱文球,廖欢 . 复杂场景下的改进 YOLOv8n 安全帽佩戴检测算法[ J ] . 软件工程, 2023 , 26 ( 12 ): 46-51.
[ 5 ] 张全,刘田甜,杨亮,等 . 基于人体关键点的吸烟打电话检测[ J / OL ] . 计算机应用与软件, 1-10 [ 2025-06-04 ] .https : ∥kns cnki net.webvpn.nepu.edu.cn / kcms / detail / 31.1260.TP.20240930.1104.002.html.
[ 6 ] 孙召龙,徐昕,朱云龙,等 . 基于 YOLOv5 的油田作业现场吸烟检测方法[ J ] . 系统仿真技术, 2021 , 17 ( 2 ): 89-93.
[ 7 ] 陈学台,欧郁强,李敏,等 . 基于 YOLOv5 及深度神经网络的电力智慧安监算法[ J ] . 机械与电子,2024 , 42( 12 ): 37-42 , 48.
[ 8 ] 曹萌迪,杨梦凡,王留毅,等 . 基于改进 YOLOv8 的复杂场景跌倒检测算法[ J ] . 现代信息科技, 2025 , 9 ( 5 ):66-71.
[ 9 ] 宋杰,徐慧英,朱信忠,等 . 基于 YOLOv8 改进的跌倒检测算法: OEF-YOLO [ J ] . 计算机工程, 2025 , 51( 7 ): 121-139.
[ 10 ] 李文杰 . 油田生产安全事故统计、原因分析及控制措施探讨[ J ] . 化工安全与环境, 2023 , 36 ( 9 ): 14-18.
[ 11 ] CHU B Q , SHAO R H , FANG Y , et al.Weed detection method based on improved YOLOv8 with neck slim [ C ] ∥2023 China Automation Congress ( CAC ), New York : IEEE , 2023 : 9378-9382.
[ 12 ] HAO S , SUN H B , MA X , et al.Small target fault detection method based on parallel attention [ C ] ∥2024 2nd International Conference on Algorithm , Image Processing and Machine Vision ( AIPMV ), New York : IEEE , 2024 : 258-261.
[ 13 ] PAN X , GE C , LU R , et al.On the integration of self attention and convolution [ C ] ∥Proceedings of the IEEE / CVF Conference on Computer Vision and Pattern Recognition , 2022 : 815-825.
[ 14 ] PAN X R , GE C J , LU R , et al.On the integration of self-attention and convolution [ C ] ∥Proceedings of the IEEE / CVF Conference on Computer Vision and Pattern Recognition , 2022 : 815-825.
[ 15 ] SHI D.Transnext : robust foveal visual perception for vision transformers [ C ] ∥Proceedings of the IEEE / CVF Conference on Computer Vision and Pattern Recognition , 2024 : 17773-17783.
[ 16 ] XU B , GAO B , LI Y H.Improved small object detection algorithm based on YOLOv5 [ J ] .IEEE Intelligent systems , 2024 , 39 ( 5 ): 57-65.
[ 17 ] 邹磊,苏家仪,黎恒,等 . 改进 YOLOv5 的密集小目标安全帽检测研究[ J ] . 物联网技术, 2025 , 15 ( 2 ): 3-8.
[ 18 ] 韩博,张婧婧,鲁子翱 .FEV-YOLOv8n :轻量化安全帽佩戴 检 测 方 法 [ J ] . 计算机测量与控制,2025 , 33( 1 ): 69-77 , 84.

相似文献/References:

[1]潘学华,梁炜皓,李金瑾,等.面向绝缘子微小缺陷的 YOLOv8s 的改进算法研究[J].机械与电子,2025,(12):10.
　PAN Xuehua,LIANG Weihao,LI Jinjin,et al.Research on Improved YOLOv8s Algorithm for Micro-defect Detection in Insulators[J].Machinery & Electronics,2025,(10):10.

备注/Memo

备注/Memo:
收稿日期: 2025-05-19
基金项目:国家自然科学基金资助项目( 52474036 )
作者简介:吴攀超 ( 1981- ),男,黑龙江大庆人,博士,讲师,研究方向为图像处理与计算机视觉;王佳硕 ( 1999- ),男,山东菏泽人,硕士研究生,研究方向为模式识别应用与图像处理,通信作者, E-mail : 83768475@qq.com ;王婷婷 ( 1982- ),女,黑龙江大庆人,博士,教授,博士研究生导师,研究方向为油气田岩石脆性分析。
更新日期/Last Update: 2025-11-12