Trajectory Identification and Working Condition Analysis of Workover Rig System Based on Pose Estimation

Wang Lijun; Zhang Dejun; Li Geng; Zhang Hui; Pu Hongbin; Chuan Ping

doi:10.12473/CPM.202507026

您当前的位置：

首页 >

文章列表页 >

Trajectory Identification and Working Condition Analysis of Workover Rig System Based on Pose Estimation

OIL AND GAS ARTIFICIAL INTELLIGENCE | 更新时间：2026-11-04

- Trajectory Identification and Working Condition Analysis of Workover Rig System Based on Pose Estimation
- China Petroleum Machinery Vol. 54, Issue 10, Pages: 1-11(2026)
- 作者机构：
  
  1.中国石油新疆油田分公司数智技术公司
  2.中国石油新疆油田分公司工程技术部
- 作者简介：
- 基金信息：
- DOI：10.12473/CPM.202507026
  CLC： TE928/319
- Received：07 July 2025，
  
  Revised：2026-07-19，
  
  Accepted：08 September 2025，
  
  Published：25 October 2026
- 稿件说明：
移动端阅览
王利君，张德君，李赓，等.基于定位姿态估计的修井机系统轨迹识别与工况分析［J］.石油机械，2026，54（10）：1-11.

Wang Lijun，Zhang Dejun，Li Geng，et al.Trajectory identification and working condition analysis of workover rig system based on pose estimation［J］.China Petroleum Machinery，2026，54（10）：1-11.
王利君，张德君，李赓，等.基于定位姿态估计的修井机系统轨迹识别与工况分析［J］.石油机械，2026，54（10）：1-11. DOI： 10.12473/CPM.202507026.

Wang Lijun，Zhang Dejun，Li Geng，et al.Trajectory identification and working condition analysis of workover rig system based on pose estimation［J］.China Petroleum Machinery，2026，54（10）：1-11. DOI： 10.12473/CPM.202507026.

摘要

在油气增储上产及高效开发需求驱动下，油气田修井作业现场亟须提升智能化监测水平。传统人工与半自动化作业模式下应用的计算机视觉监测技术难以实现修井机关键构件高精度实时定位，轨迹跟踪与

工况分析相互独立，大钩摆动偏移、载荷异常等安全风险难以自动辨识。针对该工程难题，提出了一套集成快速姿态估计与精细轨迹识别和工况分析的系统框架，利用基于姿态引导的非局部注意力机制改进姿态估计YOLO-Pose方法，设计全局粗定位与局部特征细化相结合的分层局部化姿态估计策略，该方法面向游车大钩结构定制5个专用关键点，并配套井口单点标注方案，同步完成井架、大钩双目标联合检测，依托关键点像素距离建立图像与真实物理空间尺度换算关系；引入姿态引导轻量化通道注意力机制，以设备实时姿态信息驱动特征权重动态分配，抑制井场遮挡、光影带来的无效噪声，实现轨迹跟踪与工况分析深度耦合；基于连续帧关键点坐标建立空间坐标转换模型，结合牛顿动力学模型求解大钩位移、速度、加速度等时序运动参数，构建大钩-井口垂向夹角量化判定准则，完成动态受力解析与摆动危险工况自动预警。研究结果表明：模型经300轮迭代训练后趋于稳定收敛，目标与关键点检测精度达99.91%、召回率100%，边界框与关键点

A@0.5

均为99.5%，关键点平均精度均值

mA@0.5：0.95

可达98.21%；现场实测可完整复现大钩起下管运动轨迹，运动参数曲线可清晰区分启停、匀速、静置等作业阶段，设定5°安全阈值能够精准识别大钩异常摆动状态。该方法可在复杂井场环境下实现修井机关键部件高精度定位、连续轨迹稳定跟踪与工况量化诊断，提前预判作业安全隐患。研究结果可为复杂井场作业提供智能化解决方案。

Abstract

Driven by the demands for increased oil and gas reserves/production and efficient development， there is an urgent need to enhance the intelligent monitoring level at field workover sites. Traditional computer-vision-based monitoring techniques， mainly applied in manual and semi-automated operation modes， are unable to achieve high-precision real-time positioning of key components of workover rigs. Trajectory tracking and working condition analysis are performed independently， and safety risks such as hook swing deviation and load anomalies are difficult to identify automatically. To address this engineering challenge， a systematic framework integrating rapid pose estimation， fine trajectory identification， and working condition analysis was proposed. The YOLO-Pose method for pose estimation was improved using a pose-guided non-local attention mechanism. A hierarchical localized pose estimation strategy combining global rough positioning and local feature refinement was designed， with five keypoints customized for the block-hook assembly. In conjunction with the wellhead single-point annotation scheme， the derrick and hook were detected synchronously. Based on the pixel distance of keypoints， the conversion relationship between image scale and real physical space scale was establ

ished. A pose-guided lightweight channel attention mechanism was introduced to dynamically allocate feature weights driven by real-time equipment pose information， suppressing invalid noise caused by wellsite occlusion and lighting variations， and achieving deep coupling of trajectory tracking and working condition analysis. A spatial coordinate conversion model was constructed based on continuous-frame keypoint coordinates， and combined with the Newtonian dynamics model， the time-series kinematic parameters of hook displacement， velocity， and acceleration were solved. A quantification criterion for the vertical angle between the hook and the wellhead was defined. Ultimately， dynamic force analysis and automatic warning of dangerous swing conditions were completed. The results show that the model tends to converge stably after 300 training epochs， achieving a detection accuracy of 99.91% and a recall of 100% for targets and keypoints， with both bounding-box and keypoint

A@0.5

reaching 99.5%， and keypoint

mA@0.5：0.95

reaching 98.21%. Field measurements can fully reproduce the hook motion trajectory during pipe running and pulling， and the kinematic parameter curves can clearly reflect operational stages such as start-stop， uniform motion， and idling. A safety threshold of 5° enables accurate identification of abnormal hook swing states. This method can achieve high-precision positioning of key components of workover rig， stable tracking of continuous trajectory， and quantitative diagnosis of working condition in complex wellsite environments， providing early warning of operational safety hazards. The research offers an intelligent solution for complex wellsite operations.

关键词

Keywords

references

Sircar A ， Yadav K ， Rayavarapu K ， et al . Application of machine learning and artificial intelligence in oil and gas industry ［J］. Petroleum Research ， 2021 ， 6 （ 4 ）： 379 - 391 .

Zhang Kai ， Song Zewen ， Xia Xiaolong ， et al . Enhancing safety and workover efficiency through real-time scale-adaptive video detection of critical objects at the wellsite ［J］. SPE Journal ， 2024 ， 29 （ 8 ）： 3982 - 3999 .

张菲菲，王茜，王学迎，等．油气井工程多源多模态数据融合技术与展望［J］．天然气工业， 2024 ， 44 （ 9 ）： 152 - 166 ．

Zhang Feifei ， Wang Qian ， Wang Xueying ， et al . Multi-source and multi-modal data fusion technology and its prospect in oil and gas well engineering ［J］. Natural Gas Industry ， 2024 ， 44 （ 9 ）： 152 - 166 .

王晓宇，闫继宏，秦勇，等．基于扩展卡尔曼滤波的两轮机器人姿态估计［J］．哈尔滨工业大学学报， 2007 ， 39 （ 12 ）： 1920 - 1924 ．

Wang Xiaoyu ， Yan Jihong ， Qin Yong ， et al . Attitude estimation based on extended Kalman filter for a two-wheeled robot ［J］. Journal of Harbin Institute of Technology ， 2007 ， 39 （ 12 ）： 1920 - 1924 .

Zhang Kai ， Xia Xiaolong ， Song Zewen ， et al . Trajectory recognition and working condition analysis of rod pumping systems based on pose estimation method with heatmap-free joint detection ［J］. SPE Journal ， 2024 ， 29 （ 10 ）： 5521 - 5553 .

Kocabas M ， Karagoz S ， Akbas E . MultiPoseNet： fast multi-person pose estimation using pose residual network ［C］∥ Computer Vision-ECCV 2018 ， 2018 ： 437 - 453 .

He Kaiming ， Gkioxari G ， Dollár P ， et al . Mask R-CNN ［C］∥ 2017 IEEE International Conference on Computer Vision （ICCV）， 2017 ： 2980 - 2988 .

Newell A ， Yang Kaiyu ， Deng Jia . Stacked hourglass networks for human pose estimation ［C］∥ Computer Vision–ECCV 2016 ， 2016 ： 483 - 499 .

Cao Zhe ， Simon T ， Wei S E ， et al . Realtime multi-person 2D pose estimation using part affinity fields ［C］∥ 2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）， 2017 ： 1302 - 1310 .

Pishchulin L ， Insafutdinov E ， Tang Siyu ， et al . DeepCut： joint subset partition and labeling for multi person pose estimation ［C］∥ 2016 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）， 2016 ： 4929 - 4937 .

Ren Shaoqing ， He Kaiming ， Girshick R ， et al . Faster R-CNN： towards real-time object detection with region proposal networks ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence ， 2017 ， 39 （ 6 ）： 1137 - 1149 .

Toshev A ， Szegedy C . DeepPose： human pose estimation via deep neural networks ［C］∥ 2014 IEEE Conference on Computer Vision and Pattern Recognition ， 2014 ： 1653 - 1660 .

Tompson J ， Goroshin R ， Jain A ， et al . Efficient object localization using Convolutional Networks ［C］∥ 2015 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）， 2015 ： 648 - 656 .

Maji D ， Nagori S ， Mathew M ， et al . YOLO-Pose： enhancing YOLO for multi person pose estimation using object keypoint similarity loss ［C］∥ 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops （CVPRW）， 2022 ： 2636 - 2645 .

Lin T Y ， Maire M ， Belongie S ， et al . Microsoft COCO： common objects in context ［C］∥ Computer Vision–ECCV 2014 ， 2014 ： 740 - 755 .

Redmon J ， Divvala S ， Girshick R ， et al . You only look once： unified， real-time object detection ［C］∥ Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition ， 2016 ： 778 - 779 .

Chen Yilun ， Wang Zhicheng ， Peng Yuxiang ， et al . Cascaded pyramid network for multi-person pose estimation ［C］∥ 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition ， 2018 ： 7103 - 7112 .

Larochelle H ， Hinton G E . Learning to combine foveal glimpses with a third-order Boltzmann machine ［C］∥ Proceedings of the 24th International Conference on Neural Information Processing Systems-Volume 1 ， 2010 ： 1243 - 1251 .

Wang Qilong ， Wu Banggu ， Zhu Pengfei ， et al . ECA-Net： efficient channel attention for deep convolutional neural networks ［C］∥ Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition ， 2020 ： 11534 - 11542 .

Hu Jie ， Shen Li ， Sun Gang . Squeeze-and-excitation networks ［C］∥ 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition ， 2018 ： 7132 - 7141 .

Liu Shu ， Qi Lu ， Qin Haifang ， et al . Path aggregation network for instance segmentation ［C］∥ 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition ， 2018 ： 8759 - 8768 .

Yang Wei ， Li Shuang ， Ouyang Wanli ， et al . Learning feature Pyramids for human pose estimation ［C］∥ 2017 IEEE International Conference on Computer Vision （ICCV）， 2017 ： 1290 - 1299 .

Munea T L ， Jembre Y Z ， Weldegebriel H T ， et al . The progress of human pose estimation： a survey and taxonomy of models applied in 2D human pose estimation ［J］. IEEE Access ， 2020 ， 8 ： 133330 - 133348 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Localization of Hazardous Areas during Rig Operations Based on an Improved YOLOv8 Model

Downhole Vibration Mode Identification Based on E-ASW-LA Model

Interpretable Real-Time Prediction of Drilling Parameters Based on Improved Sequential Network

Implementation of Drill Pipe Joint Positioning Under Binocular Vision

Design and Motion Characteristics of In-Situ Flexible Manipulator Robot for Mud Tank Cleaning

Related Author

Ma Mingjun

Chu Shengli

Xu Zihan

Hu Jinqiu

Liu Xin

Li Zhongwei

Yu Zihan

Li De

Related Institution

College of Safety and Ocean Engineering， China University of Petroleum

Key Laboratory of Oil & Gas Production Safety and Emergency Technology， Ministry of Emergency Management

CNPC Research Institute of Safety & Environment Technology

Key Laboratory of Oil & Gas Storage and Transportation Safety Risk Prevention， Ministry of Emergency Management

Qingdao Institute of Software / College of Computer Science and Technology， China University of Petroleum （East China）

AI问答

⁰