Point cloud method for identifying the center pose of the pin shaft end face in the automated dismantling of oil derricks
DOI:
https://doi.org/10.54097/dqx9zy05Keywords:
PointNet++, RANSAC algorithm, Normal vector coordinate system, Point cloud segmentation slicingAbstract
The derrick of oil drilling rigs is constructed by pin shafts connecting steel frames. Traditional manual disassembly methods pose safety risks and exhibit low efficiency. To achieve automated disassembly, precise identification of the end-face center poses (position and orientation) of pin shafts is essential. However, conventional registration and localization methods suffer from insufficient positioning accuracy due to incomplete point cloud acquisition of pin shafts, while existing deep learning approaches struggle to accurately calculate local object poses. To solve the above problem, this paper proposes a method that combines deep learning and traditional point cloud processing for the precise positioning of the center of the pin end face on the derrick of an oil rig. Specifically, a point cloud segmentation network based on PointNet++ is used in the deep learning part for effectively extracting the point cloud data of the pin shaft and its surrounding structures from the complex scene. In the point cloud processing part, an innovative method based on RANSAC plane fitting is proposed, in which a local coordinate system is established for the fitted plane, the attitude information is calculated, and the plane is sliced for extracting the center of the pin endface to realize the spatial localization of the pin endface.The experimental results show that the positioning error of the proposed method along the X/Y/Z axes is less than 0.5 mm, the directional error is less than 1°, and the accuracy is as high as 96%, which is better than that of the traditional method in terms of precision and accuracy. This study provides a high-precision pin shaft localization solution for automated derrick disassembly, effectively mitigating safety risks and enhancing operational efficiency.
Downloads
References
[1] Ji, X., Zhang, X., & Hu, H. (2021). Point cloud segmentation for complex microsurfaces based on feature line fitting. Multimedia Tools and Applications, 80, 4553-4578.
[2] Rusu, R. B., Blodow, N., & Beetz, M. (2009, May). Fast point feature histograms (FPFH) for 3D registration. In 2009 IEEE international conference on robotics and automation (pp. 3212-3217). IEEE.
[3] Zhong, J., Dujovny, M., Park, H. K., Perez, E., Perlin, A. R., & Diaz, F. G. (2003). Advances in ICP monitoring techniques. Neurological research, 25(4), 339-350.
[4] Chum, O., Matas, J., & Kittler, J. (2003, September). Locally optimized RANSAC. In Joint pattern recognition symposium (pp. 236-243). Berlin, Heidelberg: Springer Berlin Heidelberg.
[5] Varia, N., Dokania, A., & Senthilnath, J. (2018, November). DeepExt: A convolution neural network for road extraction using RGB images captured by UAV. In 2018 IEEE Symposium Series on Computational Intelligence (SSCI) (pp. 1890-1895). IEEE.
[6] Liu, Z., Shi, S., Duan, Q., Zhang, W., & Zhao, P. (2019). Salient object detection for RGB-D image by single stream recurrent convolution neural network. Neurocomputing, 363, 46-57.
[7] Hu, S., Hou, R., Ming, L., Meifang, S., & Chen, P. (2023). A hyperspectral image reconstruction algorithm based on RGB image using multi-scale atrous residual convolution network. Frontiers in Marine Science, 9, 1006452.
[8] Qi, C. R., Su, H., Mo, K., & Guibas, L. J. (2017). Pointnet: Deep learning on point sets for 3d classification and segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 652-660).
[9] Ge, L., Ren, Z., & Yuan, J. (2018). Point-to-point regression pointnet for 3d hand pose estimation. In Proceedings of the European conference on computer vision (ECCV) (pp. 475-491).
[10] Aoki, Y., Goforth, H., Srivatsan, R. A., & Lucey, S. (2019). Pointnetlk: Robust & efficient point cloud registration using pointnet. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 7163-7172).
[11] Qi, C. R., Yi, L., Su, H., & Guibas, L. J. (2017). Pointnet++: Deep hierarchical feature learning on point sets in a metric space. Advances in neural information processing systems, 30.
[12] He, S., Wu, Q. H., & Saunders, J. R. (2006). A group search optimizer for neural network training. In Computational Science and Its Applications-ICCSA 2006: International Conference, Glasgow, UK, May 8-11, 2006, Proceedings, Part III 6 (pp. 934-943). Springer Berlin Heidelberg.
[13] Bello, I., Zoph, B., Vasudevan, V., & Le, Q. V. (2017, July). Neural optimizer search with reinforcement learning. In International Conference on Machine Learning (pp. 459-468). PMLR.
[14] Nugroho, B., Suciati, N., & Fatichah, C. (2024). Analysis of Optimization Techniques in 6D Pose Estimation Approaches using RGB Images on Multiple Objects with Occlusion. Procedia Computer Science, 234, 1689-1696.
[15] Hoque, S., Xu, S., Maiti, A., Wei, Y., & Arafat, M. Y. (2023). Deep learning for 6D pose estimation of objects—A case study for autonomous driving. Expert Systems with Applications, 223, 119838.
[16] Le, T. T., Le, T. S., Chen, Y. R., Vidal, J., & Lin, C. Y. (2021). 6D pose estimation with combined deep learning and 3D vision techniques for a fast and accurate object gras**. Robotics and Autonomous Systems, 141, 103775.
[17] Wu, C., Fu, H., Kaiser, J. P., Barczak, E. T., Pfrommer, J., Lanza, G., Heizmann, M., & Beyerer, J. (2024). 6d pose estimation on point cloud data through prior knowledge integration: A case study in autonomous disassembly. Procedia CIRP, 122, 193-198.
[18] Govi, E., Sapienza, D., Toscani, S., Cotti, I., Franchini, G., & Bertogna, M. (2024). Addressing challenges in industrial pick and place: A deep learning-based 6 Degrees-of-Freedom pose estimation solution. Computers in Industry, 161, 104130.
[19] Zhang, H. B., Hong, J. X., Liu, J. H., Lei, Q., & Du, J. X. (2025). Images, normal maps and point clouds fusion decoder for 6D pose estimation. Information Fusion, 117, 102907.
[20] Zhuang, C., Wang, H., & Ding, H. (2024). AttentionVote: A coarse-to-fine voting network of anchor-free 6D pose estimation on point cloud for robotic bin-picking application. Robotics and Computer-Integrated Manufacturing, 86, 102671.
[21] Chen, W., Jia, X., Chang, H. J., Duan, J., & Leonardis, A. (2020). G2l-net: Global to local network for real-time 6d pose estimation with embedding vector features. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 4233-4242).
[22] Han, K., **ao, A., Wu, E., Guo, J., Xu, C., & Wang, Y. (2021). Transformer in transformer. Advances in neural information processing systems, 34, 15908-15919.
[23] Hong, J. X., Zhang, H. B., Liu, J. H., Lei, Q., Yang, L. J., & Du, J. X. (2024). A transformer-based multi-modal fusion network for 6d pose estimation. Information Fusion, 105, 102227.
[24] Periyasamy, A. S., Amini, A., Tsaturyan, V., & Behnke, S. (2023). YOLOPose V2: Understanding and improving transformer-based 6D pose estimation. Robotics and Autonomous Systems, 168, 104490.
[25] Pan, J., & Zhang, X. (2025, February). Research on 3D point cloud registration algorithm based on FPFH and ColorICP. In International Conference on Optical and Photonic Engineering (icOPEN 2024) (Vol. 13509, pp. 33-38). SPIE.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 International Journal of Advanced Engineering and Technology Research
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
