基于DI-PointNet的变电站主设备点云高精度语义分割方法

doi:10.19595/j.cnki.1000-6753.tces.240638

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (1903 KB)
输出: BibTeX | EndNote (RIS)

摘要

在变电站机器人巡检任务中,三维点云数据的高精度语义分割是关键技术之一,有助于机器人理解电力设备、障碍物和其他物体的空间布局。然而,现有的点云分割算法在变电站场景中的效果有限,准确度较低、计算复杂度高,难以实现对变电站主设备点云的准确分割,为了解决这一问题,该文提出了一种基于PointNet++的DI-PointNet算法。首先,采用双层连续变换器模块,增强点云之间的信息交互,有效地聚合长距离上下文,增大网络有效感受野;其次,通过分层键采样策略生成自注意力机制所需的键值,降低算法复杂度;最后,使用倒置残差模块,通过倒置瓶颈设计和残差连接来缓解梯度消失,有效地增加模型的深度,同时降低计算复杂度。同时,该文构建了变电站点云数据集,并对DI-PointNet算法进行了详细的消融实验,并与主流深度学习算法和电力领域典型点云分割算法进行了对比实验。实验验证结果表明,DI-PointNet对变电站主设备点云分割的平均交并比达到了82.5%,相比PointNet++算法提高了2.1%,且总体精度提高了3.4%,达到90.1%。DI-PointNet为智能电力设备巡检和维护提供了有效的解决方案。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	裴少通
	孙海超
	孙志周
	胡晨龙
	祝雨馨

关键词 ：点云语义分割, 双层连续变换器, 分层键采样, 倒置残差, 变电站

Abstract：

In substation robot inspection tasks, high-precision semantic segmentation of 3D point cloud data is one of the key technologies. Traditional point cloud semantic segmentation algorithms have certain limitations, making it difficult to handle complex 3D scenes. Deep learning methods have compensated for the shortcomings of traditional point cloud semantic segmentation algorithms and have become the main method for achieving point cloud semantic segmentation. However, existing point cloud segmentation methods for substations face issues such as high complexity, low accuracy, and gradient vanishing. To address these issues and achieve accurate segmentation of the main equipment point cloud in substations, this paper proposes a high-precision semantic segmentation method for substation main equipment point clouds based on DI-PointNet.
Firstly, on the basis of the PointNet++ network structure, a Double-Layer Consecutive Transformer (DLCTransformer) module is introduced. Key points are sampled through the DLCTransformer to enhance information interaction between point clouds and expand the effective receptive field. Secondly, a hierarchical key sampling strategy is adopted. The point cloud data is divided into the original dense point cloud space and a sparse point cloud space formed after farthest point sampling. These are then divided into multiple non-overlapping 3D windows, ultimately generating key values required for self-attention mechanism calculations, thereby reducing computational complexity, improving the model’s receptive field, and aggregating long-range context to achieve information interaction of substation-associated point clouds. Finally, an Inverted Residual Module (InvResMLP) based on residual connections and inverted bottleneck design is added to the network. This enhances the model’s ability to extract complex structural features from substation point clouds while effectively reducing the gradient vanishing problem, making the algorithm more robust in handling complex substation scenarios and improving the accuracy of semantic segmentation of substation main equipment point clouds.
Additionally, to validate the segmentation effectiveness of the algorithm, this paper uses Avia LiDAR equipment to collect point cloud images of different devices at substations such as the Baobei substation in Baoding City. The original data includes transformers, switchgear, steel towers, insulators, maintenance equipment, and others (mainly vegetation and buildings). To simplify the point cloud data while filtering noise, the original input point cloud is first subjected to grid sampling with a grid size of 0.03m. Data augmentation methods such as Z-axis rotation, scaling, perturbation, and color reduction are employed. The initial window size is set to 0.12m and is doubled after each down-sampling layer. The DI-PointNet is trained using the cross-entropy loss function and Adam optimizer with the following hyperparameters: initial learning rate of 0.001, batch size of 2, and 100 epochs. To ensure the reasonableness and accuracy of the experiments, the comparative algorithms used in this paper are trained using the same hardware platform, environment version, loss function, optimizer, hyperparameters, and training strategies as DI-PointNet.
Through ablation experiments and comparative analysis, the DI-PointNet algorithm proposed in this paper improves the OA value of substation point cloud segmentation by 3.4% compared to before the improvement, while reducing algorithm complexity. The proposed algorithm outperforms other mainstream deep learning algorithms and other point cloud segmentation algorithms in the power sector. The performance of this algorithm is close to the accuracy of manual segmentation and can achieve precise segmentation of substation point clouds.

Key words： Point cloud semantic segmentation double-Layer consecutive transformer hierarchical key sampling inverted residual substation

收稿日期: 2024-04-24

PACS:	TM407
	TP39

基金资助:

国家电网有限公司总部管理科技项目资助（5500-202316168A-1-1-ZN）

通讯作者: 裴少通男,1990年生,副教授,硕士生导师,研究方向为电气设备在线监测及故障诊断。E-mail：peishaotong@ncepu.edu.cn

作者简介: 孙海超男,2000年生,硕士研究生,研究方向为电气设备在线监测及故障诊断。E-mail：931871016@qq.com

引用本文:

裴少通, 孙海超, 孙志周, 胡晨龙, 祝雨馨. 基于DI-PointNet的变电站主设备点云高精度语义分割方法[J]. 电工技术学报, 0, (): 2492938-2492938. Pei Shaotong, Sun Haichao, Sun Zhizhou, Hu Chenlong, Zhu Yuxin. High-Precision Semantic Segmentation of Point Clouds for Primary Equipment in Substations Based on DI-PointNet. Transactions of China Electrotechnical Society, 0, (): 2492938-2492938.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.240638 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/2492938

[1] 王生杰, 马永福, 马国祥, 等. 330 kV GIS外壳异常发热机理与改进措施研究[J/OL]. 高压电器, 2024: 1-11[2024-06-24]. https://kns.cnki.net/kcms/detail/61.1127.tm.20240621.1536.002.html.
Wang Shengjie, Ma Yongfu, Ma Guoxiang, et al. Research on enclosure overheat mechanism and improvement measures of 330 kV GIS[J/OL]. High Voltage Apparatus, 2024: 1-11[2024-06-24]. https://kns.cnki.net/kcms/detail/61.1127.tm.20240621.1536.002.html.
[2] 吴霖, 马飞越, 佃松宜, 等. 气体绝缘开关设备检测维护机器人控制系统设计[J/OL]. 高压电器, 2024: 1-8[2024-06-07]. https://kns.cnki.net/kcms/detail/61.1127.TM.20240606.1446.011.html.
Wu Lin, Ma Feiyue, Dian Songyi, et al. Design of control system for GIS inspection and maintenance robot[J/OL]. High Voltage Apparatus, 2024: 1-8[2024-06-07]. https://kns.cnki.net/kcms/detail/61.1127.TM.20240606.1446.011.html.
[3] 刘栋良, 詹成根, 屈峰, 等. 无人机17kW电机振动噪声分析与巡航转速下尖端噪声优化[J]. 电工技术学报, 2024, 39(6): 1749-1763.
Liu Dongliang, Zhan Chenggen, Qu Feng, et al.Vibration noise analysis and tip noise optimization of unmanned aerial vehicle 17kW motor at cruise speed[J]. Transactions of China Electrotechnical Society, 2024, 39(6): 1749-1763.
[4] Pei Shaotong, Sun Haichao.Structural design and simulation study of intelligent defect elimination equipment for high-voltage transmission line pin defects[J]. IET Generation, Transmission & Distribution, 2023, 17(24): 5366-5377.
[5] Pei Shaotong, Sun Haichao.Design of an intelligent transformer oil sampling system[J]. Electronics Letters, 2023, 59(22): e13038.
[6] 胡晨龙, 裴少通, 刘云鹏, 等. 基于LEE-YOLOv7的输电线路边缘端实时缺陷检测方法[J/OL]. 高电压技术, 2023: 1-14[2024-06-07]. https://doi.org/10.13336/j.1003-6520.hve.20230945.
Hu Chenlong, Pei Shaotong, Liu Yunpeng, et al. Real-time defect detection method for transmission line edge end based on LEE-YOLOv7[J/OL]. High Voltage Engineering, 2023: 1-14[2024-06-07]. https://doi.org/10.13336/j.1003-6520.hve.20230945.
[7] 贾惠彬, 武文瑞, 吴堃, 等. 基于异步整形机制的智能变电站通信队列调度策略[J/OL]. 电工技术学报, 2023: 1-12[2024-06-07]. https://doi.org/10.19595/j.cnki.1000-6753.tces.231236.
Jia Huibin, Wu Wenrui, Wu Kun, et al. Research on communication queue scheduling strategy for intelligent substations based on asynchronous shaping mechanism[J/OL]. Transactions of China Electro-technical Society, 2023: 1-12[2024-06-07]. https://doi.org/10.19595/j.cnki.1000-6753.tces.231236.
[8] 潘玺安, 艾欣, 胡俊杰, 等. 考虑网络安全约束的分布式智能电网边云协同优化调度方法[J/OL]. 电工技术学报, 2023: 1-15[2024-06-07]. https://doi.org/10.19595/j.cnki.1000-6753.tces.231352.
Pan Xian, Ai Xin, Hu Junjie, et al. Network security constrained distributed smart grid edge-cloud collaborative optimization scheduling[J/OL]. Transactions of China Electrotechnical Society, 2023: 1-15[2024-06-07]. https://doi.org/10.19595/j.cnki.1000-6753.tces.231352.
[9] Jiang X Y, Meier U, Bunke H.Fast range image segmentation using high-level segmentation primitives[C]//Proceedings Third IEEE Workshop on Applications of Computer Vision. WACV'96, Sarasota, FL, USA, 1996: 83-88.
[10] Xi Xiaohuan, Wan Yiping, Wang Cheng.Building boundaries extraction from points cloud using an image edge detection method[C]//2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Beijing, China, 2016: 1270-1273.
[11] Besl P J, Jain R C.Segmentation through variable-order surface fitting[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988, 10(2): 167-192.
[12] Schnabel R, Wahl R, Klein R.Efficient RANSAC for point-cloud shape detection[J]. Computer Graphics Forum, 2007, 26(2): 214-226.
[13] 张烨, 李博涛, 尚景浩, 等. 基于多尺度卷积注意力机制的输电线路防振锤缺陷检测[J]. 电工技术学报, 2024, 39(11): 3522-3537.
Zhang Ye, Li Botao, Shang Jinghao, et al.Defect detection of transmission line damper based on multi-scale convolutional attention mechanism[J]. Transactions of China Electrotechnical Society, 2024, 39(11): 3522-3537.
[14] 金亮, 尹振豪, 刘璐, 等. 基于残差U-Net和自注意力Transformer编码器的磁场预测方法[J]. 电工技术学报, 2024, 39(10): 2937-2952.
Jin Liang, Yin Zhenhao, Liu Lu, et al.Magnetic field prediction method based on residual U-net and self-attention transformer encoder[J]. Transactions of China Electrotechnical Society, 2024, 39(10): 2937-2952.
[15] 陈光宇, 袁文辉, 徐晓春, 等. 基于残差图卷积深度网络的电网无功储备需求快速计算方法[J]. 电工技术学报, 2023, 38(17): 4683-4700.
Chen Guangyu, Yuan Wenhui, Xu Xiaochun, et al.Fast calculation method for grid reactive power reserve demand based on residual graph convolutional deep network[J]. Transactions of China Electrotechnical Society, 2023, 38(17): 4683-4700.
[16] Maturana D, Scherer S.VoxNet: a 3D Convolutional Neural Network for real-time object recognition[C]//2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Hamburg, Germany, 2015: 922-928.
[17] Wu Zhirong, Song Shuran, Khosla A, et al.3D ShapeNets: a deep representation for volumetric shapes[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 2015: 1912-1920.
[18] Boulch A, Guerry J, Le Saux B, et al.SnapNet: 3D point cloud semantic labeling with 2D deep segmentation networks[J]. Computers & Graphics, 2018, 71: 189-198.
[19] Charles R Q, Hao Su, Mo Kaichun, et al.PointNet: deep learning on point sets for 3D classification and segmentation[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, 2017: 652-660.
[20] Qi C R, Li Yi, Hao Su, et al.PointNet++: deep hierarchical feature learning on point sets in a metric space[C]//Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, CA, USA, 2017: 5105-5114.
[21] Gao Wei, Zhang Lixia.Semantic segmentation of substation site cloud based on seg-PointNet[J]. Journal of Advanced Computational Intelligence and Intelligent Informatics, 2022, 26(6): 1004-1012.
[22] Yuan Qianjin, Chang Jing, Luo Yong, et al.Automatic cables segmentation from a substation device based on 3D point cloud[J]. Machine Vision and Applications, 2022, 34(1): 9.
[23] Eldar Y, Lindenbaum M, Porat M, et al.The farthest point strategy for progressive image sampling[J]. IEEE Transactions on Image Processing, 1997, 6(9): 1305-1315.
[24] Talvitie J, Renfors M, Lohan E S.Distance-based interpolation and extrapolation methods for RSS-based localization with indoor wireless signals[J]. IEEE Transactions on Vehicular Technology, 2015, 64(4): 1340-1353.
[25] Hu Han, Hou Yongkuo, Ding Yulin, et al.V2PNet: voxel-to-point feature propagation and fusion that improves feature representation for point cloud registration[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023, 16: 5077-5088.
[26] Liu Ze, Lin Yutong, Cao Yue, et al.Swin transformer: hierarchical vision transformer using shifted windows[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada, 2021: 10012-10022.
[27] Lai Xin, Liu Jianhui, Jiang Li, et al.Stratified transformer for 3D point cloud segmentation[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, 2022: 8500-8509.
[28] 杨文杰, 裴少通, 刘云鹏, 等. 基于改进Point Net++的输电线路关键部位点云语义分割研究[J]. 高电压技术, 2024, 50(5): 1943-1953.
Yang Wenjie, Pei Shaotong, Liu Yunpeng, et al.Research on semantic segmentation of point cloud for key parts of transmission lines based on improved PointNet[J]. High Voltage Engineering, 2024, 50(5): 1943-1953.
[29] Wu Zifeng, Shen Chunhua, van den Hengel A. Wider or deeper: revisiting the ResNet model for visual recognition[J]. Pattern Recognition, 2019, 90: 119-133.
[30] Sandler M, Howard A, Zhu Menglong, et al.MobileNetV2: inverted residuals and linear bottlenecks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, 2018: 4510-4520.
[31] Su Hang, Maji S, Kalogerakis E, et al.Multi-view convolutional neural networks for 3D shape recognition[C]//2015 IEEE International Conference on Computer Vision (ICCV), Santiago, Chile, 2015: 945-953.
[32] Wang Yue, Sun Yongbin, Liu Ziwei, et al.Dynamic graph CNN for learning on point clouds[J]. ACM Transactions on Graphics, 2019, 38(5): 1-12.
[33] Wu Wenxuan, Qi Zhongang, Li Fuxin.PointConv: deep convolutional networks on 3D point clouds[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 2019: 9621-9630.
[34] Liu Zhijian, Tang Haotian, Lin Yujun, et al.Point-voxel CNN for efficient 3D deep learning[C]// Proceedings of the 33rd International Conference on Neural Information Processing Systems, Vancouver, BC, Canada, 2019: 965-975.
[35] Su Hang, Jampani V, Sun Deqing, et al.SPLATNet: sparse lattice networks for point cloud processing[C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, 2018: 2530-2539.
[36] Zhao Hengshuang, Jiang Li, Jia Jiaya, et al.Point transformer[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, QC, Canada, 2021: 16239-16248.
[37] Liu Yongcheng, Fan Bin, Xiang Shiming, et al.Relation-shape convolutional neural network for point cloud analysis[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, CA, USA, 2019: 8895-8904.
[38] Xu Yifan, Fan Tianqi, Xu Mingye, et al.SpiderCNN: deep learning on point sets with parameterized convolutional filters[C]//Computer Vision-ECCV 2018, Munich, Germany, 2018: 90-105.
[39] Chen Hui, Wang Tingting, Dai Zuoxiao, et al.Power equipment segmentation of 3D point clouds based on geodesic distance with K-means clustering[C]//2021 6th International Conference on Power and Renewable Energy (ICPRE), Shanghai, China, 2021: 317-321.
[40] Yu Hao, Wang Zhengyang, Zhou Qingjie, et al.Deep-learning-based semantic segmentation approach for point clouds of extra-high-voltage transmission lines[J]. Remote Sensing, 2023, 15(9): 2371.
[41] Zhao Wenbo, Dong Qing, Zuo Zhengli.A point cloud segmentation method for power lines and towers based on a combination of multiscale density features and point-based deep learning[J]. International Journal of Digital Earth, 2023, 16(1): 620-644.
[42] Liu Xiuning, Shuang Feng, Li Yong, et al.SS-IPLE: semantic segmentation of electric power corridor scene and individual power line extraction from UAV-based lidar point cloud[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023, 16: 38-50.
[43] Chen Chi, Jin Ang, Yang Bisheng, et al.DCPLD-Net: a diffusion coupled convolution neural network for real-time power transmission lines detection from UAV-Borne LiDAR data[J]. International Journal of Applied Earth Observation and Geoinformation, 2022, 112: 102960.