基于改进KPConv的电力场景点云鲁棒语义分割方法研究*

doi:10.7523/j.ucas.2025.064

中国科学院大学学报

基于改进KPConv的电力场景点云鲁棒语义分割方法研究^*

刘海波¹, 张苏¹, 荣经国¹, 严斐然², 高原², 杨洲², 聂胜^2,*

1.国网经济技术研究院有限公司, 北京 102209;
2.中国科学院空天信息创新研究院, 北京 100094

收稿日期:2025-09-02 修回日期:2025-12-29 发布日期:2025-12-31
通讯作者: †E-mail: niesheng@aircas.ac.cn
基金资助:
国家电网有限公司总部科技项目(1400-202456289A-1-1-ZN)资助

Robust Semantic Segmentation of Transmission Corridor Point Clouds Based on Improved KPConv

LIU Haibo¹, ZHANG Su¹, RONG Jingguo¹, YAN Feiyan², GAO Yuan², YANG Zhou², NIE Sheng²

1. State Grid Economic and Technological Research Institute Co., LTD, 102209, China;
2. Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China

Received:2025-09-02 Revised:2025-12-29 Published:2025-12-31
Supported by:
; This work was supported by the Technical Foundation of State Grid Corporation of China under Grant (52230025000F-191-ZN).

摘要/Abstract

摘要： 针对当前电力场景点云分割模型因噪声干扰强、线状结构识别难、类别不平衡等因素,导致分割性能不稳定的问题。本文提出一种面向电力场景的鲁棒性点云语义分割方法,结合KNN噪声抑制、椭球结构感知卷积与Focal Loss损失函数三大模块,有效提升了模型的抗干扰性、结构识别完整性和对少数类的判别能力。在高压输电走廊实测数据集(500 kV)和ECLAIR公开数据集上开展对比实验,结果显示本文方法在总体准确率(OA)和平均交并比(mIoU)方面均优于主流方法KPConv与RandLA-Net,其中针对500 kV 数据集,新提出算法的mIoU较RandLA-Net算法提升9.5%,较KPConv算法提升5.9%,显著提升了线状部件与小类别目标的识别效果,展现出良好的准确性、稳定性与泛化能力。

关键词: 电力场景, 语义分割, 鲁棒性, 机载激光点云, 类别不平衡

Abstract: Semantic segmentation of point clouds in transmission corridor remains a challenging task due to strong noise interference, the difficulty of accurately identifying linear structures, and severe class imbalance, all of which result in unstable segmentation performance. To address these issues, this paper proposes a robust semantic segmentation method for transmission corridor point clouds. The proposed approach integrates three key modules: k-nearest neighbor (KNN)-based noise suppression, ellipsoidal structure-aware convolution, and the Focal Loss function. These modules collaboratively enhance the model's robustness to noise, structural completeness in feature representation, and discrimination capability for minority classes. Extensive experiments were conducted on a real-world 500 kV high-voltage transmission corridor dataset and the publicly available ECLAIR dataset. Experimental results demonstrate that the proposed method achieves superior performance compared to state-of-the-art approaches such as KPConv and RandLA-Net in terms of overall accuracy (OA) and mean intersection over union (mIoU). Specifically, on the 500 kV dataset, the proposed method improves the mIoU by 9.5% over RandLA-Net and by 5.9% over KPConv. The method significantly enhances the recognition of linear components and small-category objects, exhibiting excellent accuracy, stability, and generalization capability. This work was supported by the Technical Foundation of State Grid Corporation of China under Grant (52230025000F-191-ZN).

Key words: transmission corridor, semantic segmentation, robustness, airborne lidar point cloud, class imbalance

中图分类号:

刘海波, 张苏, 荣经国, 严斐然, 高原, 杨洲, 聂胜. 基于改进KPConv的电力场景点云鲁棒语义分割方法研究^*[J]. 中国科学院大学学报, DOI: 10.7523/j.ucas.2025.064.

LIU Haibo, ZHANG Su, RONG Jingguo, YAN Feiyan, GAO Yuan, YANG Zhou, NIE Sheng. Robust Semantic Segmentation of Transmission Corridor Point Clouds Based on Improved KPConv[J]. Journal of University of Chinese Academy of Sciences, DOI: 10.7523/j.ucas.2025.064.

参考文献

[1] He T, Zeng Y H, Hu Z L.Research of multi-rotor UAVs detailed autonomous inspection technology of transmission lines based on route planning[J]. IEEE Access, 2019, 7: 114955-114965. DOI: 10.1109/ACCESS.2019.2935208.
[2] 芦肇基,沈艳霞,谭永强. 基于多结构融合WGAN的模糊绝缘子图像复原方法研究[J]. 电力系统保护与控制,2024,52(22):166-175. DOI:10.19783/j.cnki.pspc.240115.
[3] Xu C, Li Q W, Zhou Q K, et al.Power line-guided automatic electric transmission line inspection system[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 3512118. DOI: 10.1109/TIM.2022.3169555.
[4] 麻卫峰,王成,王金亮,等. 点云数据温度变化环境下输电线形态模拟[J]. 电力系统保护与控制,2021,49(9):90-96. DOI:10.19783/j.cnki.pspc.200873.
[5] 梁磊,习晓环,王成,等. 基于B/S架构的激光雷达电力巡线可视化管理与分析系统[J]. 中国科学院大学学报,2022,39(2):201-207. DOI:10.7523/j.ucas.2020.0044.
[6] Liu X N, Shuang F, Li Y, 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. DOI: 10.1109/JSTARS.2023.3289599.
[7] Xie Y X, Tian J J, Zhu X X.Linking points with labels in 3D: a review of point cloud semantic segmentation[J]. IEEE Geoscience and Remote Sensing Magazine, 2020, 8(4): 38-59. DOI: 10.1109/MGRS.2019.2937630.
[8] Yermo M, Laso R, Lorenzo O G, et al.Powerline detection and characterization in general-purpose airborne LiDAR surveys[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024, 17: 10137-10157. DOI: 10.1109/JSTARS.2024.3396522.
[9] Chen M L, Li J Y, Pan J P, et al.Insulator extraction from UAV LiDAR point cloud based on multi-type and multi-scale feature histogram[J]. Drones, 2024, 8(6): 241. DOI: 10.3390/drones8060241.
[10] Zhang J X, Lin X G, Ning X G.SVM-based classification of segmented airborne LiDAR point clouds in urban areas[J]. Remote Sensing, 2013, 5(8): 3749-3775. DOI: 10.3390/rs5083749.
[11] Chehata N, Guo L, Mallet C.Airborne lidar feature selection for urban classification using random forests[C]//ISPRS Workshop on Laser Scanning 2009. September 1-2, 2009, Paris, France. ISPRS, 2009: 135-140.
[12] Zhuang Y, Liu Y S, He G J, et al.Contextual classification of 3D laser points with conditional random fields in urban environments[C]//2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). September 28 - October 2, 2015, Hamburg, Germany. IEEE, 2015: 3908-3913. DOI: 10.1109/IROS.2015.7353927.
[13] Huang J J, Shen Y Q, Wang J G, et al.Automatic pylon extraction using color-aided classification from UAV LiDAR point cloud data[J]. IEEE Transactions on Instrumentation and Measurement, 2023, 72: 2520611. DOI: 10.1109/TIM.2023.3293543.
[14] Tang Q Y, Zhang L T, Lan G W, et al.A classification method of point clouds of transmission line corridor based on improved random forest and multi-scale features[J]. Sensors, 2023, 23(3): 1320. DOI: 10.3390/s23031320.
[15] Liu Y X, Aleksandrov M, Zlatanova S, et al.Classification of power facility point clouds from unmanned aerial vehicles based on adaboost and topological constraints[J]. Sensors, 2019, 19(21): 4717. DOI: 10.3390/s19214717.
[16] Chen C, Jin A, Yang B S, 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. DOI: 10.1016/j.jag.2022.102960.
[17] 王文曦,李乐林. 深度学习在点云分类中的研究综述[J]. 计算机工程与应用,2022,58(1):26-40. DOI:10.3778/j.issn.1002-8331.2105-0200.
[18] Caros M, Just A, Segui S.Object segmentation of cluttered airborne lidar point cloudsM//Puig V, ed. Artificial Intelligence Research and Development. Amsterdam: IOS Press, 2022: 259-268. DOI: 10.3233/FAIA220347.
[19] Wang G J, Wang L N, Wu S C, et al.Semantic segmentation of transmission corridor 3D point clouds based on CA-PointNet++[J]. Electronics, 2023, 12(13): 2829. DOI: 10.3390/electronics12132829.
[20] Zhou Y Y, Feng Z Y, Chen C L, et al.Bilinear distance feature network for semantic segmentation in PowerLine corridor point clouds[J]. Sensors, 2024, 24(15): 5021. DOI: 10.3390/s24155021.
[21] Kumar V, Nandy A, Soni V, et al.Powerline extraction from aerial and mobile LiDAR data using deep learning[J]. Earth Science Informatics, 2024, 17(4): 2819-2833. DOI: 10.1007/s12145-024-01310-w.
[22] Liu T Z, Hu B Y, Gu Y F, et al.An enhanced classification method based on adaptive multi-scale fusion for long-tailed multispectral point clouds[J]. Science China Information Sciences, 2025, 68(8): 182302. DOI: 10.1007/s11432-024-4324-6.
[23] Jiang Y Z, Gu Y F, Li X.Pillar-voxel fusion network for 3D object detection in airborne hyperspectral point clouds[J]. Science China Information Sciences, 2026, 69(1): 112301. DOI: 10.1007/s11432-024-4458-0.
[24] Peng S W, Xi X H, Wang C, et al.Point-based multilevel domain adaptation for point cloud segmentation[J]. IEEE Geoscience and Remote Sensing Letters, 2022, 19: 6500605. DOI: 10.1109/LGRS.2020.3037702.
[25] 李建,王健,王雷,等. 双重注意力机制的电力走廊点云语义分割[J]. 测绘通报,2025(4):127-133. DOI: 10.13474/j.cnki.11-2246.2025.0421.
[26] Thomas H, Qi C R, Deschaud J E, et al.KPConv: Flexible and deformable convolution for point clouds[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). October 27-November 2, 2019, Seoul, Korea. IEEE, 2019: 6411-6420. DOI: 10.1109/ICCV.2019.00651.
[27] Melekhov I, Umashankar A, Kim H J, et al.ECLAIR: A high-fidelity aerial LiDAR dataset for semantic segmentation[C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). June 17-18, 2024, Seattle, WA, USA. IEEE, 2024: 7627-7637. DOI: 10.1109/CVPRW63382.2024.00758.
[28] Hu Q Y, Yang B, Xie L H, et al.RandLA-net: Efficient semantic segmentation of large-scale point clouds[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). June 13-19, 2020, Seattle, WA, USA. IEEE, 2020: 11108-11117. DOI: 10.1109/CVPR42600.2020.01112.

基于改进KPConv的电力场景点云鲁棒语义分割方法研究^*

Robust Semantic Segmentation of Transmission Corridor Point Clouds Based on Improved KPConv

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 9

编辑推荐

Metrics

本文评价

访问统计

联系我们

[1]	廖凌岑, 刘巍, 刘士彬. 一种融合Transformer和UNet的森林覆盖信息提取方法[J]. 中国科学院大学学报, 2025, 42(3): 350-360.
[2]	王寅达, 陈嘉辉, 彭玲, 李兆博, 杨丽娜. 高分辨率关系图卷积网络遥感语义分割方法[J]. 中国科学院大学学报, 2025, 42(1): 107-115.
[3]	林大权, 范睿, 张良峰. 基于高维特征的图像对抗攻击算法[J]. 中国科学院大学学报, 2022, 39(3): 421-431.
[4]	胡莉娜, 高盛华. 基于层次化标签的人体解析[J]. 中国科学院大学学报, 2020, 37(6): 820-827.
[5]	郑敬华, 郭世泽, 高梁, 赵楠. 基于多任务学习的大五人格预测[J]. 中国科学院大学学报, 2018, 35(4): 550-560.
[6]	陈肃, 罗铁坚, 许延祥. 基于信任的推荐算法的鲁棒性分析[J]. 中国科学院大学学报, 2011, 28(2): 253-261.
[7]	丁飞, 尹红霞. 改进的基于均衡约束数学规划的分类模型[J]. 中国科学院大学学报, 2009, 26(5): 599-608.
[8]	古今, 郭立, 郑东飞. 一种基于感知特性的鲁棒性语音认证算法[J]. 中国科学院大学学报, 2009, 26(4): 474-482.
[9]	王蜀泉; 赵光恒. 基于模糊控制的卫星大角度姿态机动控制方法研究[J]. 中国科学院大学学报, 2006, 23(1): 111-117.