智能化的输电线路异物检测系统对于电力的健康监测以及电气传输安全具有重要意义。在实际应用中，由于地势环境复杂多变，不可控因素多，因此现有输电线路异物检测系统存在检测精度低、工作人员操作不便捷的问题。同时，异物入侵输电线路图片数据稀缺，进一步限制了对异物智能化检测模型的研究。因此本文针对以上问题设计了解决方案，构建了一种基于特征生成网络的输电线路异物检测系统，主要内容如下：

（1）构建基于特征生成网络的输电线路异物检测系统框架。首先分析了输电线路异物检测系统的工作环境，并结合实际巡检情况进行了需求分析，然后对系统方案进行了设计，其次针对输电线路异物检测系统存在的不足之处进行彻底的分析，构建了基于特征生成网络的输电线路异物检测系统框架，并对框架的各个模块进行了简要叙述。

（2）提出面向输电线路的多元化特征生成网络模型。异物检测模型是以数据作为驱动的，需要大量的图片作为研究支撑，而该场景的数据集十分稀缺，已经成为异物检测模型发展的瓶颈。本文通过设计图像融合算法、纹理分割算法、自筛选网络对生成对抗网络进行优化，分别构建了异物特征多元化模型和场景特征多元化模型，可以针对不同特征进行数据扩增，从而获得大量的图片数据集。实验结果表明，该模型在输电线路场景的应用性能强，扩增的图片不仅质量高，而且多样性强。

（3）设计面向输电线路的实时入侵异物目标检测模型。入侵输电线路的异物具有小目标、多种类的特点，同时易发生形变，易被遮挡，导致检测难度大，准确率低，现有检测模型无法满足该场景的需求，因此本文设计了面向输电线路的实时入侵异物目标检测模型，该模型引入了残差结构与注意力机制，并优化了非极大值抑制算法。实验结果表明，该模型具有很好的检测性能。

（4）建立面向输电线路的入侵异物智能化检测软件平台。交互性强、一体化的检测软件平台可以极大的提高输电线路异物巡检的效率。采用MVC设计模式，MySQL数据库进行设计，然后对目标追踪功能进行了研究，使得无人机可以根据异物检测模型的结果调整自身参数。实验结果表明，软件平台工作人员体验良好且功能齐全。

An intelligent foreign object detection system for transmission lines is of significant importance for power grid health monitoring and electrical transmission safety. However, in practical applications, due to the complex and ever-changing terrain environment and numerous uncontrollable factors, existing foreign object detection systems for transmission lines suffer from low detection accuracy and inconvenient operation by personnel. Additionally, the scarcity of image data of foreign object intrusion into transmission lines further limits the research of intelligent detection models for foreign objects. To address these issues, this paper proposes a foreign intrusion object detection system for transmission lines based on a feature generation network. The main contents of the system are as follows:

(1) This paper presents a framework for detecting foreign objects on transmission lines based on feature generation networks. First, the working environment of the transmission line foreign object detection system is analyzed, and a requirement analysis is conducted based on actual inspection situations. The system scheme is then designed, and the inadequacies of the transmission line foreign object detection system are thoroughly analyzed. A framework of a transmission line foreign object detection system based on a feature generation network is constructed, and a brief description of each module in the framework is provided.

(2) This paper proposes a diverse feature generation network model for foreign object detection on transmission lines. The foreign object detection model is data-driven, it requires a large number of images as research support. However, the dataset about this scenario is scarce and has become a bottleneck for the development of foreign object detection models. To address this issue, this article designs image fusion algorithms, texture segmentation algorithms, and self-screening networks to optimize the generative adversarial network. Two diverse feature generation models are constructed: one for foreign object features and the other for scene features. These models can perform data augmentation on different features, thereby obtaining a large number of image datasets. The experimental results show that the proposed model has strong application performance in the transmission line scenario, and the augmented images not only have high quality but also strong diversity.

(3) This paper designs a real-time foreign object intrusion detection model for transmission lines. The foreign objects that intrude on transmission lines have the characteristics of small targets, multiple types, and are prone to deformation and occlusion, which leads to difficulty in detection and low accuracy. Existing detection models cannot meet the requirements of this scenario. Therefore, this paper proposes a real-time foreign object intrusion detection model for transmission lines, which introduces residual structures and attention mechanisms and optimizes the non-maximum suppression algorithm. The experimental results show that the proposed model has excellent detection performance.

(4) This paper establishes an intelligent intrusion foreign object detection software platform for transmission lines. An interactive and integrated detection software platform can greatly improve the efficiency of foreign object inspections on transmission lines. The platform is designed using the MVC design pattern and MySQL database, and the target tracking function is studied so that unmanned aerial vehicles can adjust their parameters based on the results of the foreign object detection model. The experimental results show that the software platform has good user experience and complete functions for the working personnel.

[1] 刘健. 基于YOLOX的输电线路异物检测算法研究及软件设计[D].徐州：中国矿业大学,2022.

[2] Wang Lei, He Yigang, Li Lie. A Single-Terminal Fault Location Method for HVDC Transmission Lines Based on a Hybrid Deep Network[J]. Electronics, 2021, 10(3): 255.

[3] Weng Liguo, Zhang Yanghui, Yang Yong, Fang Ming, Yu Zelong. A mobility compensation method for drones in SG-eIoT[J]. Digital Communications and Networks, 2021, 7(2): 196-200.

[4] 任贵新. 输电线路覆冰及异物智能视频检测算法研究[D].哈尔滨：哈尔滨理工大学,2019.

[5] Hyeoncheol Baik, Jorge Valenzuela. Unmanned Aircraft System Path Planning for Visually Inspecting Electric Transmission Towers[J]. Journal of Intelligent & Robotic Systems, 2019, 95: 1097-1111.

[6] Jiang Bian, Xiaolong Hui, Xiaoguang Zhao, Min Tan. A monocular vision–based perception approach for unmanned aerial vehicle close proximity transmission tower inspection[J]. International Journal of Advanced Robotic Systems, 2019, 16(1): 1729881418820227.

[7] 韩冰,尚方. 面向无人机输电线路巡检的电力杆塔检测框架模型[J]. 浙江电力, 2016, 35(4): 6-11.

[8] 马飞越,刘佳豪,赵涛,佃松宜,牛勃. 电力巡检机器人运行姿态的终端滑模控制[J]. 西南大学学报(自然科学版), 2021, 43(12): 172-179.

[9] 邢志刚,万宏宇. 多旋翼无人机在输电线路巡检中的应用[J]. 黑龙江科学,2020,11(24):116-117.

[10] 郝艳捧,梁苇,潘锐健,罗兵,李立浧,张福增,王婷婷.输电线路智能带电检修关键技术研究综述[J].电力自动化设备,2022,42(02):163-175.

[11] 张勇.架空输电线路障碍物巡检的无人机低空摄影测量方法研究[D].武汉： 武汉大学, 2017.

[12] 许永波. 基于DCNN的航拍输电线路图像特征分类和覆冰检测算法研究[D].郑州：郑州大学,2019.

[13] 彭向阳,易琳,钱金菊,王柯,郑晓光,韩正伟,陈国强.大型无人直升机电力线路巡检系统实用化[J].高电压技术,2020,46(02):384-396.

[14] 彭向阳,王柯,肖祥,吴开春,包令聪,顾温国.大型无人直升机电力线路智能巡检宽带卫星通信系统[J].高电压技术,2019,45(02):368-376.

[15] 齐南珣. 输电线路巡检图像的智能缺陷检测方法研究[D].北京：华北电力大学,2018.

[16] Yang Wu, Gen Zhao, Jun Hu, Yong Ouyang, Shan X. Wang, Jinliang He, Feng Gao, Sen Wang. Overhead Transmission Line Parameter Reconstruction for UAV Inspection Based on Tunneling Magnetoresistive Sensors and Inverse Models[J]. IEEE Transactions on Power Delivery,2019,34(3):819-827.

[17] Manh Duong Phunga, Cong Hoang Quacha, Tran Hiep Dinh, Quang Ha. Enhanced discrete particle swarm optimization path planning for UAV vision-based surface inspection[J]. Automation in construction,2017,81:25-33.

[18] Naser Hossein Motlagh, Miloud Bagaa, and Tarik Taleb. UAV-Based IoT Platform: A Crowd Surveillance Use Case[J]. IEEE Communications Magazine,2017,55(2):128-134.

[19] Carol Martinez, Carlos Sampedro, Aneesh Chauhan, Jean Francois Collumeau, Pascual Campoy. The Power Line Inspection Software (PoLIS): A versatile system for automating power line inspection[J]. Engineering Applications of Artificial Intelligence, 2018, 71:293-314.

[20] Hamelin P , F Mirallès, Lambert G, et al. Discrete-time control of LineDrone: An assisted tracking and landing UAV for live power line inspection and maintenance[C]// International Conference on Unmanned Aircraft Systems (ICUAS). IEEE, 2019:292-298.

[21] 王淼,李源源,陈艳芳,刘伟东,武艺,陈利明.大型固定翼无人机在架空输电线路中的应用模式研究[J].测绘通报,2017,(S1):159-163.

[22] 梁文勇,吴大伟,谷山强,王海涛,严碧武,刘辉.输电线路多旋翼无人机精细化自主巡检航迹优化方法[J].高电压技术,2020,46(09):3054-3061.

[23] Chi Chen, Bisheng Yang, Shuang Song, Xiangyang Peng, Ronggang Huang. Automatic Clearance Anomaly Detection for Transmission Line Corridors Utilizing UAV-Borne LIDAR Data[J]. Remote Sensing, 2018, 10(4):613.

[24] Jiao Shengxi, Wang Haiyang. The Research of Transmission Line Foreign Body Detection Based on Motion Compensation[C]//First International Conference on Multimedia and Image Processing. IEEE, 2016:10-14.

[25] Ye Xuhui, Wang Dong, Zhang Daode, Hu Xinyu. Transmission Line Obstacle Detection Based on Structural Constraint and Feature Fusion[J]. Symmetry, 2020, 12(3):452.

[26] Wang Bixiao, Wu Runze, Zheng Zhe, Zhang Weilong, Guo Jinghong. Study on the method of transmission line foreign body detection based on deep learning[C]//IEEE Conference on Energy Internet and Energy System Integration (EI2). IEEE, 2017:1-5.

[27] Xia Pengfei, Yin Jin, He Jingsong, Yang Kejun. Neural Detection of Foreign Objects for Transmission Lines in Power Systems[C]//Journal of Physics: Conference Series. IOP Publishing, 2019, 1267(1): 012043.

[28] Yao Nan, Zhu Lvfu Zhu. A Novel Foreign Object Detection Algorithm Based on GMM and K-Means for Power Transmission Line Inspection[C]//Journal of Physics Conference Series, 2020, 1607(1):012014.

[29] Zhang Fan, Fan Yalei, Cai Tao, Liu Wenda, Hu Zhongqiu, Wang Nengqing, Wu Minghu. OTL-Classifier: Towards Imaging Processing for Future Unmanned Overhead Transmission Line Maintenance[J]. Electronics, 2019, 8(11):1270.

[30] Sharma Vipul, Mir Roohie Naaz. A comprehensive and systematic look up into deep learning based object detection techniques: A review[J]. Computer Science Review, 2020, 38:100301.

[31] Song Chunhe, Xu Wenxing, Wang Zhongfeng, Yu Shimao, Zeng Peng, Ju Zhaojie. Analysis on the Impact of Data Augmentation on Target Recognition for UAV-Based Transmission Line Inspection[J]. Complexity, 2020, 2020:1-11.

[32] Krizhevsky Alex, Sutskever Ilya, Hinton GeoffreyE. ImageNet Classification with Deep Convolutional Neural Networks[J]. Communications of the ACM, 2017, 60(6): 84-90.

[33] Shorten Connor, Khoshgoftaar Taghi M. A survey on Image Data Augmentation for Deep Learning[J]. Journal of Big Data, 2019, 6(1):1-48.

[34] 林本旺,赵光哲,王雪平,李昊. 基于组残差块生成对抗网络的面部表情生成[J/OL].计算机工程与应用,2023:1-12.

[35] 王通,陈延彬.基于改进生成对抗网络的动液面建模数据扩充[J].电子测量与仪器学报,2023,37(02):99-109.

[36] 王泽,姜斌,程月华,张香燕,薛琪. 一种基于生成对抗网络的卫星异常检测方法[J].空间控制技术与应用,2023,49(01):113-120.

[37] 姜世博,孙跃文,许硕,吴志芳. 基于生成对抗网络的CT图像金属伪影校正[J].哈尔滨工程大学学报,2022,43(12):1766-1771.

[38] 章仕稳. 基于生成对抗网络的行人轨迹预测方法的研究[D]. 南京：南京邮电大学, 2022.

[39] 任志斌. 高压输电线巡检机器人控制系统的研究与实现[D]. 上海：上海大学, 2009.

[40] 范黎. 基于生成对抗网络的图像数据增强技术研究及应用[D]. 浙江：浙江大学, 2022.

[41] Jia Deng, Wei Dong, Socher Richard, Li Jia Li, Kai Li, Li Fei-Fei. ImageNet: A large-scale hierarchical image database[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, 2009: 248-255.

[42] Francisco J. Moreno-Barea, Fiammetta Strazzera, Jerez JoseM. Daniel Urda, Leonardo Franco. Forward Noise Adjustment Scheme for Data Augmentation[C]// IEEE Symposium Series on Computational Intelligence (IEEE SSCI 2018). IEEE, 2018:728-734.

[43] Summers Cecilia, Dinneen Michael J. Improved Mixed-Example Data Augmentation[C]// 2019 IEEE winter conference on applications of computer vision (WACV). IEEE, 2019: 1262-1270.

[44] Zhong Zhun, Zheng Liang, Kang Guoliang, Li Shaozi, Yang Yi. Random Erasing Data Augmentation[C]//Proceedings of the AAAI conference on artificial intelligence. 2020, 34(07): 13001-13008.

[45] Konno Tomohiko, Iwazume Michiaki . Icing on the Cake: An Easy and Quick Post-Learnig Method You Can Try After Deep Learning[J]. arXiv preprint arXiv:1807.06540, 2018.

[46] Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, Yoshua Bengio. Generative adversarial networks[J]. Communications of the ACM, 2020, 63(11):139–144.

[47] Bowles Christopher, Chen Liang, Guerrero Ricardo, et al. GAN Augmentation: Augmenting Training Data using Generative Adversarial Networks[J]. arXiv preprint arXiv:1810.10863, 2018.

[48] Zolna Konrad, Zajac Michal, Rostamzadeh Negar, Pedro O. Pinheiro. Adversarial Framing for Image and Video Classification[C]//Proceedings of the AAAI Conference on Artificial Intelligence. 2019, 33(01): 10077-10078.

[49] Justin Johnson, Alexandre Alahi, Fei-Fei Li. Perceptual Losses for Real-Time Style Transfer and Super-Resolution[C]//Computer Vision–ECCV, 2016: 694-711.

[50] 徐思源. 基于StyleGAN的心脏表面三维重建方法研究[D].成都：电子科技大学,2022.

[51] 尹芹,方晖,王金东,王侃,晏天文,霍智勇. 基于样式编码的真实图像逆映射算法[J].南京邮电大学学报(自然科学版),2023,43(01):80-85.

[52] 张兴娜,杨会,刘雪宇,姜秋竹,李荣山,周晓霜,王晨.基于DenseNet算法对膜性肾病组织病理图像肾小球钉突分类研究[J].重庆医学,2022,51(24):4308-4312.

[53] 李自华. 基于GPU服务器的分布式机器学习平台研究[D]. 成都：电子科技大学,2021.

[54] 冯彦卿. 基于深度学习的视觉目标检测与跟踪技术研究[D].湖南:国防科技大学,2020.

[55] 王佩雯. 基于多尺度感知的密集人群计数方法研究与实现[D].重庆:重庆邮电大学,2022.

[56] 陈英红,杜明坤.基于Fast-RCNN与结构光纵焊缝三维形态参数检测方法[J].中国测试,2018,44(12):85-90.

[57] 杨明雄. 基于Faster-Rcnn的离线数学公式识别切割技术的研究与实现[D].成都：电子科技大学,2019.

[58] 张帆,郭思媛,任方涛,张新红,李结平.基于改进YOLOv3的玉米叶片气孔自动识别与测量方法[J].农业机械学报,2023,54(02):216-222.

[59] 白一睿,方辉,张泽.基于YOLO神经网络模型的花生智能精选系统设计[J].机械,2023,50(02):1-6.

[60] 陈德海,孙仕儒,王昱朝,雷志军.改进SSD算法的小目标检测研究[J].传感器与微系统,2023,42(03):65-68+72.

[61] 范明凯,肖满生,胡一凡,吴宇杰.基于改进SSD算法的胃部息肉图像检测[J].软件工程,2023,26(03):30-35.

[62] 沈翔. 基于RetinaNet的小目标检测提升方法研究[D].南京：南京邮电大学,2022.

[63] 宋建熙,李兴科,于哲,李西兵.改进Retina-Net的草坪杂草目标检测[J].中国农机化学报,2022,43(12):170-177.

[64] Joseph Redmon, Santosh Kumar Divvala, Ross Girshick, Ali Farhadi. You Only Look Once: Unified, Real-Time Object Detection[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 779-788.

[65] 路齐硕. 基于深度学习的目标检测方法研究[D].北京：北京邮电大学,2020.