何宇浩,宋云海,何森,周震震,孙萌,陈毅,闫云凤
HE Yuhao,SONG Yunhai,HE Sen,ZHOU Zhenzhen,SUN Meng,CHEN Yi,YAN Yunfeng

• 1:南网超高压公司电力科研院
• 2:浙江大学机械工程学院
• 3:浙江大学海南研究院
• 4:浙江大学电气工程学院

摘要(Abstract)：

由于电力缺陷数据稀缺，目前大多数缺陷检测方法都无法有效地对电力缺陷情况进行准确的检测。为此，使用小样本图像生成方法，基于改进的LoFGAN（局部融合生成对抗网络），设计基于上下文信息的小样本图像生成器，提高缺陷检测网络对细节特征的提取能力；引入基于LC-散度的正则化损失来优化图像生成模型在有限数据集上的训练效果。实验表明，小样本图像生成方法能够为电力场景缺陷情况生成有效且多样的缺陷数据，所提模型能够有效解决电力缺陷场景数据稀缺的问题。
Due to the limited availability of power defect data, most current defect detection methods are unable to accurately detect power system anomalies. To overcome this challenge, a few-shot image generation method is employed. Building upon the improved local-fusion generative adversarial network(LoFGAN), a context-aware fewshot image generator is designed to enhance the defect detection network's capability to extract detailed features. A regularization loss based on LC-divergence is introduced to optimize the training effectiveness of the image generation model on limited datasets. Experimental results reveal that the few-shot image generation method can generate effective and diverse defect data for power scenarios. The proposed model can address the issue of data unavailability in power defect scenarios.

关键词(KeyWords)： 小样本图像生成;电力缺陷;上下文信息;LC-散度
few-shot image generation;power defect;context-aware;LC-divergence

Abstract:

Keywords:

基金项目(Foundation): 浙江省科技计划项目（2022C01056）;浙江省科技计划项目（LQ21F030017）;; CCF-联想蓝海科研基金

作者(Author): 何宇浩,宋云海,何森,周震震,孙萌,陈毅,闫云凤
HE Yuhao,SONG Yunhai,HE Sen,ZHOU Zhenzhen,SUN Meng,CHEN Yi,YAN Yunfeng

DOI: 10.19585/j.zjdl.202401015

参考文献(References)：

• [1] GENG S J,WANG X L.Predictive maintenance scheduling for multiple power equipment based on data-driven fault prediction[J]. Computers&Industrial Engineering,2022,164:107898.
• [2]齐冬莲，韩译锋，周自强，等.基于视频图像的输变电设备外部缺陷检测技术及其应用现状[J].电子与信息学报，2022,44(11):3709-3720.QI Donglian,HAN Yifeng,ZHOU Ziqiang,et al.External defect detection technology of power transmission and transformation equipment based on video image and its application status[J]. Journal of Electronics&Information Technology,2022,44(11):3709-3720.
• [3]黄悦华，陈照源，陈庆，等.基于边缘计算和改进YOLOv5s算法的输电线路故障实时检测方法[J].电力建设，2023,44(1):91-99.HUANG Yuehua,CHEN Zhaoyuan,CHEN Qing,et al.Real-time detection method for transmission line faults applying edge computing and improved YOLOv5s algorithm[J].Electric Power Construction,2023,44(1):91-99.
• [4]刘君，邓毅，杨延西，等.基于深度学习的空预器转子红外补光图像积灰状态识别[J].发电技术，2022,43(3):510-517.LIU Jun,DEND Yi,YANG Yanxi,et al.Ash accumulation state identification for infrared compensation images of air preheater rotor based on deep learning method[J].Power Generation Technology,2022,43(3):510-517.
• [5]魏业文，李梅，解园琳，等.基于改进Faster-RCNN的输电线路巡检图像检测[J].电力工程技术，2022,41(2):171-178.WEI Yewen,LI Mei,XIE Yuanlin,et al. Transmission line inspection image detection based on improved FasterRCNN[J].Electric Power Engineering Technology,2022,41(2):171-178.
• [6]赵振兵，蒋志钢，李延旭，等.输电线路部件视觉缺陷检测综述[J].中国图象图形学报，2021,26(11):2545-2560.ZHAO Zhenbing,JIANG Zhigang,LI Yanxu,et al.Overview of visual defect detection of transmission line components[J].Journal of Image and Graphics,2021,26(11):2545-2560.
• [7] ALI SIDDIQUI Z,PARK U,LEE S W,et al. Robust powerline equipment inspection system based on a convolutional neural network[J].Sensors,2018,18(11):3837.
• [8] MAO X D,LI Q,XIE H R,et al.Least Squares generative adversarial networks[C]//2017 IEEE International Conference on Computer Vision(ICCV).October 22-29,2017,Venice,Italy.IEEE,2017:2813-2821.
• [9]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].杭州：浙江大学，2022.FAN Li.Research and application of image data enhancement technology based on generative countermeasure network[D].Hangzhou:Zhejiang University,2022.
• [10] BROCK A,DONAHUE J,SIMONYAN K.Large scale GAN training for high fidelity natural image synthesis[EB/OL]. 2018:arXiv:1809.11096. https：//arxiv. org/abs/1809.11096.pdf.
• [11] VASWANI A,SHAZEER N,PARMAR N,et al.Attention is all You need[EB/OL]. 2017:arXiv:1706.03762.https：//arxiv.org/abs/1706.03762.pdf.
• [12] GOODFELLOW I J,POUGET-ABADIE J,MIRZA M,et al. Generative adversarial networks[EB/OL]. 2014:arXiv:1406.2661.https：//arxiv.org/abs/1406.2661.pdf.
• [13] KARRAS T,LAINE S,AITTALA M,et al.Analyzing and improving the image quality of StyleGAN[C]//2020IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). June 13-19,2020,Seattle,WA,USA.IEEE,2020:8107-8116.
• [14] ABDAL R,QIN Y P,WONKA P. Image2StyleGAN:how to embed images into the StyleGAN latent space?[C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV). October 27-November 2,2019,Seoul,Korea(South).IEEE,2020:4431-4440.
• [15] ZHU J Y,PARK T,ISOLA P,et al.Unpaired image-toimage translation using cycle-consistent adversarial networks[C]//2017 IEEE International Conference on Computer Vision(ICCV).October 22-29,2017,Venice,Italy.IEEE,2017:2242-2251.
• [16] ANTONIOU A,STORKEY A,EDWARDS H. Data augmentation generative adversarial networks[EB/OL]. 2017:arXiv:1711.04340. https：//arxiv. org/abs/1711.04340.pdf.
• [17] ZHAO S Y,LIU Z J,LIN J,et al.Differentiable augmentation for data-efficient GAN training[EB/OL]. 2020:arXiv:2006.10738.https：//arxiv.org/abs/2006.10738.pdf.
• [18] CLOU?TRE L,DEMERS M. FIGR:few-shot image generation with reptile[EB/OL].2019:arXiv:1901.02199.https：//arxiv.org/abs/1901.02199.pdf.
• [19] LIANG W X,LIU Z X,LIU C. DAWSON:a domain adaptive few shot generation framework[EB/OL].2020:arXiv:2001.00576.https：//arxiv.org/abs/2001.00576.pdf.
• [20] NICHOL A,ACHIAM J,SCHULMAN J.On first-order meta-learning algorithms[EB/OL]. 2018:arXiv:1803.02999.https：//arxiv.org/abs/1803.02999.pdf.
• [21] FINN C,ABBEEL P,LEVINE S.Model-agnostic metalearning for fast adaptation of deep networks[EB/OL]. 2017:arXiv:1703.03400. https：//arxiv. org/abs/1703.03400.pdf.
• [22] BARTUNOV S,VETROV D P. Few-shot Generative Modelling with Generative Matching Networks[C].International Conference on Artificial Intelligence and Statistics,2018.
• [23] HONG Y,NIU L,ZHANG J F,et al. Matchinggan:matching-based few-shot image generation[C]//2020IEEE International Conference on Multimedia and Expo(ICME).July 6-10,2020,London,UK.IEEE,2020:1-6.
• [24] GU Z,LI W B,HUO J,et al.LoFGAN:fusing local representations for few-shot image generation[C]//2021IEEE/CVF International Conference on Computer Vision(ICCV). October 10-17,2021,Montreal,QC,Canada.IEEE,2022:8443-8451.
• [25] ODENA A,OLAH C,SHLENS J.Conditional image synthesis with auxiliary classifier GANs[EB/OL]. 2016:arXiv:1610.09585.https：//arxiv.org/abs/1610.09585.pdf.
• [26] LI Y H,YAO T,PAN Y W,et al.Contextual transformer networks for visual recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2023,45(2):1489-1500.
• [27] KARL P. On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to have arisen from random sampling[M]//KOTZ S,JOHNSON N L.,Eds.Springer Series in Statistics.New York,NY:Springer New York,1992:11-28.
• [28] ZHANG R,ISOLA P,EFROS A A,et al.The unreasonable effectiveness of deep features as a perceptual metric[EB/OL]. 2018:arXiv:1801.03924. https：//arxiv. org/abs/1801.03924.pdf.