基于Faster R-CNN模型的絕緣子故障檢測
陳俊杰，葉東華，產(chǎn)焰萍，陳凌睿
（國網(wǎng)漳州供電公司，福建 漳州 363000）
摘 要：絕緣子是電力系統(tǒng)中用來支撐電線和電氣隔離的重要器件，對輸配電線路絕緣狀態(tài)的在線檢測意義重大。針對現(xiàn)階段人工判別航拍圖像的不足，提出基于Faster R-CNN的絕緣子圖像故障檢測方案，闡述了卷積神經(jīng)網(wǎng)絡(luò)特征提取的原理，構(gòu)建基于Faster R-CNN的絕緣子檢測模型，利用無人機航拍的絕緣子圖像及故障樣本，對檢測模型加以訓(xùn)練與測試，分別進行絕緣子分類檢測實驗和絕緣子故障定位實驗。實驗結(jié)果表明，所提出的絕緣子故障檢測方法能夠準確對絕緣子進行檢測與分類，并定位出故障位置，且達到實時性要求。
關(guān)鍵詞：絕緣子檢測；故障定位；卷積神經(jīng)網(wǎng)絡(luò)；圖像檢測；深度學(xué)習(xí)
中圖分類號：TM216；TM855 文獻標識碼：A 文章編號：1007-3175(2020)04-0056-05
Insulator Fault Detection Based on Faster R-CNN
CHEN Jun-jie, YE Dong-hua, CHAN Yan-ping, CHEN Ling-rui
（State Grid Zhangzhou Power Supply Company, Zhangzhou 363000, China）
Abstract: Insulators are important devices used to support electrical wires and electrical isolation in power systems, and are of great significance for online test of the insulation status of transmission and distribution lines. In this paper, in view of the shortcomings of manually discriminating aerial images at this stage, an insulator image fault detection scheme based on Faster R-CNN is proposed, and the principle of feature extraction for convolutional neural networks is described, and an insulator detection model based on Faster R-CNN is constructed. Utilizing aerial insulator images and fault samples of aerial drones, the detection model is trained and tested, and insulator classification detection experiments and insulator fault location experiments are performed respectively. Experimental results show that the proposed insulator fault detection method can accurately detect and classify insulators, locate the fault location, and meet the real-time requirements.
Key words: insulator detection; fault location; convolutional neural network; image detection; deep learning
參考文獻
[1] 仝衛(wèi)國，苑津莎，李寶樹. 圖像處理技術(shù)在直升機巡檢輸電線路中的應(yīng)用綜述[J]. 電網(wǎng)技術(shù)，2010，34(12)：204-208．
[2] 朱虎，李衛(wèi)國，林治. 絕緣子檢測方法的現(xiàn)狀與發(fā)展[J]. 電瓷避雷器，2006(6)：13-17.
[3] PARK K C, MOTAI Y, YOON J R. Acoustic Fault Detection Technique for High Power Insulators[J].IEEE Transactions on Industrial Electronics，2017，64(12)：9699-9708．
[4] 黃霄寧，張真良. 直升機巡檢航拍圖像中絕緣子圖像的提取算法[J]. 電網(wǎng)技術(shù)，2010，34(1)：194-197．
[5] 徐耀良，張少成，楊寧，等. 航拍圖像中絕緣子的提取算法[J]. 上海電力學(xué)院學(xué)報，2011，27(5)：515-518．
[6] 趙振兵，金思新，劉亞春. 基于NSCT的航拍絕緣子圖像邊緣提取方法[J]. 儀器儀表學(xué)報，2012，33(9)：2045-2052．
[7] 李衛(wèi)國，葉高生，黃鋒，等. 基于改進MPEG-7紋理特征的絕緣子圖像識別[J]. 高壓電器，2010，46(10)：65-68．
[8] OBERWEGER M, WENDEL A, BISCHOF H.Visual recognition and fault detection for power line insulators[C]//19th Computer Vision Winter Workshop，2014．
[9] ZHANG Xinye, AN Jubai, CHEN Fangming.A method of insulator fault detection from airborne image[C]//2010 Second WRI Global Congress on Intelligent Systems，2010．
[10] 姜浩然，金立軍，閆書佳. 航拍圖像中絕緣子的識別與故障診斷[ J ] . 機電工程，2015，32(2)：274-278．
[11] KRIZHEVSKY Alex, SUTSKEVER I, HINTON G.Imagenet classification with deep convolutional neural networks[C]//Proceedings of the 25th International Conference on Neural Information Processing Systems，2012．
[12] REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN：Towards Real-Time Oobject Detection with Region Proposal Networks[C]//IEEE Transactions on Pattern Analysis and Machine Intelligence，2016．
[13] ZEILER M D, FERGUS R. Visualizing and understanding convolutional networks[C]//Computer Vision-ECCV 2014：13th European Conference，2014.
[14] SIMONYAN K, ZISSERMAN A. Very Deep Convolutional Networks for Large-Scale Image Recognition[C]//3rd International Conferenceon on Learning Representations，2015．
