基于感受野模塊的絕緣子實時識別定位方法
吉志朋1，張國偉1，盧秋紅2
（1 上海電力大學(xué) 自動化工程學(xué)院，上海 200082；2 上海合時智能科技有限公司，上海 201100）
摘 要：針對使用無人機進行絕緣子識別實時性的要求，以感受野模塊（RFB）網(wǎng)絡(luò)為基礎(chǔ)，提出了一種基于RFB模型改進的輕量型架構(gòu)。使用MobileNetV3網(wǎng)絡(luò)作為特征提取主干，設(shè)計了新的感受野模塊RFB-X，并使用Focal-loss損失函數(shù)解決正負樣本不平衡問題。實驗結(jié)果表明，該模型提高了絕緣子的檢測速度和準(zhǔn)確率。
關(guān)鍵詞：輕量型模型；感受野模塊；無人機；絕緣子檢測；實時性
中圖分類號：TM216 文獻標(biāo)識碼：A 文章編號：1007-3175(2020)09-0019-04
Real Time Detection of Insulator by RFB
JI Zhi-peng1, ZHANG Guo-wei1, LU Qiu-hong2
（1 School of Automation Engineering, Shanghai University of Electric Power, Shanghai 200082, China;
2 Shanghai Heshi Intelligent Technology Co., Ltd, Shanghai 2011 00, China）
Abstract: In response to the real-time requirements of using UAVs for insulator identification, based on the receptive field module (RFB) network, a lightweight architecture based on the improvement of the RFB model is proposed. Firstly, mobile MobileNetV3 is used as the main feature extraction module, then a new receptive field module RFB-X is designed, and finally the Focal-loss function is used to solve the imbalance of positive and negative samples. Experiments show that the model improves the speed and accuracy of insulator detection.
Key words: lightweight model; RFB; UAV; insulator detection; real time
參考文獻
[1] 吳濤，王偉斌，于力，等. 輕量級YOLOV3 的絕緣子缺陷檢測方法[J]. 計算機工程，2019，45(8)：275-280.
[2] 張煥坤，李軍毅，張斌. 基于改進型YOLO v3 的絕緣子異物檢測方法[J]. 中國電力，2020，53(2)：49-55.
[3] REDMON J, FARHADI A.YOLOv3:An Incremental Improvement[EB/OL].(2018-04-08)[2020-05-25]. https://arxiv.org/pdf/1804.02767.pdf.
[4] HOWARD A, ZHU M, CHEN B, et al.MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications[EB/OL].(2017-04-17) [2020-05-25].https://arxiv.org/pdf/1704.04861.pdf.
[5] 陳俊杰，葉東華，產(chǎn)焰萍，等. 基于Faster R-CNN 模型的絕緣子故障檢測[J]. 電工電氣，2020(4)：56-60.
[6] REN S, HE K, GIRSHICK R, et al.Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[C]//Neural Information Processing Systems，2015：91-99.
[7] 李偉性，鄭武略，王寧，等. 基于S S D 算法的輸電線路上絕緣子缺陷檢測方法研究[J]. 儀器儀表用戶，2019，26(8)：1-4.
[8] LIU S, HUANG D, WANG Y, et al.Receptive Field Block Net for Accurate and Fast Object Detection[C]//European Conference on Computer Vision，2018：404-419.
[9] HOWARD A, SANDLER M, CHU G, et al.Searching for MobileNetV3[EB/OL].(2019-11-20)[2020-05-25].https: //arxiv.org/pdf/1905.02244.pdf.
[10] LIN T, GOYAL P, GIRSHICK R, et al.Focal Loss for Dense Object Detection[C]//International Conference on Computer Vision，2017：2999-3007.
[11] SANDLER M, HOWARD A, ZHU M, et al.MobileNetV2: Inverted Residuals and Linear Bottlenecks[C]//Computer Vision and Pattern Recognition，2018：4510-4520.
[12] ZHANG X, ZHOU X, LIN M, et al.ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices[C]//Computer Vision and Pattern Recognition，2018：6848-6856.
[13] MA N, ZHANG X, ZHENG H, et al.ShuffleNetV2: Practical Guidelines for Efficient CNN Architecture Design[C]//European Conference on Computer Vision，2018：122-138.
[14] HU J, SHEN L, ALBANIE S, et al.Squeeze-and-Excitation Networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42(8)：2011-2023.
[15] QIN Z, LI Z, ZHANG Z, et al.ThunderNet: Towards Real-Time Generic Object Detection on Mobile Devices[C]//International Conference on Computer Vision，2019：6717-6726.
[16] SZEGEDY C, IOFFE S, VANHOUCKE V, et al. Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning[C]//National Conference on Artificial Intelligence，2016：4278-4284.
[17] CHOLLET F.Xception: Deep Learning with Depthwise Separable Convolutions[C]//Computer Vision and Pattern Recognition，2017：1800-1807.
[18] LIU W, ANGUELOV D, ERHAN D, et al.SSD: Single Shot MultiBox Detector[C]//European Conference on Computer Vision，2016：21-37.
[19] SHRIVASTAVA A, GUPTA A, GIRSHICK R, et al. Training Region-Based Object Detectors with Online Hard Example Mining[C]//Computer Vision and Pattern Recognition，2016：761-769.
