基于电容层析成像（electrical capacitance tomography，ECT）系统的流型识别技术具有非侵入性和可视化的优点，被广泛应用于各种工业过程的动态测量中。为提高ECT系统的流型识别准确率，基于神经网络模型的识别算法被广泛应用，但ECT系统难以获得大量实测样本，从而导致识别准确率低、模型泛化性差。针对这一问题，该文采用了基于深度卷积生成对抗网络（DCGAN）的数据集扩增方法，利用经过预处理的LSSVM方法重建所得的图像，搭建基于DCGAN的流型图像生成模型进行实验，并利用GAN-train和GAN-test两个客观指标对生成图像进行评价，之后使用扩增后的数据集进行基于YOLOv3的流型识别。实验结果表明，利用DCGAN方法能够有效扩增原有数据集，生成图像准确度高、丰富性强，配合YOLOv3的目标检测算法，可以对ECT系统的流型识别准确率提升有较大程度的改进。 The flow pattern recognition technology based on capacitive tomography (ECT) system is widely used in dynamic measurement of various industrial processes due to its non-invasive and visual advantages. To improve the flow pattern identification accuracy of ECT system, recognition algorithm based on neural network model is widely used. But the ECT system can't get a lot of ECT system measured samples, resulting in low accuracy, low model generalization. To solve this problem, this paper adopted the dataset amplification method based on deep convolution generative adversarial network (DCGAN), and used the images reconstructed by the preprocessed LSSVM method to build a flow pattern image generation model based on DCGAN for experiments. Gan-train and GAN-test were used to evaluate the generated images. After that, the amplified dataset is used for flow pattern identification based on YOLOv3. The experimental results show that the DCGAN method can effectively amplify the original data set and generate images with high accuracy and richness. Together with the target detection algorithm of YOLOv3, the flow pattern recognition accuracy of ECT system can be improved to a great extent. 流型识别 数据集扩增 机器学习 深度卷积生成对抗网络 YOLOv3 flow pattern identification data set amplification machine learning deep convolution generative adversarial network YOLOv3 参考文献 KRIZHEVSKY A , SUTSKEVER I , HINTON G E . ImageNet classification with deep convolutional neural networks [J]. Communications of the ACM , 2017 , 60 ( 6 ): 84 - 90 . ImageNet classification with deep convolutional neural networks 刘起超 , 周云龙 , 陈聪 . 基于CEEMDAN和概率神经网络的起伏振动气液两相流型识别 [J]. 仪器仪表学报 , 2021 , 42 ( 10 ): 84 - 93 . 基于CEEMDAN和概率神经网络的起伏振动气液两相流型识别 LIU Q C , ZHOU Y L , CHEN C , Flow pattern identification of fluctuating vibration gas liquid two phase flowbased on CEEMDAN and probabilistic neural network [J]. Chinese Journal of Scientific Instrument , 2021 , 42 ( 10 ): 84 - 93 . Flow pattern identification of fluctuating vibration gas liquid two phase flowbased on CEEMDAN and probabilistic neural network 王小鑫 , 王博 , 陈阳正 , 等 . 基于电容层析成像技术重构图像的两相流流型识别 [J]. 计量学报 , 2020 , 41 ( 8 ): 942 - 946 . 基于电容层析成像技术重构图像的两相流流型识别 WANG X X , WANG B , CHEN Y Z , et al . Two-phase flow pattern recognition based on electrical capacitance tomography reconstructed images [J]. Acta Metrologica Sinica , 2020 , 41 ( 8 ): 942 - 946 . Two-phase flow pattern recognition based on electrical capacitance tomography reconstructed images 王莉莉 , 乔立勇 , 陈宇 , 等 . 基于特征提取和RBF神经网络的ECT流型辨识 [J]. 计算机工程与应用 , 2011 , 47 ( 5 ): 175 - 178 . 基于特征提取和RBF神经网络的ECT流型辨识 WANG L L , QIAO L Y , CHEN Y , et al . Flow pattern identification based on feature extraction and RBF neural net-work for electrical capacitance tomography [J]. Computer Engineering and Applications , 2011 , 47 ( 5 ): 175 - 178 . Flow pattern identification based on feature extraction and RBF neural net-work for electrical capacitance tomography 宋蕾 , 陈德运 , 姚玉梅 , 等 . Elman神经网络在ECT系统流型辨识中的应用 [J]. 哈尔滨理工大学学报 , 2014 , 19 ( 5 ): 103 - 108 . Elman神经网络在ECT系统流型辨识中的应用 SONG L , CHEN D Y , YAO Y M , et al , Application of Elman neural network in pattern identification for electrical capacitance tomography [J]. Journal of Harbin University of Science and Technology , 2014 , 19 ( 5 ): 103 - 108 . Application of Elman neural network in pattern identification for electrical capacitance tomography 张立峰 , 佟彤 . 基于GA－ELM及电容层析成像的两相流流型辨识 [J]. 电力科学与工程 , 2020 , 36 ( 8 ): 37 - 41 . 基于GA－ELM及电容层析成像的两相流流型辨识 ZHANG L F , TONG T , Identification of two-phase flow based on genetic algorithm-extreme learning machine and electrical capacitance tomography [J]. Electric Power Science and Engineering , 2020 , 36 ( 8 ): 37 - 41 . Identification of two-phase flow based on genetic algorithm-extreme learning machine and electrical capacitance tomography XIE Q Z , DAI Z H , HOVY E , et al . Unsupervised data augmentation for consistency training [C]// Proceedings of 34th International Conference on Neural Information Processing Systems . New York : ACM , 2020 : 6256 - 6268 . Unsupervised data augmentation for consistency training NAZKI H , LEE J , YOON S , et al . Image-to-image translation with GAN for synthetic data augmentation in plant disease datasets [J]. Korean Institute of Smart Media Journal , 2019 , 8 ( 2 ): 46 - 57 . Image-to-image translation with GAN for synthetic data augmentation in plant disease datasets RADFORD A , METZ L , CHINTALA S . Unsupervised representation learning with deep convolutional generative adversarial networksv [EB/OL]. ( 2016-01-07 )[ 2022-10-20 ]. https://arxiv.org/abs/1511.06434 https://arxiv.org/abs/1511.06434 . Unsupervised representation learning with deep convolutional generative adversarial networksv IOFFE S , SZEGEDY C . Batch normalization: Accelerating deep network training by reducing internal covariate shift [EB/OL]. ( 2015-03-02 )[ 2022-10-20 ]. https://arxiv.org/abs/1502.03167 https://arxiv.org/abs/1502.03167 . Batch normalization: Accelerating deep network training by reducing internal covariate shift ZHANG R . Making convolutional networks shift-invariant again [EB/OL]. ( 2019-06-09 )[ 2022-10-20 ]. https://arxiv.org/abs/1904.11486 https://arxiv.org/abs/1904.11486 . Making convolutional networks shift-invariant again RUDER S . An overview of gradient descent optimization algorithms [EB/OL]. ( 2017-06-15 )[ 2022-10-20 ]. https://arxiv.org/abs/1609.04747 https://arxiv.org/abs/1609.04747 . An overview of gradient descent optimization algorithms DUCHI J C , HAZAN E , SINGER Y . Adaptive subgradient methods for online learning and stochastic optimization [J]. Journal of Machine Learning Research , 2011 , 12 : 2121 - 2159 . Adaptive subgradient methods for online learning and stochastic optimization KINGMA D P , BA J . Adam: A method for stochastic optimization [EB/OL]. ( 2017-01-30 )[ 2022-10-20 ]. https://arxiv.org/abs/1412.6980 https://arxiv.org/abs/1412.6980 . Adam: A method for stochastic optimization WU Z X , HE S P . Improvement of the AlexNet networks for large-scale recognition applications [J]. Iranian Journal of Science and Technology, Transactions of Electrical Engineering , 2021 , 45 ( 2 ): 493 - 503 . Improvement of the AlexNet networks for large-scale recognition applications 郑宗生 , 刘贝 , 卢鹏 , 等 . 基于AlexNet的近岸水生植物光谱分类及特征光谱分析 [J]. 中国激光 , 2023 , 50 ( 2 ): 129 - 138 . 基于AlexNet的近岸水生植物光谱分类及特征光谱分析 ZHENG Z S , LIU B , LU P , et al . Spectral classification and characteristic spectral analysis of nearshore aquatic plants based on AlexNet [J]. Chinese Journal of Lasers , 2023 , 50 ( 2 ): 129 - 138 . Spectral classification and characteristic spectral analysis of nearshore aquatic plants based on AlexNet 巩泉役 , 彭克 , 陈羽 , 等 . 基于电弧随机性和卷积网络的交流串联电弧故障识别方法 [J]. 电力系统自动化 , 2022 , 46 ( 24 ): 162 - 169 . 基于电弧随机性和卷积网络的交流串联电弧故障识别方法 GONG Q Y , PENG K , CHEN Y , et al . Ac series arc fault identification method based on arc randomness and convolutional network [J]. Automation of Electric Power Systems , 2022 , 46 ( 24 ): 162 - 169 . Ac series arc fault identification method based on arc randomness and convolutional network SIMONYAN K , ZISSERMAN A . Very deep convolutional networks for large-scale image recognition [EB/OL]. ( 2015-04-10 )[ 2022-10-20 ]. https://arxiv.org/abs/1409.1556 https://arxiv.org/abs/1409.1556 . Very deep convolutional networks for large-scale image recognition 何其霖 , 穆平安 . VGG网络与多特征融合的遮挡人脸检测 [J]. 电子测量技术 , 2021 , 44 ( 18 ): 150 - 154 . VGG网络与多特征融合的遮挡人脸检测 HE Q L , MU P A , Occlusion face detection based on VGG network and multi-feature fusion [J]. Electronic Measurement Technology , 2021 , 44 ( 18 ): 150 - 154 . Occlusion face detection based on VGG network and multi-feature fusion 孙继平 , 范伟强 . MS-ADoG域结合ReNLU与VGG-16的矿井双波段图像融合算法 [J]. 光子学报 , 2022 , 51 ( 3 ): 1 - 15 . MS-ADoG域结合ReNLU与VGG-16的矿井双波段图像融合算法 SUN J P , FAN W Q , MS-ADoG domain combined with ReNLU and VGG-16 mine dual-band image fusion algorithm [J]. Acta Photonica Sinica , 2022 , 51 ( 3 ): 1 - 15 . MS-ADoG domain combined with ReNLU and VGG-16 mine dual-band image fusion algorithm REDMON J , FARHADI A . YOLOv3: An incrementalimprovement [EB/OL]. ( 2018-04-08 )[ 2022-04-06 ]. https://arxiv.org/abs/1804.02767 https://arxiv.org/abs/1804.02767 . YOLOv3: An incrementalimprovement 杨坤志 , 闫潇宁 , 孙健 , 等 . 一种雨雾背景的DeRF-YOLOv3-X目标检测方法 [J]. 传感技术学报 , 2022 , 35 ( 9 ): 1222 - 1229 . 一种雨雾背景的DeRF-YOLOv3-X目标检测方法 YANG K Z , YAN X N , SUN J , et al . A DeRF-YOLOv3-X object detection method for rainy and foggy background [J]. Chinese Journal of Sensors and Actuators , 2022 , 35 ( 9 ): 1222 - 1229 . A DeRF-YOLOv3-X object detection method for rainy and foggy background 王非 , 王欣宇 , 周景春 , 等 . 一种基于YOLOv3的水下声呐图像目标检测方法 [J]. 电子与信息学报 , 2022 , 44 ( 10 ): 3419 - 3426 . 一种基于YOLOv3的水下声呐图像目标检测方法 WANG F , WANG X Y , ZHOU J C , et al . An underwater object detection method for sonar image based on YOLOv3 model [J]. Journal of Electronics & Information Technology , 2022 , 44 ( 10 ): 3419 - 3426 . An underwater object detection method for sonar image based on YOLOv3 model 张官荣 , 赵玉 , 李波 , 等 . 基于YOLOv3的SAR舰船图像目标识别技术 [J]. 电光与控制 , 2022 , 29 ( 9 ): 107 - 110 . 基于YOLOv3的SAR舰船图像目标识别技术 ZHANG G R , ZHAO Y , LI B , et al . Ship target detection in SAR image based on YOLOv3 [J]. Electronics Optics & Control , 2022 , 29 ( 9 ): 107 - 110 . Ship target detection in SAR image based on YOLOv3 祖鑫萍 , 李丹 . 基于无人机图像和改进YOLOv3-SPP算法的森林火灾烟雾识别方法 [J]. 林业工程学报 , 2022 , 7 ( 5 ): 142 - 149 . 基于无人机图像和改进YOLOv3-SPP算法的森林火灾烟雾识别方法 ZU X P , LI D , Research on forest fire smoke identification method based on UAV images and improved YOLOv3-SPP algorithm [J]. Journal of Forestry Engineering , 2022 , 7 ( 5 ): 142 - 149 . Research on forest fire smoke identification method based on UAV images and improved YOLOv3-SPP algorithm