作者：王昕¹, 赵飞², 蒋佐富¹, 尚将³, 吴瑞文¹

作者单位：1. 台州宏创电力集团有限公司, 浙江 台州 318000;
2. 华北电力大学电气与电子工程学院, 河北 保定 071000;
3. 国网浙江台州市黄岩区供电有限公司, 浙江 台州 318000

关键词：图像识别;电力设备;迁移学习;卷积神经网络

摘要：

一线生产单位一般不具备建立大量电力设备图像数据集的条件，因此在使用深度学习模型协助完成对电力设备图像的识别过程中受到限制。通过对电力设备进行三维建模和多角度渲染，获得大量模拟电力设备图像，解决深度学习模型卷积神经网络在学习过程中数据集不足的问题。同时，通过迁移学习的方式将经过模拟电力设备图像训练的卷积神经网络应用于对真实电力设备图像的学习中，提高学习效率和精度，最终取得93.5%的识别准确率。该方法为一线生产单位将卷积神经网络应用于电力设备图像识别分类任务提供一种解决办法。

Power equipment image recognition method based on transfer learning and convolutional neural network
WANG Xin¹, ZHAO Fei², JIANG Zuofu¹, SHANG Jiang³, WU Ruiwen¹
1. Taizhou Hongchuang Power Group Limited Company, Taizhou 318000, China;
2. School of Electric and Electrical Engineering, North China Electric Power University, Baoding 071000, China;
3. State Grid Zhejiang Taizhou Huangyan Power Supply Co., Ltd., Taizhou 318000, China
Abstract: Front-line production units generally do not have the conditions to establish a large number of power equipment image data sets, so they are limited in the use of deep learning model to assist in the process of power equipment image recognition. Through three-dimensional modeling and multi-angle rendering of power equipment, a large number of images of power equipment were obtained, which solved the problem of insufficient data set in the learning process of deep learning model convolutional neural network. At the same time, the convolution neural network trained by the simulated power equipment image is applied to the learning of real power equipment image by means of transfer learning, which improves the learning efficiency and accuracy and finally achieves the recognition accuracy of 93.5%. This method provides a solution for the first-line production units to apply the convolutional neural network to the power equipment image recognition and classification task.
Keywords: image recognition;electrical equipment;transfer learning;convolutional neural network
2020, 46(5):108-113  收稿日期: 2019-04-10;收到修改稿日期: 2019-05-13
基金项目: 国家自然科学基金项目（51177047）
作者简介: 王昕(1970-),男,浙江台州市人,高级工程师,研究方向为电力设备在线监测
参考文献
[1] 王丛, 刘高, 刘平原. 具有分级报警功能的电力巡检系统[J]. 自动化与仪器表, 2019(2): 183-185, 189
[2] 马力, 王致, 张丹, 等. 基于深度学习的人脸识别技术在电力巡检机器人中的应用研究[J]. 自动化与仪器仪表, 2019(2): 36-38
[3] 彭向阳, 王柯, 肖祥, 等. 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(2): 368-376
[4] 李军锋. 基于深度学习的电力设备图像识别及应用研究[D]. 广州: 广东工业大学, 2018.
[5] 李军锋, 王钦若, 李敏. 结合深度学习和随机森林的电力设备图像识别[J]. 高电压技术, 2017, 43(11): 3705-3711
[6] LECUN Y L, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324
[7] SMIRNOV E A, TIMOSHENKO D M, ANDRIANOV S N. Comparison of regularization methods for imagenet classification with deep convolutional neural networks[J]. Aasri Procedia, 2014, 6(1): 89-94
[8] 王致, 马力, 洪永健, 等. 基于电力巡检机器人巡视系统的设备故障诊断研究[J]. 自动化与仪器仪表, 2019(1): 68-71
[9] 程海燕, 韩璞, 董泽, 等. 无人机航拍电力部件图像准确分类[J]. 计算机仿真, 2018, 35(8): 424-428
[10] SCHMIDHUBER J. Deep learning in neural networks: an overview[J]. Neural Netw, 2015, 61: 85-117
[11] 李策, 陈海霞, 汉语, 等. 深度学习算法中卷积神经网络的概念综述[J]. 电子测试, 2018(23): 61-62
[12] 徐星, 宋小鹏, 杜春晖. 基于深度学习的验证码图像识别[J]. 测试技术学报, 2019, 33(2): 138-142
[13] SAINATH T N, KINGSBURY B, SAON G, et al. Deep convolutional neural networks for large-scale speech tasks[J]. Neural Networks, 2015, 64: 39-48
[14] 袁功霖, 侯静, 尹奎英. 基于迁移学习与图像增强的夜间航拍车辆识别方法[J]. 计算机辅助设计与图形学学报, 2019, 31(3): 467-473