基于轻量化YOLOv3和Tesseract OCR的电力设备标志牌识别技术  

Electrical Sign Recognition Technology Based on Simplified YOLOv3 and Tesseract OCR

作　　者：李思妍 台升 张宇航 邱才明 LI Siyan;TAI Sheng;ZHANG Yuhang;QIU Caiming(School of Electronic Information and Electrical Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;School of Telecommunications,Huazhong University of Science and Technology,Wuhan 430074,China)

机构地区：[1]上海交通大学电子信息与电气工程学院,上海200240 [2]华中科技大学电信学院,湖北武汉430074

出　　处：《智慧电力》2021年第7期79-85,108,共8页Smart Power

基　　金：国家重点研发计划资助项目(2018YFF0214704)。

摘　　要：针对“无人化”变电站中电力设备标志牌的实时监测识别问题,提出了一种基于轻量化YOLOv3和Tesseract OCR的电力设备标志牌识别方法,针对标志牌特定的图像特征,设计了端点定位策略解决图像预处理的透视形变问题,利用Tesseract OCR引擎训练标志牌专属语言库,实现了高精度的电力设备标志牌检测与文字识别一体化。实验结果验证了所提方法对电力设备标志牌实时检测识别的有效性。Aiming at"unmanned"real-time monitoring and recognition of electrical signs in substations,a simplified YOLOv3 and Tesseract OCR method is proposed.In the light of the particular electrical sign image features,an end point positioning strategy is designed to treat perspective transformation during image pre-processing,and Tesseract OCR engine is used to train the exclusive electrical sign language database,realize the sign detection and recognition task with high accuracy.The experiment results prove the effectiveness of the proposed method.

关 键 词：标志牌检测 文字检测 轻量化YOLOv3 智能电网技术 

分 类 号：TM764[电气工程—电力系统及自动化]