基于深度学习的指针式仪表自动读数方法  

Automatic identification for reading of pointer-type meters based on deep learning

作　　者：杨诗琪 吴佳仪 陈墨楠 付陈修 赵宁[1] 王婧[4] Yang Shiqi;Wu Jiayi;Chen Monan;Fu Chenxiu;Zhao Ning;Wang Jing(School of Quality and Technical Supervision,Hebei University,Baoding 071002,China;School of Cyberspace Security and Computer,Hebei University,Baoding 071002,China;Beijing ConST Instrument Technology Co.,Inc.,Beijing 100094,China;Baoding Cigarette Factory,Hebei Baisha Tobacco Co.,Ltd.,Baoding 071000,China)

机构地区：[1]河北大学质量技术监督学院,保定071002 [2]河北大学网络空间安全与计算机学院,保定071002 [3]北京康斯特仪表科技股份有限公司,北京100094 [4]河北白沙烟草有限责任公司保定卷烟厂,保定071000

出　　处：《电子测量技术》2023年第5期149-156,共8页Electronic Measurement Technology

基　　金：国家自然科学基金(62173122);国家级大学生创新创业训练计划项目(202210075012);河北省自然科学基金(F2021201031)项目资助

摘　　要：指针式仪表广泛应用于石油化工、工业制造和烟草行业动力部门等领域。鉴于人工巡检频率较低、部份仪表安装环境恶劣等因素,致使工厂生产过程存在安全隐患,巡检人员人身安全难以保障。本文基于现有的工业生产过程的监控摄像系统提出了一种基于YOLO V3目标检测与DeepLab V3+图像分割技术的指针式仪表读数自动识别方法。通过引入YOLO V3目标检测模型检测并切割出仪表表盘子图像。结合图像特点与实际需求,改进了DeepLab V3+模型,加入腐蚀操作,有效提取了子图像中的刻度线与指针信息。通过OCR技术提取子图像仪表量程,根据刻度线与指针的相对位置关系,计算得到仪表读数。实验结果表明该方法平均相对误差为2.17%,算法稳定可靠且处理速度快。Pointer-type meters are widely applied into the petrochemical,industrial manufacturing,and the power sector of the tobacco industry.Given the low frequency of manual inspection and the hostile environment of part installed meters,the security risks exist in the procedures of the factories production,and personal safety of inspectors is difficult to be guaranteed.Based on the existing surveillance camera system of the procedures of the factories production,this paper proposes an automatic identification method of reading of pointer-type meters based on YOLO V3object detection and DeepLab V3+image segmentation techniques.YOLO V3is introduced to detect and incise the sub-images of the gauge dial.According to the characteristics of the meter images and actual needs,this paper improves the structure of DeepLab V3+and add corrosion treatment.So the scale lines and pointers of the sub-images are located effectively.The gauge range is extracted from the sub-images by OCR technology.And in combination with the relative positional relationship of the scale lines and pointers,the identification of meters reading can be computed.It is proved by experiment that the average relative error of reading of the meter image identified is 2.17%,and the proposed method can meet the applications requirements.

关 键 词：目标检测 语义分割 YOLO V3 DeepLab V3+ 指针式仪表 

分 类 号：TP216[自动化与计算机技术—检测技术与自动化装置] TP391.4[自动化与计算机技术—控制科学与工程] TH86[机械工程—仪器科学与技术]