摘要
针对变电站指针式仪表人工抄表存在易受工作人员主观因素影响,常发生漏检、误检,存在工作环境辐射危害大等问题,设计一种由巡检机器人和图像处理技术组成的变电站指针式仪表自动读数系统。该系统以STM32芯片为核心,由工字电磁传感器、双自由度云台、高清变倍相机以及ESP8266等组成的巡检机器人,能精确定位在仪表前对仪表图像进行采集并传回计算机终端。由计算机终端通过改进的区域卷积神经网络(Faster Region Convolutional Neural Networks,Faster R-CNN)目标检测算法提取出仪表盘高清图像,去除冗余信息后经霍夫变换检测出指针中心线位置,得出仪表读数。
In view of the problems that the manual meter reading of substation pointer meter is easily influenced by the subjective factors of staff,which often occurs leakage and mis-checking,and the radiation hazard of working environment,we design an automatic reading system of substation pointer meter composed of inspection robot and image processing technology.The system is composed of an inspection robot with STM32 chip as the core,an I-word electromagnetic sensor,a doubledegree-of-freedom gimbal,a high-definition zoom camera and ESP8266,which can precisely position itself in front of the meter to collect the meter image and transmit it back to the computer terminal.The computer terminal extracts the highdefinition image of the instrument panel by the improved Faster Region Convolutional Neural Networks(Faster R-CNN)target detection algorithm,removes the redundant information and then detects the centerline position of the pointer by the Hough transform to obtain the instrument reading.
作者
李芋汶
李东阳
周舒涛
毛子安
LI Yuwen;LI Dongyang;ZHOU Shutao;MAO Zi'an(College of Mechanical and Electrical Engineering,Sichuan Agricultural University,Ya'an Sichuan 625014,China)
出处
《信息与电脑》
2023年第3期167-169,共3页
Information & Computer
