The Research On Smart Meter Automated Inspection Technology Based On Machine Vision

Smart meters play a crucial role as a terminal sensing unit in national smart gridconstruction. In smart meter verification link, its parameters settings, controlinstructions and other key commands implementation need for verification, and itsnameplate appearance, LCD display quality need to be checked, which uses todetermine this meter whether is in line with national standards and quality requirements.Currently in the artificial test mode, verification inefficient and workers abnormaloperation phenomena have occurred from time to time easily cause power measuringinstruments appearance injured, indicators declined and other situation, causing themeter quality declined. In order to ensure quality, it must put an end to operationirregularities causing power measuring instruments cosmetic damaged, the qualitydeclined and other conditions in the artificial test. Therefore, the paper achieves smartmeters automatic detection by introducing the machine vision technology, The smartmeters automatic detection is consists of three main areas: smart metersidentification(SMI), smart meters LCD screen appearance detection(SMAD), smartmeters programming cap screws identification and precise positioning(SMIP).The SMI is to read the smart meters factory numbers, in this paper, combining theimage one-dimensional barcode recognition and digital identification method based onthe histogram projection is able to guarantee the stable and accurate results of the smartmeter identification.The SMAD is the destructive and nondestructive testing for the quality of LCD, thisarticle achieves accurate segmentation of the LCD screen area first, design qualitytesting standards based on the characters information of LCD to accomplish the LCDappearance detection.The SMIP is to realize the smart meters detection robot trigger the smart meters feecontrol infrared signal safe and effective, this paper is based on the screws texturefeature information and contrast image matching recognition method and texture gravitymethod, find that the latter can only stably, accurately position the programming capscrews location.By achieving the algorithms on automatic detection smart meters, the system ismounted on the Mianyang Fude smart meters detection robot FDV6, and real-timeoperation in Chongqing electric power research institute, the test results show that the system is good able to meet the smart meters automation testing requirements of theState Grid Corporation.