Reseach On Automatic Recognization Technologies Of Instrument

Gauge has played an increasingly important role in various parameter measuring and monitoring. In common cases, the data collection for old gauge is highly human dependent, which not only reduces the efficiency but also raises danger to the data-collecting worker where the gauge is placed in hazardous environment. It is desired to develop automatic identification in gauge data collection to displace manual reading. In this thesis, based on machine vision and internet of things, gauge panel identification has been investigated and experiment validation has been carried out.A comprehensive literature review has been conducted and the research background of this thesis has been stated in detail. The introduction of theories related to machine vision and internet of things have been made, as well as its developments and applications. Upon those, a feasible realization solution of the thesis has been proposed, which is specifically compromised of the following steps: the image collection, image preprocessing and identification, as well as its data transmission. The image preprocessing consists of color space conversion, grey processing and its binaryzation, enhancement, tilt correction, localization, segmentation and image thinning. Various preprocessing algorithms have been compared and the most suitable one has been selected. Image identification algorithm has been carefully chosen and its adaption to the proposed solution has been made.The gauge automatic identification system has been realized with the efforts in software, hardware and optical layout design and optimization. The software has been comprised of client and server. ARM development board has been selected as hardware platform while 3G module has been employed to data transmission. Besides, a forward duel-side illumination has been adopted as the result of optical analysis and optimization.Sets of experiment tests have been carried out, including functionality test, versatility test, illumination and tilt dependence tests, which has demonstrated the feasibility of the gauge automatic identification system. It is concluded that this system has provided an effective solution for the gauge information-upgradation of old factories.