Multimeter Digital Image Recognition Technology And Application Research

According to the request of JJG124-2005Ammeter, voltmeter, power meter and resistance meter testing procedures, in order to ensure the accuracy of the digital multimeter, the related power utilities need to send the digital multimeter to the Electricity Board to proof, but the proof result by manpower can’t be satisfied. In view of this matter, this thesis has designed and made the digital multimeter automatic verification come true, on the base of digital multimeter image recognition technology research.This paper mainly includes:Firstly, according to the request of the user and the national testing rules on multimeter, this thesis has designed and made the digital multimeter automatic verification system based on the digital image recognition technology.Secondly, on the base of the research method of digital character image preprocessing, this thesis analyze the character of the digital character, put forward to compromising the architectural feature from the threading method and the statistical property from the coarse grid method, and this thesis prove that compromising feature is better than the single feature in terms of anti-interference.Thirdly, using the fusion character from the above, this paper use two numerical recognition methods, template matching method and BP neural network method to have simulation experiment and analyzing comparison on the image recognition of numeric characters. The result indicates that the accuracy and instantaneity of BP neural network method is much better than the template matching method.Lastly, this thesis analyzed the function of the digital multimeter automatic verification system according to the national testing rules, the testing results are satisfied.