Design And Implementation Of The Character Recognition System Of The Dial For Digital Multimeters

The Character Recognition System of the Dial for Digital Multimeters(DMMS)is an electrical signal measuring instrument,widely used in various fields,in real life greatly convenient for equipment and debugging of the circuit.But electrical signal types differ in thousands ways,various properties,the electrical signal of the unit is also each are not identical,character size and display position also each are not identical,hand-held multimeter had mostly digital interface,it is difficult to directly connect with the computer.To design a high-performance DMMS face recognition system,recognition dial characters shown in the content,realize the automation of measurement process,has important application value for electronic measurement,not only broaden the use of the instrument,improve the reliability of testing data,at the same time also improved the detection efficiency of testing,and for fault timely analysis provides a reliable guarantee.This paper first discusses the character recognition technology development present situation,analyzes the DMMS dial with general character recognition character recognition system design process(license plate recognition,image text recognition,web text recognition,etc.)the difference between,points out the existing digital multimeter dial character recognition system design process of uneven light source,background interference noise big,the wind rate of problems and challenges.Secondly,according to user requirements,design the overall design scheme of character recognition system,in order to reduce the effects of the external light to identify internal optical structure is used to light up the dial is designed.Again,to improve the accuracy of the dial character recognition,reduce the background interference,to dial DMMS captured by the regional division,the entire screen is divided into digital area,unit area and background area etc,and the result of the division of a tilt correction,character illumination compensation,adaptive threshold binarization operation such as pretreatment,and analysis and comparison of various image features representation method,horizontal projection and vertical projection,the local gray features,directional line element features,side outline,stroke density characteristics,peripheral features and penetration,through analysis,this study applied in character segmentation part Hough change and characteristics of the projection,at the end of character features extraction using local gray features,namely the partition density characteristics of the method to complete the feature extraction of a single character.Compared the effectiveness of each method,through the statistics of relevant information on the vertical direction and horizontal direction,finally to extract image partition density feature;Finally,template method is discussed and the advantages and disadvantages of text recognition method based on neural network,because the BP neural network has strong ability of resisting noise and interference,the paper finally determines the text recognition using BP neural network to complete the technical proposal,the area is extracted using the previous partition density characteristics of the training for DMMS dial character recognition of BP neural network,finally completed a DMMS dial the design of the character recognition system,and which has the function of the system in accordance with user requirements in a Lab VIEW platform of VI(virtual instrument).The experimental results show that this design DMMS face recognition system,not only can adapt to different environment,can also identify the DMMS different working conditions,and the network response speed is less than 2 seconds,the recognition accuracy of 99% or more.