Research On Image Recognition Method Of Direct Reading Meter Reading

With the rapid development of computer technology,image recognition technology is also in continuous development and improvement.Image automatic recognition technology has been applied in various fields of production and life,which has liberated people from some repetitive and mechanical working environment,greatly improved the efficiency of production,and brought great convenience to people's life.For example,automatic license plate recognition,handwritten digital character recognition and face recognition have been put into use,which all bring convenience to people's daily life.Direct-reading meter reading recognition is the use of image processing technology,to enhance the meter image,enhance the visual effect of image,you can dig up information cannot be captured in the human visual characteristics,and then extracted into the instrument,the feature information of the image,and uses digital character recognition technology,automatic accurate readings of meter image,will people liberated from a single meter and repetitive work,not only can liberate the labour force,but also improve the accuracy of the readings,direct reading instrument is widely used in today's society is undoubtedly has the very vital significance.This paper aims to design a simple,fast and high recognition rate identification system,which can be applied to various industrial sites,even some special work sites that cannot be accessed by workers.For this purpose,this paper mainly does the following work:(1)Instrument reading image preprocessing.In the image acquisition stage,considering the influence of various complex environments on the image,the poor quality meter reading image is enhanced,the over-bright or over-dark image is processed by histogram equalization,the uneven illumination meter reading image is processed by local threshold,and the tilted meter reading map is rectified.(2)Digital character segmentation and feature extraction of instrument reading image.In order to better identify the meter reading in the future,the preprocessed image needs to be segmented,and the digital characters in the meter reading area are divided into single digital characters and normalized.At the same time,feature extraction is carried out on the normalized numeric characters,and the extracted features are taken as the feature quantity of numeric characters,which is conducive to the later reading recognition.(3)Instrument reading image digital character recognition.In this paper,three different recognition methods,template matching,support vector machine andconvolution neural network,were selected to identify the meter reading,and the results after recognition were compared and analyzed,so as to design a simple,fast and high recognition rate recognition system.