Gas Meter Reading Recognition Algorithm Based On Digital Image Processing

With the construction of urbanization in China,gas demand is growing,the existing gas meter reading methods(such as the current pulse or IC card prepaid)can not meet the needs of gas flexible and convenient settlement.To complete the gas billing management and other functions has great significance without entering the user home.Digital character recognition is an important research direction in the field of image processing.At present,how to realize the digital character recognition of direct reading gas meter dial more accurately and quickly is still a research hotspot and difficulty.In order to improve the accuracy and speed of recognition,this paper presents a meter recognition algorithm based on rough set and support vector machine optimized by improved quantum particle swarm optimization(QPSO)algorithm.As the primitive gas meter image is preprocessed and the mixed features is extracted to obtain the eigenvector attribute of the corresponding ten numbers characters,the rough set is used to reduce the attribute of the extracted features,and the minimum attribute reduction is obtained.QPSO has the advantages of less control parameters and faster convergence rate than genetic algorithm.It is applied to the optimized rough set.In order to overcome the shortcomings of QPSO,the artificial colony search operator can increase the ability to jump out of the local value,and introduce the immune mechanism of the immune algorithm to keep the convergence rate of the particle.There are two parameters that affect the recognition accuracy and speed of the support vector machine.The improved quantum particle swarm optimization algorithm is used to optimize the support vector machine parameters.The simulation results show that a meter recognition algorithm based on rough sets and support vector machine optimized by improved quantum particle swarm optimization algorithm has a higher accuracy and speed.The main work of the paper includes the following aspects:(1)This paper analyzes the research status of character recognition technology and digital character recognition algorithm at home and abroad,and studies the optimization algorithms of digital character recognition based on rough set and support vector machine.Combining with the characteristics of digital characters in gas table,digital character recognition algorithm model is established.The pre-processing technique of the digital image of the gas meter dial is analyzed.Considering the actual situation that the acquired image may contain noise and lack of light,the image is enhanced by the histogram equalization technique,and the noise is obtained by removing the noise.The feature data of the characters is extracted by the mixed feature extraction method,and the individual data characters are represented by the feature data,and the sample data is provided for the next step.(2)Based on the study of quantum particle swarm optimization(QPSO),the advantages and disadvantages of the algorithm are analyzed.The quantum particle swarm optimization algorithm may fall into the local value problem.Artificial colony search operator and immune algorithm are introduced to increase the search range of the latter iteration,jump out of local value and keep the algorithm converging speed.(3)The rough set attribute reduction algorithm is studied,and the attribute vector reduction of the extracted digital character of the dial is reduced.The improved quantum particle swarm optimization algorithm is used to optimize the rough set,and the minimum attribute reduction is obtained.The corresponding digital characters of the gas meter dial are more concise,Accurate and refined characteristic attribute data.(4)SVM meter number recognition algorithm optimized by the improved particle swarm optimization is proposed to obtain the relative optimal parameters of the support vector machine,and the accuracy of the identification of the gas meter character image is improved.