Design And Implementation Of Meter Reading Identification System

Current user readings in the water meter,usually directly to use the naked eye to identify.For the meter reading,the workload is undoubtedly heavy,but there is a certain degree of misunderstanding and management deficiencies.With the rapid development of digital image processing and artificial intelligence,water meter reading recognition technology has sufficient technical support to enable the automatic determination of meter readings,combined with mobile Internet and emerging equipment,the technology applied to the mobile phone to make it Can quickly spread.This paper mainly uses artificial intelligence and image processing technology to preprocess the water meter image taken by the mobile phone,then detect and identify the pointer and the digital part of the water meter,and finally realize the automatic interpretation of the water meter.The main achievements of this paper are as follows:(1)Water surface image preprocessing.Using the bilateral filter to achieve the image denoising;the use of GrayWorld algorithm to achieve image brightness balance;the use of color histogram equalization algorithm to achieve image contrast enhancement;the use of image processing Hough transform,polygon fitting,affine transformation and other images Tilt correction.(2)Realize the water meter pointer extraction and identification.Using the color space transformation and morphological transformation of the image,the water pointer is extracted,the center of each pointer is found based on the Hough transform,and the direction of each pointer is found in combination with the improved step length method.Finally,Correlation calculates the number of pointers.(3)Water meter digital extraction and identification.In this paper,we use the polygon fitting algorithm to extract the digital border,and combine the contour detection to segment the numbers.A shading removal algorithm is designed to remove the shadow around the numbers.A half-word recognition algorithm is designed for the half-word and finally Use convolutional neural networks to identify numbers.(4)Based on the above three tasks,a water meter reading identification system is designed and implemented.The system includes mobile terminal display part,background business processing part and database part.Mobile end supports iOS and Android dual-end system display,the background includes water meter reading and user management module.The water meter reading recognition module is the core of this paper and the system.The system can accurately identify the water meter digits and pointers,and the results back to the user,reducing the workload of the meter reading,has some practical significance.The paper ends with several areas for improvement.