High Efficiency Water Meter Code Word Recognition System Based On Low-cost Platform

With the development and construction of the city,there is a shortage of water resources.Water conservation has become an important measure related to the national economy and the people's livelihood.In order to strengthen the management of water consumption by relevant departments,water meters came into being as a measure product.The water meter used to read and record manually,not only consumption of huge human cost,but also the accuracy and timeliness of data acquisition can not be guaranteed,easy to stimulate the conflict between water users and water management departments.With the continuous innovation of science and technology,intelligent water meter has gradually entered our sight.However,because of its high price,difficult maintenance,and the problems of power supply and electromechanical synchronization,it has not been widely used in the market.Therefore,the implementation of quick and convenient low cost intelligent transformation of mechanical water meters is particularly important in the construction of smart cities.The water meter code recognition system under the low cost platform is based on the OCR(Optical Character Recognition,optical character recognition)technology,and uses image processing and image recognition to realize the front end recognition of the direct reading mechanical water meter.A circuit board with image acquisition module is installed on the original water meter cover,which is cheap,simple,accurate and safe.The main contents of this paper are as follows:1.the preprocessing of the water meter picture has been realized.Because of the large picture taken by the camera,it takes up the memory greatly.In order to save space,it is necessary to divide the collected water meter image into block operation and grayscale processing.At the same time,due to environmental interference,improper illumination control,water meter accumulation and other factors,the noise information in the image should be removed.2.the correction of the water meter picture and the localization of the code area have been realized.It is a crucial step to accurately identify the code of water meter and find their location accurately.By observing the digital characters of the digital characters in the water meter,the edge of the image is obtained by the canny edge detection algorithm,and then the image tilt angle is searched by Hough line detection and then the image is rotated to correct the image.Finally,the location of the code word region is realized by morphological processing and pixel distribution.3.the single character segmentation and character recognition in code word area have been realized.First,the single character segmentation in the code word region is realized by pixel distribution and pixel projection.There are two cases of single character after segmentation,one is a complete digital character,and the other is the 2.5 character that is formed with the rolling of the word wheel.The sliding template matching algorithm solves the recognition problem of the half-word character while reducing the memory occupation,and adds secondary recognition after the recognition is completed.No rate.4.through FLASH reading and writing technology to ease the main chip RAM space shortage,so that low cost intelligent water meter recognition work is carried out normally.