Abstract: This paper studies a novel direct-reading meter reading terminal based on machine vision, and constructs a water resources information management system based on it. The system's structure, overall characteristics, basic hardware structure, and the specific implementation of each functional module are described in detail. Experimental results show that the system can effectively identify the meter readings.
Keywords: Machine vision; Automated meter reading; Character recognition; GPRS
1 Introduction
With technological advancements and improved living conditions, the demand for intelligent housing is increasing, leading to the implementation of individual water meters per household. The implementation of multi-rate water pricing and the need for data statistics and billing management by water management departments necessitate a transformation of traditional mechanical water meters and manual meter reading methods, achieving digital and automatic meter reading. Current remote automatic meter reading systems complete digital automatic meter reading through three steps: sensor transformation, data accumulation and temporary storage, and data transmission. This requires modification of existing meters, where magnetic materials are embedded in a pointer or gear of the water meter to convert the metered readings into electrical pulse signals. These pulse signals are then collected, accumulated, stored, and transmitted to a host computer for remote meter reading. However, the modification process is complex, and the transmission of pulse signals is susceptible to interference, resulting in low reliability. Therefore, the metering results are difficult to perfectly match the data displayed on the meter. A critical drawback is the inability to verify the accuracy of the pulse metering results; comparison with the meter readings is necessary, requiring manual correction when needed. Throughout this process, the water meter cannot operate without a power source, resulting in low reliability and a large maintenance workload.
To address the aforementioned shortcomings, improvements to water meter sensing technology are necessary. Therefore, this paper integrates key machine vision technologies such as image acquisition, image processing, and pattern recognition into the design of an embedded meter reading terminal, attempting to employ a novel technical approach to achieve direct-reading automatic meter reading and resolve existing problems in current automatic meter reading technologies.
For details, please click: Research on Direct-Reading Meter Reading System Based on Machine Vision
