Digitaization of Analogue Meter Reading Using Convolution Neural Network

Authors

  • Azmin Raziq Rizaman School of Electrical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Johor, Malaysia.
  • Hazlina Selamat School of Electrical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Johor, Malaysia.
  • Nurulaqilla Khamis Department of Data Science, Universiti Malaysia Kelantan, City Campus, Pengkalan Chepa, 16100 Kota Bharu, Kelantan, Malaysia.

DOI:

https://doi.org/10.15282/mekatronika.v3i1.7149

Keywords:

Analogue meter, Digitization, Convolution neural network, Deep learning

Abstract

Analogue meter is a device that has been widely used in a various industry to monitor and obtain the reading of the measurement. Based on the conventional approach, the meter reading will be done continuously by the meter reader that might cause high tendency of human error during the observation. To minimize this fallacy, this approach taken in this paper enables the automation of this the process by obtaining the reading from an analogue meter using an image processing technique and send the output to the central database for further processing. By implementing this approach, observation efficacy can be improved. This paper describes the process on how to obtain the digitized reading of an analogue meter using images captured by a camera. The images are then processed using an image processing method and the Convolutional Neural Network (CNN) is used to determine the reading of the meter. Data is then sent to the MySQL database, as this approach was easily implemented and managed either on-premises or via the cloud. The use case in this study was based on the analogue meter for domestic electricity supply in Malaysia and results show that the meter reading can accurately be recognized using the proposed approach.

Downloads

Published

2021-06-18

How to Cite

[1]
A. R. . Rizaman, H. . Selamat, and N. Khamis, “Digitaization of Analogue Meter Reading Using Convolution Neural Network”, Mekatronika: J. Intell. Manuf. Mechatron., vol. 3, no. 1, pp. 28–34, Jun. 2021.

Issue

Section

Original Article