Optimization of Deep Learning Method for Automatic Recognition of Electric Energy Meter Manufacturers and Models

The manufacturing and application of electricity meters are constantly expanding with the development and expansion of the power industry. In order to better utilize the function of electricity meters and avoid accidents and malfunctions, this article believed that it is crucial to confirm the manufacturer, model, and even more relevant information of electricity meters through automatic identification. Moreover, this article also used deep learning methods to assist the automatic recognition system in identifying targets. In the final experiment of this article, a comparative experiment was conducted on the accuracy of conventional recognition systems and intelligent recognition systems based on convolutional neural networks (a deep learning technology). The results showed that the average accuracy of the conventional system was 78.79%, while the average accuracy of the intelligent system was 89.05%. Therefore, this study found that deep learning can improve the recognition accuracy of the automatic recognition system for electricity meter related information.