Research on fault and display defect detection system of watt-hour meter based on image processing

With the rapid development of smart grids, electric energy meters, as an important component of power equipment, have a direct impact on the reliability and efficiency of power systems. In response to the limitations of traditional fault detection methods in terms of efficiency and accuracy, an automatic detection system for display defects in smart meters based on image processing technology is researched and designed. An efficient fault detection framework is constructed by combining convolutional neural network (CNN) feature extraction and support vector machine (SVM) classification algorithms. Experimental results demonstrate that the system performs excellently in multiple performance indicators. In practical applications, the fault detection accuracy reaches 93.8%, with detection delay and system response time reduced to 180 milliseconds and 220 milliseconds, respectively. Compared with other combined algorithms, the processing capability of the system is significantly improved, capable of processing 1650-meter images per hour. In tests involving various fault types, the system achieves F1 scores of 0.90 and 0.86 for "Blurry Display" and "Cracked Digits" faults, respectively, demonstrating excellent classification ability and high robustness. The proposed smart meter fault detection system in this study can effectively enhance the accuracy and real-time performance of fault identification, providing reliable technical support for the operation and maintenance of smart grid equipment and possessing significant application value.