Wrong Wiring Detection of Electricity Meter Based on Image Processing

Abstract

Electricity meter wrong wiring is one of the common defects of electric energy metering device, which will cause safety and economic problems. Traditional wrong wiring detection of electricity meter mainly relies on manpower resources. In this paper, an automatic vision-based detection method is proposed for detecting the wrong wiring of electricity meter. Firstly, based on our own dataset, we modify and fine-tune the object detection model YOLOv3 to detect the meter type and the wiring area location. Then by using multiple image processing techniques including color segmentation, contour extraction and similarity analysis, we detect the wiring color sequence and compare it to the expected sequence to get the final result. The experiment result shows that the accuracy and real-time of this method is satisfactory.