Ammeter Inspection With Densely Connected Object Detection Network

摘要

It is very important to understand the safety status of the electric metering devices in our daily life. This paper proposes a method based on deep learning to detect and evaluate electric metering devices. The detection Convolutional Neural Network (CNN) model is designed with the end-to-end architecture of You Only Look Once version 3(YOLOv3), but the backbone network is substituted with Densnet201 based on the dense connection idea. The detection objects include ammeter components of various sizes. The experimental results show that the proposed method can effectively detect the electric metering device in real time.