A Multi-Target Detection Based Framework for Defect Analysis of Electrical Equipment

Recognizing and analyzing the effects of power equipment is an important way to ensure the reliable operation and safety of power systems. Manually inspecting the status of power equipment in the power Internet of Things (IoTs) is labor intensive when massive image data is needed to be inspected and analyzed. In this paper, a multi-target detection framework is proposed, which can be used as an edge service for defect analysis of power equipment over the power IoTs. The framework adopts a lightweight multi-layer convolutional neural network for multi-target detection based on YOLO, which is to learn the multi-target features in training data set for classifying and detecting image objects. The related experiments results show that the proposed framework for defect analysis of electrical equipment has excellent precision, and improves the performance of defect analysis application.