An AIoT and improved YOLOv5 algorithm-based high-voltage cable distributed intelligent monitoring method and system

Abstract

The application provides a high-voltage cable distributed intelligent monitoring method and system based on AIoT and improved YOLOv5 algorithm, and belongs to the technical field of high-voltage cable intelligent monitoring. The method comprises the following steps: collecting original sample images of high-voltage cables in a cable tunnel, and performing image data expansion and preprocessing; constructing a high-voltage cable anomaly detection model based on a YOLOv5 model; embedding a SimAM attention module in the backbone network of the YOLOv5 model in the high-voltage cable anomaly detection model, and replacing a C3 module with a C3-GhostDynamic module; training the high-voltage cable anomaly detection model using the preprocessed images, obtaining the trained high-voltage cable anomaly detection model; collecting real-time high-voltage cable images in the cable tunnel, and identifying cable anomalies through the trained high-voltage cable anomaly detection model. The application realizes high-precision identification of various abnormal states such as personnel intrusion and smoke in the cable channel, and improves the management level of the high-voltage cable channel.

Images