Detection of Mine Water Supply and Drainage Pipe Rupture Based on Improved YOLOv5
GAO Yongjun
GUO Yunfei
TANG Wenjie
Abstract：The rupture of mine water supply and drainage pipes will have adverse effects on timely water supply and drainage underground.A mine water supply and drainage pipe rupture detection method based on improved YOLOv5 was proposed to address the problem of manual inability to quickly and accurately locate and detect the rupture of mine water supply and drainage pipes.The original backbone network of YOLOv5 was replaced with FasterNet to reduce the model size of YOLOv5,accelerate computation speed,and improve detection accuracy.The C3 module of YOLOv5 Neck was introduced into the MHSA(Multi-head Self Attention Mechanism)structure to enhance the feature extraction capability at the site of water pipe rupture.The experimental results showed that compared with the original YOLOv5 model,the improved YOLOv5 had a 4.9 percentage point increase in mAP,reaching 92.3％,a 2.3 percentage point increase in Precision,reaching 97.1％,and a 27.52％increase in running speed.In terms of model size,YOLOv5 has reduced its parameter count by 18.67％,computational complexity by 43.03％,and model size by 19.52％.This can provide reference for the rapid and accurate detection of mine water supply and drainage pipe ruptures.
Keywords：YOLOv5FasterNetMHSAobject detectionwater pipe rupture
Publication Date：2025-02-28
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 68-73 )