Deep learning-driven online safety monitoring and early hazard warning system for smart mines
Year: 2026
Type: article
Abstract: The rapid advancement of smart mining technologies has significantly increased the demand for intelligent systems capable of ensuring continuous safety monitoring and timely hazard prediction in underground mining environments. Traditional safety monitoring methods are mostly dependent on manual monitoring, single sensor monitoring, or some rules-based methods, with the disadvantages of late monitoring of hazards, limited situational awareness, and low prediction precision when the scene changes in the mining operations. To overcome these limitations, the paper proposes DLOSM-EHWS (Deep Learning-Driven Online Safety Monitoring and Early Hazard Warning System for Smart Mines), an intelligent framework that integrates deep learning with real-time monitoring to identify hazardous conditions and provide early warning alerts for improving mine safety. The proposed framework uses visual and sensor information to continuously monitor the working environment, miners and equipment, allowing abnormal events, as well as potential hazardous situations, to be detected at an early stage. Experimental results demonstrate the effectiveness of the proposed DLOSM-EHWS framework, achieving an accuracy of 98.76%, precision of 98.42%, recall of 98.15%, F1-score of 98.28%, and a false alarm rate of 1.42%. The findings demonstrate reliable hazard detection, accurate prediction of the safety-risk, and efficient real-time online safety monitoring in the smart mining environment. The proposed DLOSM-EHWS framework enhances proactive risk management, supports intelligent decision-making, and contributes to safer, more resilient, and sustainable smart mining operations. Not applicable.
Source: Discover Computing
Authors: Ran Shi, Jinchun Xue
Institution: Jiangxi University of Science and Technology
Language: English
Topic: Industrial and Mining Safety
Subfield: Safety, Risk, Reliability and Quality
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,520
OpenAlex ID: W7233584902
