Perception Method of Hazardous Information in Coal Mine Under Multi-sensor Conditions

XU Juan

doi:10.16545/j.cnki.cmet.2025.05.010

Abstract：Underground hazardous information encompasses multi-dimensional aspects such as gas concentration,geological activity,and personnel status,making it challenging to comprehensively perceive such information in coal mine underground.To address this information perception dilemma,a multi-sensor approach for hazardous information perception in coal mine underground was proposed.By meticulously analyzing the operational links of the mine,constructing an initial hazard source database,and defining and adjusting the objective function to quantify risks,the most pressing major hazard sources in coal mine underground were ultimately identified.Through the integration of various hazardous information in coal mine underground using multi-sensor technology,a feature signal fusion and conversion method was employed to efficiently extract multi-dimensional data.By integrating multiple sensors,it is possible to simultaneously capture multi-dimensional information such as gas concentration,geological activity,and personnel status,compensating for the limitations of single sensors and providing a more comprehensive environmental perception.By calculating the weight vector and combining multi-sensor data fusion with L1 regularization,accurate calculation and optimization of the weights for hazardous information perception in coal mine underground were achieved,ultimately classifying risk levels based on the weight values.Experimental results showed that the sensitivity of hazardous information perception using the multi-sensor method consistently remained above 90％,with a false negative rate below 1％.Therefore,it is demonstrated that the multi-sensor method provides precise and comprehensive dynamic risk assessment technical support for mine safety prevention and control.