基于FocalLoss的煤矿井下人员安全视觉监测方法

A Visual Monitoring Method of Personal Safety under the Coal Mine Based on FocalLoss

监测矿井工作人员违规进入或误入煤矿井下危险区域是煤矿安全管理的重要内容,对工作人员违规进入或是误入危险区域的行为及时监测和报警是减少事故的重要手段。针对煤矿井下危险区域的监测,提出一种基于FocalLoss的人工智能方法,实现对进入危险区域的人员及时报警。经过实验测试与分析,这种方法对人员检测的平均精度达到95.6%,检测速度达到9.9 f/s,优于对比算法,具有较高的准确性和实时性。

The illegal or accidental trespass by miners into dangerous areas of coal mines is an important content of coalmine safety management,it is an important means to reduce accidents by monitoring and alarming the behavior of miners who enter illegally or stray into dangerous areas.And thus,the monitoring of dangerous areas in coal mines is of critical important.An artificial intelligence method based on Focal Loss is proposed to realize timely alarm for people entering the dangerous area.After experimental testing and analysis,this method achieved an average accuracy of 95.6%for personnel detection,with a detection speed of 9.9 f/s,which is superior to the comparison algorithm and has high accuracy and real-time performance.