基于深度学习的煤矿防尘口罩佩戴检测

Detection of Wearing of Mask for Dust Suppression in Mine Based on Deep Learning

煤矿井下存在大量有害气体及粉尘,井下工作人员长期吸入有害气体或粉尘后会对身体造成严重危害.提出一种基于深度学习的煤矿防尘口罩佩戴检测方法,对数据集进行预处理后,采用YOLO算法进行迭代训练,得到最优权重.通过与其他算法进行对比,本文算法mAP值为92.5%,检测速度为12 ms,相比于其他目标识别算法检测精度更高、速度更快,表明该算法能够更好地识别防尘口罩,满足实际应用需求.

There are a lot of harmful gases and dust in the coal mine.The long-term inhalation of harmful gases or dust by underground workers will cause serious harm to their bodies.In this paper,a detection method of wearing of mask for dust suppression in mine based on deep learn⁃ing is proposed.After preprocessing the data set,the YOLO algorithm is used for iterative train⁃ing to obtain the optimal weight.Compared with other algorithms,the mAP value of this algo⁃rithm is 92.5%and the detection speed is 12ms.Compared with other target recognition algo⁃rithms,this algorithm has higher detection accuracy and faster speed,which shows that this al⁃gorithm can better identify mask for dust suppression and meet the needs of practical applica⁃tion.