基于深度学习的矿井下作业人员安全帽佩戴检测

Detection of Wearing Safety Helmets for Underground Coal Mine Workers based on Deep Learning

文章基于深度学习方法来研究安全帽佩戴的检测方法,对自建安全帽数据集的预处理后,采用YOLO算法训练来获取一个最优检测模型;通过对模型测试,可以得到文章所使用的YOLO算法对矿井下安全帽佩戴检测能够达到一个比较好的检测精度,实际mAP值为90.68%,相较于其他单阶段检测算法来说有着更好的检测效果,更加符合实际运用的检测精度要求。

Wearing safety helmet in the mine is an important guarantee for the life and safety of workers in the mine,but many situations of not wearing safety helmet can be found in the mine accident;therefore,it is necessary to find a method that can effectively detect the helmets wearing situation in real time.This paper based on the deep learning method to study the detection method of helmet wearing.After preprocessing the self-built helmet data set,use the YOLO algorithm training to obtain an optimal detection model;Through the model test,it can be obtained that the YOLO algorithm used in this paper can achieve a relatively good detection accuracy for the detection of the helmet wearing in the mine.The actual mAP value is 90.68%,which has a better detection effect compared with other single-stage detection algorithms and is more in line with the actual detection accuracy requirements.