【摘要】 针对施工人员未佩戴安全帽的现象,为了实时检测施工人员是否佩戴安全帽,在YOLO v3算法的基础上,通过深度可分离卷积对Darknet网络结构进行改进,得到新的特征提取网络D-Darknet,减少了模型参数量,降低了模型计算量.试验结果表明,改进后的YOLO v3算法在保证检测准确率的同时,检测速度提高了近2倍.更多还原

【Abstract】 Safety helmets play an important role in ensuring the safety of construction personnel, but it often occurs that construction personnel do not wear safety helmets. On the basis of YOLO v3 algorithm, this paper improves the structure of Darknet network through deep separable convolution network, obtains a new feature extraction network D-Darknet, reduces the number of model parameters and the computational cost of the model. The experimental results show that the improved YOLO v3 algorithm improves the detection speed by nearly 2 times while ensuring the detection accuracy.更多还原