摘要
佩戴安全帽可以有效保障建筑施工人员的人身安全。针对目前安全帽检测算法检测效率低,漏检率高的问题,本文提出了一种基于改进YOLOv5网络模型的安全帽检测算法。首先,在YOLOv5的骨干网络中插入混合注意力模块,降低特征提取网络的数据维度,使网络关注于图片中安全帽特定区域,提高网络的安全帽检测性能;然后,使用双向特征金字塔结构代替原网络中的特征金字塔结构,融合不同层级的特征并保留特征图中浅层信息,提高模型的计算效率;最后,使用EIoU作为网络的损失函数,提高改进模型的识别准确率。实验结果表明,改进模型在扩充的安全帽数据集上检测精度达到83.1%,每秒检测速度为48.1,相比于原始算法模型精度提高了4.5%,检测速度提高了1.5帧。
Wearing a safety helmet can effectively ensure the personal safety of construction workers.Aiming at the problems of low detection efficiency and high missed detection rate of the current helmet detection algorithm,this paper proposes a helmet detection algorithm based on the improved YOLOv5 network model.First,a hybrid attention module is inserted into the backbone network of YOLOv5 to reduce the data dimension of the feature extraction network,so that the network can focus on the specific area of the helmet in the image,and improve the performance of the network′s helmet detection;then,the bidirectional feature pyramid structure is used to replace the original network.The feature pyramid structure fuses the features of different levels and retains the shallow information in the feature map to improve the computational efficiency of the model;finally,EIoU is used as the loss function of the network to improve the recognition accuracy of the improved model.The experimental results show that the detection accuracy of the improved model on the expanded helmet dataset reaches 83.1%and the detection speed per second is 48.1.Compared with the original algorithm,the model accuracy is increased by 4.5%,and the detection speed is increased by 1.5 frames.
作者
王子元
王国中
顾嘉城
WANG Ziyuan;WANG Guozhong;GU Jiacheng(School of Electronic and Electrical Engineering,Shanghai University of Engineering Science,Shanghai 201620,China)
出处
《智能计算机与应用》
2023年第8期169-174,共6页
Intelligent Computer and Applications
基金
国家重点研发计划(2019YFB1802702)。
