基于YOLOv8改进的安全帽佩戴检测算法研究  

Research on improved helmet wear detection algorithm based on YOLOv8

作　　者：任海 王涛 贾利云[1] REN Hai;WANG Tao;JIA Liyun(Hebei University of Architecture,Zhangjiakou,Hebei 075000)

机构地区：[1]河北建筑工程学院,河北张家口075000

出　　处：《河北建筑工程学院学报》2024年第3期222-228,共7页Journal of Hebei Institute of Architecture and Civil Engineering

摘　　要：安全帽的佩戴对建筑工地和工厂的工人来说至关重要,佩戴安全帽是施工场地的重要防护措施。如何有效地监测和确保安全帽的佩戴一直是企业监管的一大挑战。基于这一问题,在YOLOv8算法的基础上,通过改进YOLOv8网络的主干网络和头部网络来提高检测精度。实验结果表明,改进的算法检测准确率可达74.4%,与YOLOv8相比提升了0.4个百分点。这个改进对于提高安全帽检测的准确性具有重要意义。Wearing a safety helmet is crucial for construction site and factory workers.Wearing a safety helmet is an important protective measure at the construction site.How to effectively monitor and ensure the wearing of safety helmets has always been a major challenge for corporate supervision.First,this paper improves the detection accuracy based on the YOLO v8algorithm by replacing the backbone network and head network of the YOLOv8network.Experimental results show that the detection accuracy of the improved algorithm reaches 74.4%,Compared to YOLOv8,there has been an improvement of 0.4percentage points.This improvement is used to improve the accuracy of helmet detection and is of great significance meaning.

关 键 词：深度学习 目标检测 安全帽检测 YOLOv8 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]