基于改进YOLOv5的安全反光背心检测方法研究

Research on the Detection Method of Safety Reflective Vest Based on Enhanced YOLOv5

为了解决安全反光背心传统检测方法效率低下的问题,文章提出了一种基于改进YOLOv5的安全反光背心检测方法。该方法采用GhostNet对主干网络进行轻量化改进,同时引入CBAM注意力模块以提高检测精度,最后应用Python的PySide6库搭建GUI界面并进行交互设计。通过开发检测系统实现了安全反光背心的穿戴检测和检测结果展示。实验结果表明:提出的优化模型mAP@0.5值达到89.02%,具有较高的检测精度。与原YOLOv5模型相比,参数量和计算量分别下降24.8%和56.3%,显著降低了检测模型的计算资源需求,适合嵌入式平台部署检测,从而提高了该模型的实用性。研究为安全反光背心检测提供了智能方法,同时为施工安全管理数字化提供了新的思路。

To address the inefficiencies of traditional safety reflective vest detection methods,this paper proposes a safety reflective vest detection method based on enhanced YOLOv5.The method uses GhostNet for lightweight modifications of the backbone network,while integrating the CBAM attention module to enhance detection accuracy.Finally,a GUI interface is built using the PySide6 library in Python,focusing on interactive design.By developing a detection system,the wearing detection and result display of the safety reflective vest are achieved.The experimental results demonstrate that the proposed optimized model achieves an mAP@0.5 of 89.02%,indicating high detection accuracy.Compared with the original YOLOv5 model,reductions in parameter and computation are at 24.8%and 56.3%respectively,significantly decreasing computational resource requirements of detection model and making it suitable for deployment detection on embedded platform,thereby improving the practicality of the model.This study provides an intelligent method for the detection of safety reflective vest,and also offers novel insights for digital of construction safety management.