Journal of Engineering, Project, and Production Management, 2026, 16(4), 2026-147
TransSPCW-Net: A Personal Protective Equipment Detection Model
Undergraduate Student, School of Automation, Beijing Institute of Technology, Beijing 102488, P. R. China, E-mail:dajinhao.da@outlook.com
Project Management
Received January 28, 2026; revised April 15, 2026; accepted May 10, 2025
Available online May 29, 2026
Abstract: Construction site safety is an important issue for protecting workers health and reducing safety accidents. However, the traditional manual monitoring method often suffers from incomplete coverage and delayed response. To better overcome these problems, it is crucial to adopt real-time detection of both humans and Personal Protective Equipment (PPE) using intelligent recognition systems. To overcome these problems, this paper designs a model integrating the C3TR module, the SPP module, the CBAM attention mechanism, and the Wise-IoU (WIoU) loss function based on YOLOv8, named TransSPCW-Net. The TransSPCW-Net integrates these key innovations: the C3 Transformer (C3TR) module is used in conjunction with Spatial Pyramid Pooling (SPP). The joint module enhances the ability to extract features of all objections and reduce the impact of occlusion. The Convolutional Block Attention Module (CBAM) attention mechanism is designed to reduce background noise and focus more attention on the human body. Then the WIoU loss function is adopted to enhance the attention to the ordinary quality sample, thereby strengthening the precision of the prediction box. Experimental results show the improvements. The precision of the new model increased 1.377% from 94.932%, and recall increased 2.686% from 90.976%. Furthermore, it means Average Precision (mAP) increased 1.45% from 96.073%. These results demonstrate the model’s robust performance in detecting PPE and humans in complex construction environments.
Keywords: Construction site safety, TransSPCW-Net, PPE detection, CBAM, WIoU-v3, C3TR.
Copyright © Journal of Engineering, Project, and Production Management (EPPM-Journal).
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.
Requests for reprints and permissions at eppm.journal@gmail.com.
Citation: Da, J. (2026). TransSPCW-Net: A Personal Protective Equipment Detection Model. Journal of Engineering, Project, and Production Management, 16(4), 2026-147.
