Scholarlink

This paper presents the helmet wear identification Deep Learning model, With YOLO-ESCA, depends on an enhanced YOLOv5, to address the issue of workers improperly wearing helmets. This model could track the laborer helmet use in the real world by UAV images, other methods, and it can impulsively minimize video running detection outcomes also. The machine learning algorithm underwent training using a self-constructed dataset including 4,400 photos. An enhanced version of the original YOLOv5 is proposed to rectify its deficiencies and incorporating the systematic intersection of combination function of loss (EIOU-loss), Soft-NMS non-max suppression, and also convolutional block attention module (CBAM), along with the inclusion of small target detection layer to augment model production.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

YOLO-ESCA: A Better Real-Time Safety Helmet Standard Compliance Detection System with More Video Analytics and the Capacity to Manage Uncertainty

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Bejugam Savya,
  • N. Chandramouli,
  • Rasagnya Reddy Avala

摘要

This paper presents the helmet wear identification Deep Learning model, With YOLO-ESCA, depends on an enhanced YOLOv5, to address the issue of workers improperly wearing helmets. This model could track the laborer helmet use in the real world by UAV images, other methods, and it can impulsively minimize video running detection outcomes also. The machine learning algorithm underwent training using a self-constructed dataset including 4,400 photos. An enhanced version of the original YOLOv5 is proposed to rectify its deficiencies and incorporating the systematic intersection of combination function of loss (EIOU-loss), Soft-NMS non-max suppression, and also convolutional block attention module (CBAM), along with the inclusion of small target detection layer to augment model production.