Datasets:
Upload PPE.zip
Browse filesPPE Detection Dataset
A high-quality, annotated dataset for Personal Protective Equipment (PPE) detection.
This dataset is designed for training object detection models (YOLOv8/YOLOv5/Detectron2/etc.) in industrial safety monitoring, construction site compliance, and workplace hazard prevention.
📦 Dataset Summary
The dataset contains images labeled across 6 PPE classes, annotated manually using Roboflow.
Class Count
Vest 4,418
Safety Shoe 2,006
Mask 2,763
Helmet 2,703
Goggles 1,431
Gloves 2,693
Total Objects 16,014
🖼️ Sample Use Cases
Construction site safety compliance
Industrial worker monitoring
Automatic PPE checking in factories
Surveillance analytics for safety audits
Real-time alerts using edge devices (Jetson, Raspberry Pi, etc.)
📁 Dataset Structure
dataset/
├── train/
│ ├── images/
│ ├── labels/
├── valid/
│ ├── images/
│ ├── labels/
├── test/
│ ├── images/
│ ├── labels/
└── data.yaml
The dataset follows YOLO format, where each .txt file contains:
<class_id> <x_center> <y_center> <width> <height>
—all normalized between 0 and 1.
🧠 Classes
0 - Vest
1 - Safety Shoe
2 - Mask
3 - Helmet
4 - Goggles
5 - Gloves
🚀 Training Example (YOLOv8)
from ultralytics import YOLO
model = YOLO("yolov8n.pt")
model.train(
data="data.yaml",
epochs=50,
imgsz=640,
batch=16
)
📤 Exported From
Annotated on Roboflow
Exported in YOLO format
🔒 License
Recommended: CC-BY 4.0
(Users may reuse commercially or academically but must give credit.)
🙌 Author
Siddhesh D. Narnavre
If you use this dataset, please credit the author.