Worksite Safety Monitoring Dataset

The Construction Site Safety Dataset is specifically curated for training and evaluating deep learning models in classifying workplace safety compliance on construction sites. This dataset focuses on detecting the presence of essential Personal Protective Equipment (PPE), such as helmets and reflective vests, to determine whether images meet workplace safety standards.

Dataset Overview

• Training Set: 810 images
• Validation Set: 100 images
• Test Set: 100 images
  The images were sourced from various online platforms and carefully labeled by experts to ensure high-quality annotations. This dataset is designed exclusively for image classification tasks and does not include bounding box annotations.

Safety Compliance Definition
An image is classified as safe only if every individual in the image is wearing:

1. A construction vest (orange, blue, or other reflective types).
2. A safety helmet.

If any individual in the image is missing these safety essentials, the image is classified as unsafe.

Applications

• Training AI models to classify safe vs. unsafe construction site scenarios.
• Automating workplace safety compliance monitoring using image analysis.
• Conducting research in industrial safety and workplace compliance for better enforcement of standards.

Key Features

• Binary Classification: Images are categorized as either safe or unsafe.
• Diverse Environments: Includes varied construction scenarios to improve the generalizability of machine learning models.
• Expert Labeling: All images are manually labeled for high accuracy and relevance.

Licensing
This dataset is provided under the MIT License, enabling users to freely use, modify, and distribute it with proper attribution.

Why Choose This Dataset?
This dataset is ideal for developing AI-powered safety solutions, promoting workplace compliance, and conducting research on industrial safety measures. Whether you’re an AI developer, researcher, or safety professional, this dataset offers a reliable foundation for your projects.