OpenData | CrowdSource data for training AI and share datasets

CrowdSource data for training AI and share datasets

vyn Excavation-safety Dataset


Videos and Images Captured by the vyn Community

The vynExcavation-safety dataset includes videos and images of excavation sites, where observations can be made on personal protection equipment, site barricading and safety signs, working conditions etc.

The data set includes 7 videos and 27 images grouped into four observation classes.

OBSERVATION CLASS NUMBER OF IMAGES
Site Barricading
12 images
Excavation Site
2 images
Signage
11 images
Working at Height
2 images

Created on: 03/09/2020, Last Modified: 05/12/2020

Download

You can download the dataset from AWS S3 using the AWS command line interface (AWS CLI)

License

Public Domain

All images included in the dataset are marked as dediated to the public domain and have been crowdsourced from the vyn community and captured using vyn (a smart video capture and analytics tool).

We strive to follow FAIR Data Principles (findable, accessible, interoperable, reusable) - to the extent possible - and will continue to promote best practices in sharing of research data.


Images Scraped from the Web

The Excavation-safety dataset includes images of excavation sites, where observations can be made on personal protection equipment, site barricading and safety signs, working conditions etc.

The data set includes 108 images grouped into five observation classes.

OBSERVATION CLASS NUMBER OF IMAGES
Worker at Height
39 images
Site Barricading & Safety Signage
31 images
PPE (Hard Hat)
15 images
Excavation Site
16 images
Signs
7 images

Created on: 29/08/2020, Last Modified: 31/08/2020

Download

You can download the dataset from AWS S3 in one of the following ways:

License

Public Domain

All images included in the dataset are marked as dedicated to the public domain by the original creators. We strive to follow FAIR Data Principles (findable, accessible, interoperable, reusable) - to the extent possible - and will continue to promote best practices in sharing of research data.

We have used the following search engines to curate the images - Creative Commons Search, Piqsels, Pixabay, Google Custom Search Engine, Google Advanced Search, and Roboflow.


Third Party Public Datasets

We have aggregated data on publically available reusable datasets (marked for Public domain) related to safety in excavation sites. We do not host this content and have included the links only for reference.

vyn does not verify that the dataset is properly marked for public domain (or CC-licensed) or that the attribution information is accurate and complete. Please follow the link to the source of the content to independently verify the license and citation requirements before reuse.

  1. Pictor-v3 dataset referenced in Deep Learning for Site Safety: Real-time Detection of Personal Protection Equipment (PPE)

  2. Hard Hat Workers Dataset shared by Northeastern University - China and published by Harvard Dataverse

  3. Hard Hat Workers Dataset provided by K&K Technologies, Inc. and available at MakeML

We have used the following search engines to curate public datasets - Google Dataset Search Engine and Roboflow.