Datasets
Standard Dataset
PPE Usage Dataset
- Citation Author(s):
- Submitted by:
- Marcos Eduardo ...
- Last updated:
- Fri, 12/13/2024 - 13:48
- DOI:
- 10.21227/d8ne-tf43
- Data Format:
- License:
- Categories:
- Keywords:
Abstract
PPE Usage Dataset
This repository provides the Personal Protective Equipment (PPE) Usage Dataset, designed for training deep neural networks (DNNs). The dataset was collected using the EFR32MG24 microcontroller and the ICM-20689 inertial measurement unit, which features a 3-axis gyroscope and a 3-axis accelerometer.
The dataset includes data for four types of PPE: helmet, shirt, pants, and boots, categorized into three activity classes: carrying, still, and wearing.
We obtained unprocessed accelerometer and gyroscope data for each PPE during utilization under the three identified conditions: carrying, still, and wearing. The data were sampled at a 25Hz rate, which follows the rate used in other datasets, such as the so-called ``Daily & Sports Activities''.
For the ``still'' class, data were collected in various positions, including facing down, facing up, hanging, and other orientations. For the ``carrying'' class, data were gathered in different positions while carrying the PPEs, such as holding them in the hands or inside a bag, during activities like walking, standing, and going up and down stairs. For the ``wearing'' class, data were obtained during activities such as walking at different speeds, sitting, lying down in various positions, going up and down stairs, kneeling, and standing.