Model Overview

This model analyzes construction site imagery to automatically detect and identify personal protective equipment (PPE) being used.

Personal Protective Equipment (PPE) Detection

The Personal Protective Equipment (PPE) Detection model analyzes aerial construction site images and quickly detects and labels PPE used on site. Identifying unsafe behaviors among construction workers is critical for enhancing site safety. This model provides key insight to owners, contractors, and construction site managers, allowing them to prioritize safety and conduct regular jobsite safety training.

Key features include:

  • Fast analysis – Designed to ingest and analyze large sets of data, this model can identify and locate multiple forms of PPE in a dataset, including head gear, gloves, boots, safety goggles, safety vests, fall protection harnesses, and much more.
  • Confidence Level Settings – We give you the flexibility of choosing the confidence level you’d like the model to use for the PPE detection.
  • Annotated images and detailed reports – Bounding boxes will appear around detected PPE, enabling you to quickly locate where they are within the image. The annotated images are viewable directly from our platform, for the utmost convenience, or you can download them for later viewing. The model also generates a detailed report that lists each detection along with their confidence level for further analysis.

This model detects:

  • People
  • Head Gear (Type 1 and Type 2)
  • Boots
  • Gloves
  • Full Body Harness
  • Safety Vest (Class 2 and 3)
  • Self-retracting Lifeline
  • Anchorage Point
  • Fall Limiter
Required Inputs

PNG files

JPEG files

Note: Oblique to level camera angles are required.

Expected Output

Annotated images

A summary report in XLSX format

Cookie Statement

We collect cookies to improve your experience with BRYX.