Build a Safety Gear Detector

For people working in hazardous conditions, wearing appropriate safety gear is critical. This solution observes workers as they pass in front of a camera, identifies them using facial recognition, and determines if they have adequate safety protection.

Target Operating System Ubuntu* 16.04 LTS
Time to Complete 45 minutes

GitHub* (C++)   GitHub (Python*)

What You Will Learn

This application uses the inference engine included in the Intel® Distribution of OpenVINO™ toolkit for facial recognition and identification of authorized personnel and their use of appropriate safety protection.

Gain insight into the following solutions:

  • Computer vision applications for IoT
  • Inference to detect people and objects
  • Industrial market IoT

Learn to build and run an application with these capabilities:

Monitor restricted work zones for unauthorized personnel.
Determine if appropriate safety gear equipment is being worn by authorized personnel.
Record and report the number of safety violations detected.

How It Works

Using a combination of different computer vision techniques, this application detects whether workers within a video frame are wearing safety gear, and if not, applies a red bounding box over the image of the worker. The video displays the total number of workers without safety gear.

  1. The inference engine in the Intel Distribution of OpenVINO toolkit uses a trained neural network to process the video and detect people in the video frame. If detected, the image of the person is cropped.
  2. Within the cropped image, traditional computer vision techniques detect the presence of specific colors, namely the yellow of a hard hat and the orange of a safety vest.
    • If appropriate wear is recognized, a green bounding box highlights the worker. 
    • If safety gear is missing, a red bounding box highlights the worker and the application counts them.
  3. The final image frame displays workers not wearing safety gear and the overall statistics in the bottom left corner.

flow chart graphic of how the gaze monitor application works