Personal Protective Equipment Detection | nvidia | AWS |

Overview

The intent of this project is to build an accurate Computer Vision Intelligence model that can detect PPE violations (Hard-hat, reflective Vests, Goggles) at construction sites and other industries to improve workplace safety and compliance with regulations

Solution Overview

null

Workplace injuries are a big source of avoidable personal loss and financial loss

null

Safety protocols like PPE requirements are routinely violated and traditional measures like safety training, manual monitoring, and safety bill-boards aren't very effective

null

Machine Learning models on TensorFlow or Yolo can be trained using annotated data using 'Transfer Learning'.

null

These models can be optimized using Hardware accelerators like Intel OpenViNo or Nvidia TensorRT to be able to run in low cost, low power SBCs (Single Board Computers)

null

These models can be hosted in the Cloud (AWS, Azure, or Google) to run on video stream from IP cameras to scale easily across any number of regular CCTV cameras.

Business Value

null

An automated, real-time Computer Vision Intelligence-based monitoring system that can detect and notify violations to improve compliance.

null

The RoI on these automated systems is very high since they use existing CCTV systems and don't require expensive hardware or upfront capital.

null

The value of lives saved, and injuries prevented through this automated PPE monitoring is beyond financial measurements.

Technology

Free Consulting