Key Insights: Developed a computer vision–powered driver monitoring system to detect unsafe behaviors in mining vehicles and send real-time alerts via a private network.

About the Client

A major mining company based in Perú, operating 24/7 in a high-risk environment where safety is a critical priority.

The Challenge

In mining operations, even brief driver distractions can lead to serious accidents. The client wanted to detect and respond immediately when unsafe behaviors occurred, such as eating while driving, using a phone, or operating without both hands on the steering wheel.

With hundreds of vehicles in continuous operation, they needed an automated, real-time solution to minimize risks, protect workers, and maintain uninterrupted production.

Marvik’s Approach

We designed and implemented a driver safety application for mining vehicles, deployed on NVIDIA Jetson Edge devices and integrated with Microsoft IoT Hub.

Key capabilities included:

• Real-time detection of unsafe behaviors using computer vision.
  
  
• Automatic alerts sent through an onsite private 4G network for immediate action.
  
  
• Integration with the company’s operational infrastructure for seamless response workflows.
  

The solution was built with Python, TensorFlow, Docker, NVIDIA Jetson, React, Azure, Microsoft IoT Hub, and Azure DevOps.

The Results & Impact

• Enabled real-time intervention when unsafe driving behaviors were detected.
  
  
• Reduced the likelihood of workplace accidents, costly shutdowns, and equipment damage.
  
  
• Supported the company’s commitment to safety in an environment where the International Labor Organization reports over 340 million workplace accidents every year.
  

Why This Matters

By bringing AI to the edge in mining operations, we turned safety monitoring from a reactive process into a proactive safeguard, helping protect people, equipment, and productivity.