Detecting fatigue and distraction even through a mask

By Josh Savit, Product Manager-OAS

The global pandemic has changed the landscape of the mining industry, increasing the stress in an already stressful environment. Many mines are mandating that their employees wear face masks; even vehicle operators. While masks reduce the risk of COVID-19 transmission, they also present a new challenge for safety, especially for fatigue and distraction detection.

A system ahead of its time

In 2019, Hexagon launched HxGN MineProtect Operator Alertness System Light Vehicle (OAS-LV), addressing the need for a fatigue and distraction solution for pickups, busses, and other vehicles outside of those classified off-highway. It’s based on the same proven technology used in OAS-HV, which protects operators of haul trucks. OAS features the most advanced fatigue and distraction algorithm. A client requested that the solution identify fatigue and alertness events while the operator wears a dust mask, or other face cover. Hexagon saw the value and included it as part of the OAS 6.1 release for both the heavy and light vehicle versions.

The challenge

Most face trackers rely on limited data points, requiring the full face to be visible. A face mask therefore renders face trackers inoperable. The ability to focus is stuck in a continuous loop, trying to establish the needed points. Events generated are false positives because the system can only lock on to a greatly reduced number of data points.

The solution

OAS leads the industry in computer vision and expands detection beyond just eye closure to include distraction, inattentiveness, daydreaming – all major causes of risk to the operator. The OAS 6.1 algorithm utilizes dynamic tracking, constantly readjusting and mapping the face with every movement, removing the data point requirement. This is possible due to the OAS machine learning library. With more than 400,000 images to reference, the algorithm can create millions of unique models. By modeling beyond a standard dust mask to include everything from turtlenecks to respirators, the library grows. This allows Hexagon’s face tracker to adapt and focus on key facial components; eyes and forehead, for example.

The result

Enhancements to the algorithm combined with the continual growth of the machine learning capabilities mean operators can wear face masks while still being monitored for signs of fatigue and distraction. Subsequently, overall performance of the system has improved. Increased accuracy and ability to track the face – partially concealed by a mask or not – reinforces trust in OAS by operators and operations.

Hexagon’s Commitment to Safety 

The MineProtect portfolio is committed to the safety of the men and women who put in the long hours and hard work. This commitment aligns with our customers’ needs and drives us to continually improve our systems and to not accept the status quo. We are ready to meet the ever-changing needs of the mine safety environment.

  • Recent Posts