Road cameras to detect drunk drivers are closer than you think

Drinking and driving is a fatal mix. There are many roadside checks to prevent this, but it would be better to detect drunk drivers as soon as they get into their car, not after they have been driving dangerously for some time. That is why a new facial tracking system has been created that uses a camera mounted on the vehicle.

There are already several experimental systems integrated into cars to detect driver intoxication. From breathalyzers (new cars must have this pre-installed from July 2024). Others work by evaluating drivers’ driving patterns, which allows those people to drive drunk while the data is still being collected.

In response to these limitations, scientists at the Edith Cowan University from Australia set out to develop a system that would provide early intervention using conventional, inexpensive hardware.

The researchers began by having 60 volunteers use an indoor driving simulator while a conventional video camera recorded images of their face. Each person drove with three levels of alcohol intoxication: sober, low and serious.

A machine learning-based algorithm was then used to examine the recordings, looking for telltale visual features exhibited by all (or most) of the test subjects at each of the three levels. Certain facial movements, gaze direction or head position, were found to be relatively consistent indicators.

When the algorithm was then tested on more facial videos of drunk drivers, it was found to be 75% accurate at determining how intoxicated each person was. That figure should improve as the technology develops further, which could involve training it on a larger data set.

The system could be integrated into a smartphone app (although it would not be the first poison detection app), so that users could test themselves or people traveling with them.

“This research confirms that it is possible to detect levels of intoxication with a simple camera,” said lead scientist Dr Syed Zulqarnain Gilani. “The next step in our research is to define the image resolution required to employ this algorithm. If low-resolution videos are sufficient, this technology could be used for surveillance cameras installed on the roadsand the police can use it to prevent alcohol consumption. driving.»

More information: IEEE XPlore

Happy drivers

Client testimonials

Lisa D.


Theuth has completely changed how I follow automotive news. Relevant articles, thorough reviews, and an active community!

Hanna A.


I never miss a live show. It's like having front-row seats at every exhibition or race, but from my couch!

Andrew R.


As a novice in the world of motorbikes, Theuth's guides and analysis have been invaluable for my first purchase.

Ignite the drive, explore the ride