ML21.Traffic
A nature preserve suffers from chemical waste dumping. They had a year’s worth of license plate recognition data, which recorded which vehicles go where in the park at what time. The goal was simple: catch the car dumping the trash.
Detecting abnormalities
From amongst all the detected license plate, our software was automatically able to categorize cars into several categories:
But one car stood out! As early as the first pass over the data, our detection software detected a single vehicle that:
Indeed, the deviant car was responsible for dumping the garbage!
Looking for the only car dumping chemical waste from over 170,000 data points from a license plate recognition system? In the 2017 IEEE VAST Challenge, co-founder Bram Cappers single-handedly outperformed more than 50 six-person industry teams working on a case for more than three months.
Results
Where more than 50 industry teams, consisting of up to 6 people, took three months to complete this challenge. CodeNext21 used our state-of-the-art anomaly detection software to find the outliers within hours!
Although we can proudly say that we achieved first place in this challenge, unfortunately it was an offline challenge. But imagine if we could realize this business case: a park ranger could have been deployed at the vehicle even before they could dump their waste, preventing large-scale damage to the local natural ecosystem.
