Four computer science engineers in Colorado are looking to tackle the state’s roadkill problem. Because of the various forests, ski resorts, and mountainous roads, Colorado is well known for its disturbingly high number of wildlife-caused vehicle crashes. But this animal detector could soon help with that problem.
The detector is a small device that can detect deer in a vehicle’s path and then send a warning to the driver, allowing them to avoid the animal. The entire goal here is to reduce the number of people and animals injured by accidents and crashes caused by wildlife wandering into the street.
The animal detector is known as Project Deer. While it mostly focuses on deer at the moment, automakers and suppliers are already working on a mandate that would require vehicles to be able to detect animals and automatically brake to avoid collisions by 2029.
What makes Project Deer so intriguing, though, is that it claims to use a unique approach in its solution. Instead of just looking for deer using a camera, it also claims to use an artificial intelligence model to model thermal images of the deer. This should, hopefully, make for a much stronger recognition system.
Colorado’s Department of Transportation estimates nearly 4,000 crashes each year involve wildlife, with an estimated cost of $80 million in property damage thanks to those crashes. So if these engineers can really pull off their animal detector, then it could help tell folks when animals are at risk of being hit.
This would drastically help cut down on crashes and wildlife-related accidents, especially in places like Colorado, where so much of the roadways go through forests and mountainous areas where deer, elk, and other animals are extremely likely to cross traveler’s paths.
And considering AI continues to improve slowly by slowly—and we’ve even seen algorithms that can tell when AI is hallucinating—using AI to help pick up on the thermal images of these animals could be a huge boon in helping to detect them.
What makes this kind of animal detector especially helpful, too, is that it would be much cheaper to add as a third-party device than having to buy one of the top-end luxury vehicles that typically get this kind of new tech first. The engineers estimate it would cost roughly $1,000 to $2,000. While that isn’t cheap, it would still be much cheaper than buying a new car with the tech.
