New AI Powered Camera System Can Spot Reindeer Crossing Train Tracks, Day Or Night
Dec. 1, 2017 – AI tech companies Boulder AI and ENGINE27 have created a smart camera system that is capable of detecting herds of reindeer as they approach train tracks. The system, known as ReindeerCam, can warn train operators to reduce speeds when reindeer are present, reducing or eliminating the needless slaughter of countless reindeer. This system was developed in response to a recent series of high profile reindeer killings in Norway, where herds of reindeer in remote areas have been struck by train.
As reported by the The Guardian and others, the Norwegian reindeer accidents occurred after train engineers did not receive critical messages that reindeer herds from nearby ranches were in the area and that slow speeds were recommended. In all, there have been eight train incidents that killed more than 100 reindeer in the previous week.
“The needless death of any animal is a tragedy, especially considering that we have the technology to prevent this from happening,” said Darren Odom, CEO of Boulder AI. “We’re pleased to work with ENGINE27 to develop a system that uses our AI-capable camera and their proprietary machine learning algorithms to prevent unnecessary reindeer deaths.”
The new ReindeerCam system combines AI software developed by Belgian company ENGINE27 with the DNNCam™, a unique AI camera developed by Colorado company Boulder AI.
The ReindeerCam is a network of cameras that can monitor vast stretches of train tracks in remote wilderness, day or night. Each camera is capable of executing a TensorRT based machine learning algorithm that can recognize reindeer. When reindeer are detected near the tracks, an alert can be sent within 1/10th of a second.
The system is powered using solar cells paired with battery storage, and the communication between the system and the train operator can be completed via either satellite modem or cellular modem. Because the DNNCam™ hardware powering the system is both waterproof and dustproof, it can survive the harsh weather in areas where reindeer herds live.
“The DNNCam™ is the perfect compliment to our reindeer recognition algorithm. If we were using a standard machine vision camera, we’d have to setup a high speed internet connection for each camera along the railway, as most standard AI cameras are not capable of executing software,” said Geert Van Herbruggen, CEO of ENGINE27. “Without the on-board data processing capability inside each camera, the ReindeerCam system would need a high speed Internet connection to transmit data for processing. This would likely be cost prohibitive in remote areas.”
About Boulder AI
Boulder AI is a Colorado based Vision as a Service (VaaS) company headed by Darren Odom, CEO and Dan Connors, PhD, CTO. The company develops AI algorithms for a variety of applications as well as AI capable hardware, including the DNNCam™. To learn more about Boulder AI, visit https://www.boulderai.com.
ENGINE27 is an Antwerp, Belgium based technology startup founded by software experts Geert Van Herbruggen, CEO and Bart De Boeck, PhD, CTO. Their primary mission is to enable cities to gather insights by exploring data through AI frameworks. By pushing technology to the limits, ENGINE27 can resolve a variety of challenges, especially around transportation, safety, and waste management. For more information visit http://www.engine27.be/.