AI Ski Lift Surveillance

The Future of Ski Lift Surveillance

At the heart of many ski resorts are the ski lifts, transporting skiers and snowboarders to the top of the slopes. However, these lifts are not without their challenges. Human limitations, such as the varying levels of alertness and reaction times of lift attendants, can sometimes lead to oversights. PyxisAi, a groundbreaking surveillance system specifically designed for ski lifts, can detect potential risks a staggering six times faster than even the most alert human. This offers ski resorts an "always-on" extra set of eyes, ensuring critical lift situations are promptly addressed.

How PyxisAi works

The Pyxis AI surveillance system is like an extra vigilant lift attendant, but one that never tires or gets distracted. It continuously watches guests as they load and unload ski lifts. Using advanced AI and computer vision, it can instantly alert staff when a guest might be in danger of failing to load or unload properly. The goal is simple but profound: to dramatically improve ski lift staff reaction times, thereby enhancing both lift operations and guest safety.

Here is a link to the PyxisAI video for more information on how the system works.

The story behind the startup

Scott Queen, one of the startup founders, conceived the idea after witnessing a young skier, a little girl, struggling to unload from the lift. The lift attendant, engrossed in a book, remained oblivious until Scott intervened. This close call convinced Scott of the need for a vigilant system that could preempt such situations.

Proof of Concept

Queen tested out his idea by videoing skiers getting off chairlifts with his phone and using it to teach his model to categorise them into normal usage or potential problems like a skier falling, not unloading or dropping a ski pole. This proof of concept demonstrated that the system would work and more intensive model training followed.

AI monitoring

The PyxisAi system uses video surveillance cameras to monitor lifts and links this to a server to process the video and pass it into the AI model to classify.  Depending on how the AI model classifies what it sees, it will notify the lift operator so they can stop the lift if necessary. Interestingly the AI model response time is around 40 milliseconds, compared with a typical human response time from visualising and reacting to an incident of 250 milliseconds, which works out around 6 x faster.

Challenges

Having worked as a ski instructor in Coronet Peak, Queenstown NZ for a couple of seasons, I can see this is an extremely useful surveillance system. The challenge is for the AI to be trained sufficiently to differentiate the multitude of ways small children can move on a chairlift without being unsafe to identifying dangerous situations which need intervention. Lift workers in NZ were always super vigilant and although an extra pair of eyes is never a bad idea, I cannot remember any situation when a liftie was ever distracted.

PyxisAI today

In terms of achievements, PyxisAi is already making waves in the industry. In December 2021, the system saw its first commercial installation at the renowned Winter Park Resort in Colorado. The company is named after PyxisAi, a constellation that symbolizes a mariner’s compass, guiding and protecting those on their journey.

In a world where technology is constantly pushing boundaries, PyxisAi is a great example of how AI can be harnessed for safety and efficiency. As ski enthusiasts head up the mountains, they can do so with the added assurance that PyxisAi is watching over them.

Read more about this technology in National Ski Areas Association Journal Winter 2020

‍