People worldwide are now helping NASA classify coral reefs simply by playing a video game called NeMo Net. The game piloted nearly two years ago to map Guam’s own reefs.
At the forefront of the conversation of the recent Islandwide Beautification task force meeting were the advancements of three NASA optical technologies developed by Dr. Ved Chirayath, director of the AMES laboratory of advanced sensing for NASA in Silicon Valley.
Amongst the arising technologies included Fluidcam and fluid sensing, MiDar, and NeMo-Net. During the first airborne fluid lensing campaign, Dr. Chirayath partnered with Dr. Romina King of the University Guam and his core team to help map corals in Guam.
According to Chirayath, one of the most challenging aspects when observing shallow marine systems is looking through the surface of the water due to ocean wave distortions. Thankfully, the fluid cam instrument helps to remove those distortions.
“The fluid lensing solution is this new animation here, that removes that surface and reveal features down to the half centimeter size or so, so this is the first technology we have at NASA or in anywhere in the world, that can peel back the ocean surface and reveal a diver’s view of the ecosystem, and for corals, in particular, it’s very important to get high resolution,” Chirayath said.
The base layer of data collected from fluid cam and Midar was used to develop the video game NeMO Net, which stands for the Neural Multi-Modal Observation & Training Network for Global Coral Reef Assessment. Players can classify coral reefs and other shallow marine environments by painting 2D-3D models of coral and rate other players’ classifications.
“When you play NeMO net, you level up slowly, and you start off in Guam. So, this is what the overview of the game looks like. You can see some of the data sets there from the aircraft. We take all that input data, feed it to users that get fed and trained in the game, and then sent to our supercomputer, where it is compared to the model. If it’s acceptable quality, then it gets ingested into the model, and then we can train this machine learning system to autonomously map those corals based upon that training data.”
According to Chelsa Muna-Brecht, director of the Guam Department of Agriculture, she plans to incorporate NeMo Net into her middle school curriculum design, aiming to introduce and engage students to learn more about coral reefs.
NeMo Net is available on Windows Beta, IOS, Android, and Mac. For more details on how you can help NASA classify coral reefs, visit NeMoNet.info