NASA needs the public’s help to identify and categorize the world’s coral reefs. The space agency announced its new citizen scientist project on Thursday, explaining that while it is able to use its technology to get 3D imagery of the ocean’s coral reefs, actually identifying them and categorizing them requires a bit of human effort and something as simple as a tablet.
NASA’s Ames Research Center in California has spent the last several years creating instruments capable of spying coral reefs in the ocean despite the optical distortion caused by water. The result is 3D images of the seabed from around the world, some images of which include coral reefs and marine plants.
Here to complement that effort is NASA’s new “NeMO-Net,” a game designed to let citizen scientists help identify and classify each captured coral reef. The game’s name refers to the neural network behind it, the Neural Multi-Modal Observation and Training Network. According to the space agency, even young children can play its new game and help sort the ocean data in the process.
By playing this game, the public helps train the NeMO-NET supercomputer to recognize coral reefs from the surrounding elements. Over time, the machine learns what each type of coral reef is from the players’ classifications, paving the way for it to eventually make these determinations without human help.
The game involves low-resolution imagery, but the supercomputer itself will be used to classify coral reefs at an ‘unprecedented resolution’ after the training process. The resulting ocean floor maps will be used by scientists to help conservationists preserve the corals, which are threatened by climate change, pollution, and more.