USGS EROS Interns Use AI/ML for Cloud Masking

This summer, USGS Earth Resources Observation and Science (EROS) interns Carson Price and Hunter Hagedorn worked on using artificial intelligence/machine learning techniques to create a cloud masking algorithm for Landsat 9 imagery. This graphic compares the expected result to the prediction created by their model. “On the left there is what we expect to get. That’s the actual labels that we use to train. On the right is what our model predicted—so this image is one of our best ones. It very accurately represents the one on the left and by the human eye, it’s very difficult to tell what’s actually different.”