There are waves in near-Earth space that help accelerate and scatter particles in Earth’s radiation belts. These are plasma waves, and they include chorus, hiss, and lightning-generated whistlers. One of these, plasmaspheric hiss, is still not well understood—scientists aren’t sure where it comes from. It might be created locally or come from other waves that travel into the region. To help solve this mystery, we used a computer vision AI model to automatically detect and label different types of waves in satellite data from NASA’s Van Allen Probes mission. Instead of analyzing the data by hand, the model uses images built from wave features like strength, direction, and shape to tell the wave types apart. We ran this model on seven years of data to build a large database of wave events. The results show where and when each wave type is most common, and how their patterns change with solar activity. This new method can also be applied to data from other space missions, helping scientists better understand how these waves behave—and where plasmaspheric hiss might come from.