Poster
in
Affinity Workshop: WiML Workshop 1
Machine Learning API in NASA’s WorldView Satellite Image Search System
Kai Priester
SpaceML, an extension of the Frontier Development Lab, is a volunteer-based developer community that builds open-source computing tools for the bettering of space science and exploration. NASA’s WorldView Satellite Image Search System is a SpaceML project that was born out of the all-too-familiar issue of having abundant data but no efficient way of studying it. For example, if a researcher needed to find a specific type of hurricane cloud pattern, they would input one instance and the system would return all the recorded cases of that phenomena. My team was assigned to develop the system’s RESTful Application Programming Interface (API) which would handle user requests and communicate with the data store to perform image similarity searches. We have accomplished our goal of building an API that allows any user service to perform similarity searches on satellite imagery by utilizing scalable, high-performance, and developer-friendly frameworks and a machine learning integration.