|
Briefing
|
- MistNet – Researchers from University of Massachusetts Amherst, Cornell Lab of Ornithology, and others developed machine learning system that can automate processing of massive radar datasets to study migration of birds
- Dataset – Analyzed and mapped radar data for past 24 years from 1998 to 2018 from 143 radar stations in U.S., estimate of over 200 million images and hundreds of terabytes of data
- Technology – Based on neural networks for images, while incorporating architecture components uniquely built for radar data analysis
- Applications – Include science and conservation, such as studying how climate change impacted historical timing of bird migration, estimating flying velocity and traffic rates of migrating birds, as well as processing weather radar data and data from citizen science projects like eBird, animal tracking devices and earth observation instruments
- Publication – Research supported by National Science Foundation is published in scientific journal Methods in Ecology and Evolution
|
