The South African Polar Research Infrastructure (SAPRI), one of 13 research infrastructures developed by the Department of Science and Innovation (DSI), was established in 2021, with the final contract between the DSI and the National Research Foundation signed in November.
Title: Machine learning for vegetation monitoring
Presenter: Dr Glenn Moncrieff (Fynbos Node)
The simultaneous rise of computing power, explosion of ecological data, and advancement of machine learning have created a wealth of new opportunities for monitoring vegetation at very large scales with low latency. I highlight research conducted by SAEON’s fynbos node combining machine learning and data on vegetation distribution, abundance and function to improve vegetation monitoring capabilities. These examples include using object detection to locate rare tree species in aerial imagery, mapping land cover change in near-realtime from satellite imagery, and automated detection of hydraulic failure in plants under water stress. Our rich and unique ecosystems and large archive of freely available environmental datasets that are well suited to the application of machine learning methods present a great strategic opportunity to conduct cutting-edge research that is of societal relevance.