UNSW’s REMAP ecosystem mapping tool uses machine learning and archival Landsat data to map ecosystem change.
REMAP, developed by UNSW scientists, is designed to facilitate fast analyses of Landsat data.
The app uses machine learning to develop a map: users train REMAP to classify specific ecosystems types by identifying a few pixels from Google Earth, or by uploading their own field data.
From that little bit of data, REMAP can apply that information to recognise ecosystems in a selected area and then returns results that let users know the final extent of the ecosystem and how much it has changed over time.
Project lead, Dr. Nicholas Murray from UNSW’s School of Biological Earth and Environmental Sciences, said the aim was to democratise potential for ecosystem monitoring.
“In the past it has been a technical process to produce high-quality maps suitable for tracking environmental change such as deforestation and ecosystem loss. It really has been sitting in the hands of experts,” he said.
“We aimed to remove the technical steps required to monitor ecosystems from space. Now, if I want to map an area the size of Sydney using satellite data that would require a fraction of the time in REMAP.”
“We wanted to empower people to map just how much the ecosystems around them have been changing,” he said.
UNSW says that REMAP is now being used to monitor ecosystems in over 140 countries worldwide.
Stay up to date by getting stories like this delivered to your mailbox.
Sign up to receive our free weekly Spatial Source newsletter.