marți, 3 mai 2022

Neural network can read tree heights from satellite images





Neural network can read tree heights from satellite images


Researchers at ETH Zurich have created a high-resolution global vegetation height map from satelite-images. It has been designed to provide key information for fighting the climate changes and species extinction, as well as for sustainable regional development planning.

EcoVision Lab is a team of researchers in the ETH Zurich Department of Civil, Environmental and Geomatic Engineering which is specialised in analysing and preparing precisely this kind of environmental data. The team was founded in 2017 by professor Konrad Schindler and professor Jan Dirk Wegner. Nico Lang, a member of the team, in his doctoral thesis, developed an approach for deriving vegetation height from optical sattelites.




 

How does it work?

 

It is trained using  a laser data scanning from space. First of all it captures images from the 2 Copernicus Sentinel-2 Satellites operated by the European Space Agency. These 2 satellites capture every location on Earth every five days with a resolution of 10x10 meters per pixel. But because for training the algorithm must also have acces to the correct answer, it was trained on datasets provided by the tree height derived from space laser measurments from NASA`s GEDI Mission. GEDI is a NASA mission to measure how deforestation has contributed to atmospheric  CO2 concentrations. A full-wavehorm LIDAR was attached to the International Space Station to provide the first global, high resolution observations of forest vertical structure, but GEDI was designed to capture data between 51.6 North and 51.6 South, covering only 4 percent of the world surface. Trained on these datasets, the algorithm can automatically estimate the vegetation height from the more than 250,000 images (some 160 terabytes of data) needed for the global map.

 

To allow this research to continue, the map and its source code will be made publicly accessible

 

 

 

Sources:

https://nlang.users.earthengine.app/view/global-canopy-height-2020

 https://ethz.ch/en/news-and-events/eth-news/news/2022/04/neural-network-can-read-tree-heights-from-satellite-images.html

https://techxplore.com/news/2022-04-neural-network-tree-heights-satellite.html

https://www.academicgates.com/news/story/neural-network-can-read-tree-heights-from-satellite-images/11751

https://gedi.umd.edu/


Niciun comentariu:

Trimiteți un comentariu

Neural network can read tree heights from satellite images

Neural network can read tree heights from satellite images Researchers at ETH Zurich have created a high-resolution global vegetation height...