Using very high resolution (< 3 m pixel) satellite imagery to map individual trees across continents: Lessons learnt from Africa and Europe and current progress of OzTreeMap

Pricing

• Free

Dates and Times

Showing 2 of 2 dates for this event.
Dates available from 05 Dec 2023 until 05 Dec 2023

Contact

Tim McVicar

More information

Florian Reiner will present on the mapping of individual trees at the continental scale (Africa and Europe). He is a visiting scientist from the University of Copenhagen, where he is part of Assoc. Prof. Martin Brandt's lab which specialises in mapping tree cover, canopy height, biomass and carbon at the individual tree level over large areas. Their team has developed new methods for mapping single trees with deep learning and very high resolution satellite data, resulting in research outputs such as the mapping of 15 billion individual trees and their carbon stocks across the Sahel using WorldView imagery (Nature 2023, https://doi.org/10.1038/s41586-022-05653-6 ). Florian has expanded this work to the continental scale using PlanetScope data, developing a tree cover map at 1 m for all Africa (30,370,000 km2) which includes both forest and non-forest trees (Nature Communications 2023, https://doi.org/10.1038/s41467-023-37880-4 ).

Dr Nicolas Pucino (Google Scholar) will introduce the current state of OzTreeMap, which extends for all Australia what University of Copenhagen and NASA researchers have implemented in Africa and Europe. He will explain the challenges encountered and solved to date and discuss future OzTreeMap directions. OzTreeMap extends previous research by incorporating point-cloud data acquired with LiDAR and aerial photogrammetry instruments to create training data (canopy height models and crown delineations) for the deep learning models. These point-cloud data are primarily from Geoscience Australia’s ELVIS platform.

Nick is an Australian National University PostDoc mainly funded by CSIRO. He recently obtained a PhD in remote sensing of coastal morphodynamics from Deakin University, Victoria, where he mapped sediment dynamics using deep learning models at several geographical scales. He has now enthusiastically moved to studying terrestrial ecosystems where is applying his geospatial analysis, remote sensing and deep learning skills on OzTreeMap.