Ariane Roberge

Université de Montréal
M.Sc. candidate

Supervisor: Etienne Laliberté
Erik Veneklaas, University of Western Australia
Start: 2022-09-01
Personal page


Mapping shrublands from drone imagery in Western Australia
The Southwest Australian Floristic Region (SWAFR) is defined as a biodiversity hotspot along with 35 other areas around the world. Biodiversity hotspots are regions where we find the greatest amount of endemic species that are being threatened, especially by anthropic activities (Myers and al., 2000). The SWAFR’s biodiversity is mostly threatened by massive habitat loss and fragmentation due to agriculture and mining expansion, but also by root diseases and the introduction of invasive species (Coates and al. 2014). Given the current climate crisis, it is becoming critical to develop much more time-efficient methods to study ecosystems. It is in this context that my research project becomes relevant. My project aims to develop remote sensing approaches using UAVs coupled with artificial intelligence (AI) (computer vision) to monitor post-mining sites in hyperdiverse kwongan scrub vegetation of Western Australia. The details of the project are not yet established, but we would like to look at the 3 key species that are planted at the restoration sites. Building an AI model capable of identifying all plant species present in this biodiversity hotspot would be nearly impossible given the number of species present. However, automated identification of the three key species would allow accurate monitoring of plant mortality rates in order to be able to compare the effectiveness of different restoration strategies.


drone imagery, ecology, cartographie, mapping, Écologie, imagerie par drone, AI, remote sensing, télédétection