Top of page

Student/Staff Portal
Global Site Navigation

School of Science

Local Section Navigation
You are here: Main Content

Related Contents

Assessing Carbon Storage and Fire Risk in Peri-Urban Western Australia Using Drone and Satellite Imagery with Machine Learning

Project Outline

This research explores how vegetation in peri-urban Western Australia contributes to both carbon storage and bushfire risk. Utilizing drone and satellite imagery alongside machine learning, the study develops spatial models for estimating biomass and fire risk. The project addresses a critical gap in understanding the dual ecological and hazard roles of peri-urban vegetation and supports better decision-making for land management and climate resilience.

Urban expansion into vegetated landscapes in fire-prone areas such as Western Australia has created complex land management challenges. Peri-urban vegetation is both a critical carbon sink and a potential fuel source for bushfires. Understanding how to manage these areas to balance ecological services and fire safety is vital. This report presents a methodology integrating high-resolution remote sensing and AI-based analytics to assess carbon storage and wildfire susceptibility in peri-urban environments.

While much research has been conducted on either fire risk or carbon storage separately, few studies integrate both in a spatially explicit, high-resolution framework. There is limited application of drones and advanced machine learning for combined ecological and hazard mapping in peri-urban Australian contexts. This project fills a gap by fusing multiple data sources and analytical techniques to develop practical land management tools.

The research will aim to

  • Quantify carbon storage in peri-urban vegetation.
  • Assess spatial variation in bushfire risk.
  • Integrate multisource remote sensing with machine learning.
  • Support evidence-based land and emergency management.

The study will focus on  peri-urban zones around Perth. The research will use Drone surveys using multispectral cameras will be conducted alongside the acquisition of satellite imagery (Sentinel-2, Landsat). Ground truth data on vegetation types and biomass will be collected. Machine learning models, such as Random Forest and XGBoost, will be trained to estimate carbon storage and fire risk. Output layers will be validated and mapped in a GIS environment.

The research will engage with Government agencies such as WA Department of Fire and Emergency Services (DFES), Department of Biodiversity, Conservation and Attractions (DBCA), and local councils for data access and application of outputs.

It’s expected that this research will benefit local agencies  by producing actionable maps and data tools to support fire risk mitigation and climate-smart vegetation planning. The interdisciplinary approach bridges environmental science, remote sensing, and machine learning to address a timely and regionally relevant challenge.

Essential Skills

  • Skills in Database Programming, Artificial Intelligent systems and
  • Advanced Computer Programming
  • Skills in Image Processing techniques
  • Advanced programming skills in R, python or other languages
  • Good oral and report writing skills
  • Able to work in team environment
  • Good project management skills

Desired Skills

  • Experience in Environmental or Geospatial sciences
  • Drivers License
  • Drone pilot license

Project Area: Computer Science; Biological and Environmental Science

Supervisor(s): Dr Leisa Armstrong and Professor Dean Diepeveen

Project Level: Masters / PhD

Funding: Applicant should apply for ECU HDR or RTP Scholarship

Start date: Any

Skip to top of page