Sustainability and Environment
The environmental industry is concerned with our natural environment and the direct and indirect impacts of human activity.
Key areas of responsibility include:
- Environmental health and biodiversity
- Depletion of natural resources
- Water conservation
- River health
- Waste management
- Pollution
- Impacts of climate change
- Recovery of mine sites.
Machine learning techniques can be used to collate and analyse environmental data. This allows environmental specialists and decision-makers to better understand, predict and manage natural resources and the effect of human activity on the environment.
There is a range of data sources that allow specialists to study the environment including:
- Satellite remote sensing
- Visible and infrared imagery (satellite, aerial and ground-level)
- Weather observations
- Topographical and land-use maps
- Field deployed sensors
- Results from the analysis of physical samples.
Machine learning techniques (specifically deep learning and computer vision) can be used to:
- Collate and analyse environmental data from multiple sources
- Transform visual images into descriptions of the world
- Identify changes over time (e.g. geo-temporal change detection)
- Model complex interactions to provide predictive and pre-emptive models
- Track and analyse the impacts of climate change
- Automatically produce geolocation data, vegetation and land usage maps
- Identify anomalies that may indicate illegal activities such as land clearing, environmental pollution and illegal water use.
-
Climate Change
Climate change is one of the biggest issues facing our world today. Climate change has the potential to impact and disrupt all areas of life including:
- Changes to rainfall patterns
- An increase in the frequency and intensity of extreme weather events (including storms, floods, heat waves, cold snaps)
- More bushfires and wildfires that are exacerbated by higher temperatures and less rainfall
- Food security (changes to rainfall patterns and an increase in the frequency and intensity of extreme weather events will impact crop yield, affecting the price and availability of food for domestic consumption and international trade)
- Health concerns including heat stress and increased spread of infectious disease
- Sea level rise.
Understanding how climate change will affect the environment and the implications on human life is of fundamental importance. Machine learning can be used to collate and analyse environmental data and to develop better climate models that more accurately predict regional trends. This information can be used to help generate reliable information that is needed by decision makers to manage the impacts of climate change.
AIML has worked with TERN and Optimatics to develop machine learning techniques to:
- Automatically determine biomass (total amount of biological material) by analysing a set of offset panoramic photographs of wooded sites
- Predict climate change induced ecosystem flux based on novel image analysis methods
- Model water supply and distribution
- Model wastewater disposal.
How machine learning can be applied will depend on the specific application and the data that is available for training. Contact AIML for more information on how machine learning can be used in your environmental application.
AI for the environment in action
New funding rewards Adelaide machine learning capability and the power of citizen science to boost bushfire management across Australia
Researchers from AIML and USC will apply data collected through a new citizen science app to help predict the likelihood of bushfires in Australia and minimise their devastating effects.
Accelerating crop farmers’ adaptation to climate change
A valuable world-first application of the University of Adelaide’s AI image-analysis technology is in the agricultural sector.