This webinar is for ALGIM Council / CCO members only.
Urban forests—the network of trees, vegetation, and green spaces—are vital for healthier, more liveable, and more resilient communities. Recognising this, Auckland Council developed the Urban Ngahere (Forest) Strategy in 2019, built on three pillars: knowing, growing, and protecting. A key challenge, however, is consistently monitoring the urban forest in a rapidly intensifying environment.
The Auckland Council Ngahere AI project addresses this challenge by leveraging Geospatial Artificial Intelligence (GeoAI) and high-resolution aerial imagery. A GeoAI model trained to mimic human vision enables automated vegetation mapping across multiple aerial datasets. Applied to imagery from 2017 and 2024/25, the model produces detailed vegetation maps at 7.5 cm resolution, providing new insights into how Auckland’s urban forest is changing over time.
In this webinar, we will share how this GeoAI project was delivered—from project governance and collaboration to the technical steps involved in building the model. We will walk through the iterative development process, including how subject matter expertise and ongoing validation were used to continuously improve model performance and accuracy.
The project establishes a scalable and reusable approach that can be extended beyond forest mapping to other domains. It has also sparked broader interest in GeoAI within the organisation, opening up opportunities for future innovation. By continuing to leverage high-quality imagery and AI, Auckland Council is building a stronger foundation for data-driven decision-making and delivering better environmental outcomes for its communities.
About the Presenter:
Dr Joe Zhao
Senior Geospatial Specialist
Auckland Council
Joe Zhao is a Senior Geospatial Specialist at Auckland Council, where he leads innovative projects that apply advanced geospatial technologies to support urban management and decision-making.
He managed and delivered the three-year Auckland Council Ngahere GeoAI project, overseeing its end-to-end development, including the design, training, and implementation of the GeoAI model for urban vegetation mapping.
Joe holds a PhD in Forestry with a strong focus on computer vision and machine learning. His research background spans remote sensing, forestry, and the application of AI/ML algorithms to extract insights from geospatial data. He brings a unique combination of domain expertise and technical capability, bridging environmental science and cutting-edge AI solutions.
.png)