Date modified: 2023-12-07
Share this page:
GeoAI
A time-enabled topographic feature series.
On This Page
Learn about the GeoAI Series, the foundational geospatial data created using Artificial Intelligence (AI).
About GeoAI
The GeoAI Series is a collection of geographic features that are:
- natural or artificial, like roads, buildings, lakes and rivers, and more
- identified from various sources such as aerial or satellite images
- automatically extracted by machine learning algorithms
An important distinction of this series is that multiple data extractions can exist simultaneously over the same geographic area at multiple points in time.
Please note that due to the nature of the extraction process, most features carry minimal amount of information and will have had only a summary validation and correction process prior to being published.
Why is it Useful?
Canada is big, really big! The ability to identify trends and changes over time provides a wealth of opportunities to data users.
The typical mapping processes of strict attribution, validation, and correction for the continued maintenance of foundational datasets. Introducing the GeoAI solution!
By training and using AI to identify specific features in high-resolution images, most of the mapping process has been automated and is much more efficient. Compared to human digitization techniques, the GeoAI system is over 1000 times faster!
While the results are not always as accurate as results from manual techniques, the efficiency gained enables the ongoing creation of massive amounts of data and opens the door to a wealth of new applications.
Use Cases
Land Use Classification
Classify different land use types, such as forested areas, waterbodies, buildings, roads, etc. Being able to interpret these classes automatically can tremendously speed up the process, and accuracy, of creating base layers.
Change Detection
Detect and monitor changes in geospatial data over time. For example, changes in land use due to development, damage to infrastructure after a flood event, or reductions of vegetation coverage after a wildfire.
The provided example highlights first the difference between a typical mapping product, such as CanVec, and GeoAI, along with an analysis of urban development for Quebec City between 2006 and 2022 with a hotspot analysis on differences of detected buildings between the two dates from the GeoAI Series.
Development of High-Resolution Analysis-Ready Datasets
By providing raw geometries freely and openly, GeoAI features can be leveraged and integrated in projects aimed at the development of analysis-ready datasets which go through additional manual adjustments, attribution, and quality control and assurance processes. Furthermore, GeoAI features generated from recent source datasets can help identify areas in need of updates so that data providers can focus their efforts in areas of higher impact.
GeoAI Map and Dataset Gallery
GeoAI Data Index
Find areas that currently have GeoAI coverage and download the data.
GeoAI Feature Series
Access the GeoAI metadata and download the available datasets.
Urban Development Use Case
Explore the results of a change detection analysis in Québec City.
What's Next for GeoAI
This data series is still very new. A lot of effort was required to get here, and this is not a stopping point for the development of the GeoAI series. Firstly, we are looking to establish more collaboration to accelerate data creation from both current and historical sources. Secondly, we will continue working on improving the modelling and post process aspects of GeoAI to provide higher quality features. Check back regularly to keep informed on the latest updates and announcements about GeoAI.
Want to Collaborate?
Email geoinfo@nrcan-rncan.gc.ca and get in touch with our team.
Product Background
Since 2019, the Canada Centre for Mapping and Earth Observation (CCMEO) has invested significant strategic efforts in developing cloud-based geospatial data management systems and advanced AI.
These efforts led to the launch and open distribution of the Geo-Deep-Learning project (access the geo-deep-learning github), aimed at enabling the use of Convolution Neural Networks (CNN) with georeferenced datasets.
Natural Resources Canada later implemented image pre-processing and geo-deep-learning tools in an operational pipeline which is used to automatically process very large quantities of aerial or satellite optical imagery (the use of other data sources is currently in development) to extract foundational geospatial features.
To test out the systems, validate the results, develop the dataset structure, and highlight use cases, various pilot projects were carried out under the leadership of the Canadian Council on Geomatics (CCOG). Following the completion of these pilot projects, the series was formally approved by the CCOG as a GeoBase Initiative data series.
