Hatfield is at the forefront of applying AI and computer vision techniques for environmental monitoring, including the development of novel AI and computer vision systems for our clients.
We apply our extensive knowledge of Earth observation sensors (including satellite, airborne, and terrestrial sensors), leading-edge computational methods, and partner with AI and computer vision experts to prototype, design, and deliver customized AI-driven solutions.
Our expertise includes:
- Object Detection: Identifying and locating objects of interest in satellite images using Convolutional Neural Network (CNN) approaches, such as marine mammals, exemplified by our smartWhales space-based detection system, which was published in Marine Mammal Science.
- Time Series Analysis: Using 1D CNN approaches to analyze time-series generated by remote sensing data, such as spectral vegetation indices. We apply these methods to identify forest disturbance and recovery, for example in a research project with Global Mangrove Watch.
- Semantic Segmentation: Using deep learning to categorize satellite images into different classes based on feature characteristics, such as texture, colour, and shape. We apply these methods to identify linear disturbance in forest landscapes in research to support Woodland Caribou habitat management, and to identify intertidal vegetation to support monitoring the distribution of eelgrass.
- Cloud-based Solutions: Hatfield develops cost-effective scalable cloud-based solutions that integrate AI model operation, sensor management, data visualization, and management tools, such as Hatfield’s GEOAnalytics Canada platform.
Why Hatfield?
Hatfield is a trusted advisor to government, private sector, and non-government clients on the use of remote sensing technology for environmental monitoring and management around the world. If you are keen to explore AI and deep learning, we can help you navigate the options and bring together a project design and required partnerships for success.
