Foundation models improve global critical infrastructure classification accuracy
Infra-Bench CLS: A Global, Open-Source Benchmark for Critical Infrastructure Classification with Earth Observation Foundation Models
Computer Vision and Pattern RecognitionMachine Learning
Summary
It can be hard to know where important buildings like airports and train stations are around the world, especially in developing countries. The authors created a new test called Infra-Bench CLS to see how well powerful satellite image models can recognize different types of critical infrastructure, like airports, train stations, and power stations. They found these models did much better than older methods on most facility types but struggled on some power-related facilities. This work shows these new models can help find and classify key infrastructure globally using satellite images.
What this means in practice
- •For infrastructure monitoring teams: Use foundation models to improve global mapping and classification of key infrastructure assets from satellite images for better monitoring and planning.
- •For energy sector analysts: Focus on developing or refining models to detect power sector facilities from satellite data to address current low classification accuracy in this area.
Authors
Justin Guthrie, Edward Oughton, Konrad Wessels, Matthew Rice, Isaac Corley
Abstract
Critical infrastructure location data is often incomplete and unevenly distributed globally, especially in developing regions. Earth observation foundation models are proposed as a new step in enabling us to more efficiently understand the natural and built environment, raising questions as to their effectiveness in performing challenging downstream tasks. Yet, foundation models remain largely untested for detecting and classifying the facility-scale critical infrastructure that underpins a range of important societal and economic functions. Subsequently, Infra-Bench CLS is introduced as a benchmark to test foundation models on 18,756 Sentinel-1 SAR and Sentinel-2 multispectral facility-scale critical infrastructure asset images covering seven continents and 13 infrastructure classes, with results reported for the 10 retained classes. Using linear probing and fine-tuning for two training dataset levels (1.0x and 0.3x), seven foundation models are evaluated (SatlasPretrain S2, SatlasPretrain S1, CROMA, Prithvi-EO-2.0, AlphaEarth Foundations, OlmoEarth v1.1-Base, and DINOv3 ViT-L/16). When comparing macro F1 scores to a ResNet-18 supervised baseline of 39.2 percent, the best foundation model achieved 57.9 percent, a 48 percent improvement. Top performing classes were airports (F1 85.3 percent), train stations (F1 82.1 percent), and data centers (F1 77.6 percent). By contrast, many of the power sector classes perform poorly (F1 27.5-46.2 percent). These findings suggest foundation models can enable superior critical infrastructure classification, but future work should evaluate performance on higher-resolution imagery, particularly for poorly performing sectors, such as power.