180-5 Artificial Intelligence Advancements & Unmanned Aerial Systems Revolutionizing Geological Investigations

Session: Geoscience Outreach Efforts to Broaden Participation (Posters)

Presenting Author:

Authors:

Abstract:

Advancements in artificial intelligence and unmanned aerial systems (UAS) are poised to revolutionize geological investigations in connection with the development, upgrading and maintenance of aviation-related infrastructure. This research presents an artificial intelligence framework that fuses LiDAR, hyperspectral imaging, and ground-penetrating radar data (acquired by fixed-wing and rotary UAS) to generate continuous 3D models of sub-runway bedrock/soil conditions, compaction, and groundwater flow and enable predictive maintenance and hazard mitigation. Model attributes derived from MIT’s (Massachusetts Institute of Technology) geodynamic resources depend on site-specific regional geology, sedimentary processes, and soil mechanics. Machine learning techniques include convolutional neural networks for surface anomaly detection and gradient-boosting regression trees for subsurface characterization. These techniques have been adapted from state-of-the-art methodologies in MIT's machine learning curriculum. Remote sensing approaches combine active (LiDAR, radar) and passive (multispectral) modalities for near real-time identification of runway surface defects such as cracks; rutting; and frost heave, with detection latencies under two seconds (MIT, 2008). Data-fusion pipelines merge geophysical inversion outputs with live sensor feeds into interactive operations dashboards, facilitating proactive scheduling/initiating of geotechnical investigations; maintenance interventions; and climate-resilience planning. Preliminary field trials in diverse climatic and regional geology-based environments demonstrate that the integrated artificial intelligence framework can predict subsurface anomalies and minimize unscheduled runway closures. This approach combines a plethora of crucial field-based pertinent information including regional geology, aviation infrastructure management, and artificial intelligence to improve airport safety and efficiency and to support scalable deployment at global air transport hubs. Future work will explore federated learning for cross-site model generalization, real-time integration of seismic and meteorological data streams, and autonomous decision-making protocols to ensure critical infrastructure resilience.

Geological Society of America Abstracts with Program. Vol. 57, No. 6, 2025

© Copyright 2025 The Geological Society of America (GSA), all rights reserved.