Using an AI-Powered Decision Support Algorithm for Underground Monitoring with UAV-Integrated GPR

Accurate, rapid, and reliable identification of subsurface structures and anomalies is crucial for infrastructure management, engineering projects, disaster response, archaeological research, and geotechnical investigations. However, traditional ground-penetrating radar (GPR) analysis primarily relies on expert experience, making the process time-consuming and subject to user variability.

This study presents an AI-driven decision support system developed through a university–industry partnership. It automatically analyzes data from a GPR system mounted on an unmanned aerial vehicle (UAV). The system transmits GPR data collected via the UAV to dedicated software, which conducts a multi-stage analysis using AI-supported algorithms developed as part of ongoing research.

The algorithm offers an integrated decision-making tool that not only processes radar images but also performs data quality control, detects subsurface anomalies, interprets potential structures, evaluates technical feasibility for the scenario, and automatically generates reports—all without user input. A key feature is its modular design, allowing easy adaptation across various sectors and applications. Consequently, it can produce scenario-specific reports for infrastructure detection, geotechnical surveys, archaeology, post-disaster assessments, engineering, and more.

This study details the system architecture, explains the core operation of the AI-assisted decision mechanism, and shares example analysis results from diverse fields. By showcasing automatically generated reports across different scenarios, the software demonstrates that it is versatile, scalable, and suitable for multiple disciplines. This approach aims to reduce reliance on experts, accelerate analysis, and ensure consistent decision support outputs across applications.