Reincarnate is pleased to share a new publication presented at NDT-CE 2025 – The International Symposium on Nondestructive Testing in Civil Engineering: “Beyond Corrosion: Autonomous Real-Time Detection of Structural Cracks and Degradation Using Robotic Sensing and Digital Twins”. The publication is featured in the e-Journal of Nondestructive Testing – ndt.net and it explores how AI-powered multi-sensor systems and robotic platforms can transform infrastructure inspection by enabling real-time detection of cracks and hazardous gas leaks.
Smart Robotics and AI for Safer, Sustainable Infrastructure
The article, co-authored by Ghezal Ahmad Jan ZIA, Cemal HEKIM, Sabine KRUSCHWITZ (all from our Consortium partner Bundesanstalt für Materialforschung und -prüfung) highlights the development of a climbing robot equipped with optical, thermal, and gas sensors, supported by AI models, to carry out autonomous inspections of hard-to-reach structures. This innovation demonstrates how digital twins and real-time data can improve monitoring, reduce risks for human inspectors, and contribute to longer-lasting, more sustainable infrastructure.
Multi-Sensor Innovation
At the heart of the research is a real-time inspection system that integrates three types of sensors:
- Optical RGB cameras for surface crack detection,
- Thermal infrared cameras for subsurface and hidden damage, and
- Gas sensors to monitor environmental hazards.
This multi-sensor, AI-driven approach provides faster, safer, and more reliable inspections compared to traditional single-sensor or manual methods. It represents a significant step toward more sustainable infrastructure management, aligning with Reincarnate’s mission to extend the lifespan of built assets while reducing risks and environmental impact.
Key Benefits of the New Approach include:
- Multi-sensor fusion: Optical (RGB), thermal (IR), and gas sensing combined in one platform.
- AI-driven analysis: MobileNetV2 for crack classification, U-Net for thermal segmentation, and Random Forest models for gas anomaly detection.
- Robotic integration: A custom-built vertical climbing robot equipped with vacuum adhesion for safe access to hard-to-reach areas like bridge columns or industrial tanks.
- Real-time performance: Crack and gas detections processed in under a second, with live visual overlays and quantitative metrics.
Each sensing modality is powered by specialized AI models — MobileNetV2 for surface crack detection, U-Net for thermal segmentation, and Random Forests for gas anomaly analysis — ensuring both precision and efficiency. Mounted on a custom-built vertical climbing robot, the system can safely access areas that are otherwise dangerous or difficult for human inspectors. Most importantly, all detections are processed in real time, with results displayed in under a second as visual overlays and quantitative metrics.
This publication highlights the potential of combining robotics with AI-powered multi-sensor systems to transform the way civil infrastructure is inspected. By enabling real-time detection of cracks, thermal anomalies, and gas leaks, the approach supports safer working conditions and more efficient monitoring. It also offers a valuable framework for extending the lifespan of built assets and guiding the shift toward sustainable, technology-driven maintenance practices.
By linking robotic sensing with AI-driven analysis, the research paves the way for integrating digital twin technologies into infrastructure inspection, offering practical tools that enhance safety, extend asset lifespans, and contribute to more sustainable management of the built environment.