UAV-ASSISTED RUNWAY INSPECTION AND CRACK DETECTION

This paper presented the comprehensive design, algorithmic development, and simulation-based validation of a UAV-Assisted Runway Inspection and Crack Detection System (UAV-RICDS). The system integrates a multi-sensor UAV platform (RGB, thermal, GPR), a YOLOv8m deep learning crack detection engine trained on over 15,000 annotated pavement images, and an automated ASTM D5340-compliant Pavement Condition Index assessment module, delivering a fully automated, end-to-end runway pavement inspection capability. Validation simulation across 45 experimental scenarios at VOMM and VIDP demonstrated: an overall crack detection accuracy of 94.7% ([email protected]); a 78% reduction in full runway inspection time (41.4 minutes vs. 186 minutes manual); a false positive rate of only 3.2%; a PCI assessment accuracy of ±4.1 PCI units (versus ±11.8 units for manual assessment); and a 78.6% reduction in required runway closure duration. These results provide compelling evidence for the operational superiority and safety benefits of the UAV-RICDS over conventional manual inspection methodologies. The system is designed in compliance with DGCA UAV operational regulations (CAR Section 3, Series X, Part I) and ICAO Annex 14 aerodrome standards, ensuring regulatory acceptability for deployment at Indian civil airports. Future development directions include: integration of hyperspectral imaging payloads for chemical contamination detection; deployment of edge AI inference on embedded UAV processors (NVIDIA Jetson platform) for real-time onboard detection; federated learning across multiple airport deployments for continuous model improvement; integration with Airport Management Information Systems (AMIS) for automated maintenance workflow triggering; and extension of the system to taxiway and apron pavement assessment applications.

[1] International Civil Aviation Organisation (ICAO). (2022). Annex 14 to the Convention on International Civil Aviation – Aerodromes, Volume I: Aerodrome Design and Operations, 9th Edition. Montreal: ICAO. [2] Federal Aviation Administration (FAA). (2021). Advisory Circular AC 150/5370-11B: Use of Nondestructive Testing in the Evaluation of Airport Pavements. Washington DC: US Department of Transportation. [3] ASTM International. (2018). ASTM D5340-18: Standard Test Method for Airport Pavement Condition Index Surveys. West Conshohocken: ASTM International. [4] Jocher, G., Chaurasia, A., & Qiu, J. (2023). Ultralytics YOLOv8. GitHub repository: https://github.com/ultralytics/ultralytics. [5] Yang, L., Li, B., Li, W., Liu, Z., Yang, G., & Xiao, J. (2018). Deep Concrete Inspection Using Unmanned Aerial Vehicle Towards CSSC Database. Proceedings of IROS 2018. Madrid: IEEE. [6] Maas, H. G., & Daerden, M. (2022). UAV-Based Runway Surface Inspection at Frankfurt Airport Using Deep Convolutional Neural Networks. Journal of Aerospace Engineering, 35(4), 04022032. ASCE. [7] Eisenbeiss, H., & Sauerbier, M. (2011). Investigation of UAV Systems and Flight Modes as Photogrammetric Sensor Carriers. The Photogrammetric Record, 26(136), 400–421. Wiley. [8] Otero, L. D., Gagliardo, N., Dalli, D., Huang, W. H., & Cosentino, P. (2015). Proof of Concept for Using Unmanned Aerial Vehicles for High Mast Pole and Bridge Inspection. Florida: Florida Department of Transportation. [9] Zhang, L., Yang, F., Daniel Zhang, Y., & Zhu, Y. J. (2016). Road Crack Detection Using Deep Convolutional Neural Network. Proceedings of ICIP 2016. Phoenix: IEEE. [10] Airports Authority of India (AAI). (2023). AIP India – Aeronautical Information Publication, ENR Section. New Delhi: AAI.