Abstract
This paper presents an end-to-end framework for automated inspection and condition assessment of airport runway pavement using UAV-acquired imagery. The proposed approach integrates Unmanned Aerial Vehicle (UAV)-based data collection, deep learning-based pixel-level semantic segmentation of surface defects, and Geographic Information System (GIS)-based spatial aggregation to generate a georeferenced digital representation of airfield pavement condition. Multiple safety-critical defect types are detected and localized at pixel resolution, while spatially referenced processing enables a Pavement Condition Index (PCI)-inspired condition assessment based on defect density within predefined sampling units. The framework is validated through a real-world case study at Zadar Airport, where the entire runway was surveyed using high-resolution UAV imagery. The results demonstrate the system's capability to identify and map multiple defect categories across the full runway extent and to produce a coherent, runway-scale condition map supporting maintenance prioritization and decision-making. Overall, the proposed solution provides a scalable, data-driven alternative to traditional manual runway inspection workflows and establishes a practical foundation for digital condition monitoring of airport pavement infrastructure.