Multi-Scale Image Analysis of Aggregate Packing and Its Influence on Asphalt Mixture Fatigue Life

Abstract

Multi-scale image analysis has emerged as a critical bridge between the microscopic morphology of aggregates and the macroscopic fatigue performance of asphalt mixtures. This paper develops a system-level examination of how imaging architectures, computational pipelines, and governance arrangements shape the translation of aggregate packing descriptors into pavement fatigue predictions. The argument is that fatigue life is not simply a function of binder content or aggregate gradation; it is mediated by spatial arrangement, contact topology, void distribution, and the robustness of the analytical infrastructure used to observe these features. The paper reviews micro-scale, meso-scale, and macro-scale imaging modalities, including X-ray computed tomography, scanning electron microscopy, optical surface scanning, and field-based vision systems, and considers their integration into digital twins for pavement asset management. It then analyzes structural trade-offs among durability, permeability, workability, and sustainability, showing that aggregate packing optimization can improve fatigue resistance while creating new vulnerabilities in compaction, moisture susceptibility, and production variability. The discussion extends to data governance, fairness, policy, and deployment, emphasizing that multi-scale imaging systems are socio-technical infrastructures rather than neutral measurement tools. Issues of data ownership, algorithmic accountability, standardization, and geographic equity are examined alongside technical questions of resolution, calibration, and model transferability. The paper concludes that robust multi-scale imaging can support longer-lasting, lower-carbon pavements, but only if technical accuracy is matched by institutional capacity, transparent validation, and inclusive decision-making.