Impact of granular packing heterogeneity on macroscopic properties via automated 3D feature extraction algorithm: A case study on surface texture variability in asphalt mixtures

Three-dimensional granular packing systems exhibit significant structural heterogeneity, which limits macroscopic property optimization of engineering materials like asphalt mixtures. Therefore, this study aimed to reveal how heterogeneity drives performance variation, using surface texture as the research focus. To this end, one database containing 100 digital asphalt mixtures and 47 structural parameters were constructed utilizing the digital image processing technology. Subsequently, the collinearity analysis was conducted to filter structural parameters, whose heterogeneity was quantified and validated by conducting parametric uncertainty analysis. Finally, the impact of particle skeleton/void structure heterogeneity on texture was quantified by significance analysis. The results show that among all parameters, multifractal parameters and void characteristics have a significantly greater impact. At the same time, the coefficient of determination close to 1.0 ensures the reliability of the influence model, while also enabling the feasibility of controlling macroscopic property variation through the regulation of structural features.