Image Processing of Aggregate Skeleton Structure of Asphalt Mixture for Aggregate Uniformity Quantification

Nonuniform or segregated asphalt mixture has been critical in asphalt pavement construction, accelerating pavement damage and reducing pavement service life. Existing uniformity quantification methods have not considered the aggregate skeleton structure that impacts the mixture loading and durability. Under the isotropic hypothesis, this paper proposed a nondestructive, real-time, and rapid approach to measuring asphalt mixture uniformity using image processing technology (IPT). A total of 1,188 images were first collected from a new construction site and then preprocessed to identify the coarse aggregate. The aggregate particle and skeleton structure factors were obtained, including particle quantity and position distribution, particle axial length ratio, particle orientation, and distance between particles. We used the entropy weight method to combine these four factors to propose a uniformity quantification index (UQI). Finally, UQI was compared with two existing uniformity quantification indices in evaluation consistency. The results show that the uniformity level ranked by the UQI was more consistent with the visual observation than the other two indices. Among the four factors, the particle quantity and position distribution, particle axial length ratio, and distance between particles correlate well with the asphalt mixture uniformity. Moreover, for the considered asphalt mixtures, stable quantification results need a minimum image size of 250 by 556 mm. Even smaller image sizes produced misleading uniformity quantification results.