Characterization of Homogeneity of Asphalt Concrete Using 2D Cross-Sectional Images

Abstract

Asphalt concrete is fabricated from a three-phase mixture consisting of aggregate, asphalt binder, and air voids. Homogeneity of asphalt mixtures affects performance characteristics of asphalt concrete pavements. In this study, homogeneity is determined from 2D vertical section images of gyratory compacted specimens. Because the aggregate structure dominates the performance of asphalt concrete, the mixture homogeneity is determined via statistical distribution of aggregate particles. The cross-sectional images are obtained using a flatbed scanner and then processed to detect various particle shape parameters and locations. Aggregate shape parameters are used for the purpose of determining aggregate gradation curves based on 2D cross-section images of asphalt concrete specimens. An algorithm is developed to detect all the particles in the cross-sections, and then the detected aggregates are redistributed at random orientations and locations to create synthetic images of the same cross-sections. In this way, a large number of sectional images representing the distribution of the same aggregate size fractions are generated. Statistical analyses are performed on four different mixture specimens to determine the range of particle size distributions functions. Results show that the particle distribution of the specimens fall within the range of distribution functions for homogenous mixtures.