Proposition of an aggregate-binder interface index based on the adhesiveness test assessed by digital image processing

Aggregate-binder adhesiveness impacts asphalt mixture performance. Evaluation of this interface within mixes typically involves visual assessment and digital image processing (DIP), often resulting in an average value for this complex property. Current computational simulations overlook the possibility of incomplete aggregate adhesion to the binder, a common occurrence in cases of adhesive/cohesive failure between materials. This study proposes an interface index determined from adhesiveness test results in order to improve computational simulations. For this purpose, twelve adhesiveness tests, following Brazilian standards, including mineral and steel-slag aggregates, and neat and modified binders, were conducted. Software with a graphical user interface (GUI) was developed to quantify the percentage of aggregate area coated by binder (%Coated) post-testing. Descriptive statistics and probability distributions were applied to %Coated values, in addition to the quality of measurements using DIP. The DIP values exhibited sensitivity across all examined covering scenarios, with statistical tests suggesting a high likelihood of accurately representing the %Coated parameter from the assessed probability distributions. In conclusion, the developed software enables satisfactory quantification of %Coated. Furthermore, this study indicates that the adhesion test results must be represented considering the data variability, with the Beta distribution being a good way to represent the behavior. •The Beta distribution of probability represents the result of the aggregate-binder adhesiveness test.•100 particles are a reasonable number of observations to represent the Brazilian adhesiveness test.•The adhesion test can be represented with a mean value if the samples are close to 100 or 0 % of the coated aggregate area.•Steel slag aggregates, when compared to crushed mineral aggregates, demonstrated more uniform adhesion.•The binder brightness significantly influences the digital image processing with only samples with modified binders.