Concave distribution characterization of asphalt pavement surface segregation using smartphone and image processing based techniques

Abstract

Surface segregation of bituminous mixtures is a criterion of pavement quality and largely affects the characteristics of the pavement. Simple yet effective evaluation of the surface segregation will enable pavement engineers to tailor timely strategies to mitigate the problem. In this paper, a more efficient image processing method with the aid of smartphone imaging was adopted to rate the segregation level of asphalt pavement surface. Twenty-seven asphalt mixture specimens with vast differences were prepared to acquire different surface images using three types of smartphones. A field test section was chosen to validate the practicability. Furthermore, the Fractal Dimensions (FD, DBC-FD) and Percentage of Concave Distribution Area (PCDA) were used to characterize Concave Distribution Characteristics (CDC) of asphalt pavement surface. Texture Depth (TD) and Mean Texture Depth (MTD) were gained through the sand patch method. It was found that it is an encouraging approach to evaluate the surface segregation based on CDC. The image processing technique relying on the selected smartphone type was proposed by the error rate of reliability analysis, which was not more than 3% compared to the other used smartphones. A newly developed indicator called e was presented to stand for PCDA. The coefficient of determination between e, FD, DBC-FD and TD/MTD are respectively 0.7958, 0.7882, and 0.7585. In the field validation, the coefficient of determination between e and TD/MTD reaches to 0.8546. Therefore, it was demonstrated that the proposed image processing method can be a promising approach to rate the segregation of asphalt pavement surface.