Assessment of aggregate structure in porous asphalt using X-ray computed tomography

Abstract: X-ray computed tomography is a technique that has been successfully utilized to characterize internal microstructure of asphalt mixtures. The city of Linköping (Sweden) developed an action plan to reduce noise; accordingly KTH and VTI developed a method to determine air void content in porous asphalt using x-ray computed tomography and an image processing software called ImageJ. The present study is a continuation of the previously mentioned work and focuses on the investigation of aggregate structure in porous asphalt by x-ray computed tomography. First the previously proposed method to estimate air void content was validated. Then, the assessment of aggregate structure including qualitative and quantitative analysis was completed. Qualitative evaluation was performed to determine the quality of slices in regard to establishing challenging areas, and the extent of beam hardening present in the X-ray images. This evaluation produced quality slices for each sample in Y-direction to work as a reference to establish general threshold ranges and image enhancement procedures, as well as identifying the interface between the top and bottom layer of the porous asphalt. The quantitative analysis consisted of analysing aggregate structures in the porous asphalt and developing a method to estimate the aggregate size distribution in porous asphalt layers. The gradation curves from the quantification of aggregate size distribution in all directions (X, Y, Z) were compared to the gradation curves from laboratory sieving tests previously performed on the drilled asphalt cores. To determine the accuracy of the method a perimeter analysis was performed to evaluate the suggested method to measure the aggregates. The results obtained indicate that to quantify the aggregates in porous asphalt, enhancement of the images is needed, as well as morphological operations to deal with beam hardening and overlapping stones due to unsuccessful separation of aggregates when thresholding. This further indicates that ImageJ Fiji is more suitable for complex cases, such as cases where aggregates and mastic are hard to distinguish from each other, rather than the original ImageJ software. This is due to the extended plugins available in ImageJ Fiji, where more options of for example filtering and enhancing of images are available. The results also suggest that the proposed method is a suitable method to determine the aggregate size distribution in porous asphalt pavements, as it allows to quantify the aggregate distribution and produces realistic results with slight inaccuracies due to the analysis being performed in 2D. Future development will be focused on automizing the determination of air void structure and aggregate size distribution, but also on the development of procedures to determine other relevant parameters such as mastic and binder content to establish a complete methodology to investigate the internal structure of a porous asphalt pavement, as well as 3D analysis to determine these parameters.