Characterising Pavement Surface Macrotexture using 3D Modelling Techniques
Currently Hot Rolled Asphalt (HRA) surfacings make up over 50% of the pavement surface on the Irish national road network. HRA is a dense, gap graded bituminous mixture with a layer of cold coated chippings applied to the surface mat and then rolled. Well manufactured HRA is a positively textured surface i.e. the surface macrotexture is formed by the peaks and ridges of the coated chippings projecting above the surface level of the mastic. Detailed visual inspections of relatively new HRA schemes constructed in Ireland indicated that the chippings were in many cases clustered together forming voids below surface level. This type of surface is characterised as having ‘non-positive’ texture, which should not be a feature of HRA.
Currently, HRA surfacings make up over 50% of the pavement surface on the national road network. According to a European study, well designed HRA may offer between 17 and 25 years of service life on heavily trafficked European roads.
Macrotexture acceptance criteria for HRA at the time was based on texture depth only.
Surface texture takes two forms:
‘positive’ texture: a cluster of angular peaks or series of ridges above a datum level, typical of chip sealing, hot rolled asphalt with chips, slurry and microsurfacings and brushed concrete. Figure 1 illustrates positive texture in a HRA surfacing material.
‘negative’ texture: a network of voids or depressions below the general level, typical of thin surface course systems, and porous asphalt. This ‘negative’ texture should not be a feature of a HRA surfacing.
Detailed visual inspections of relatively new HRA schemes constructed in Ireland indicated that the coated chippings were in many cases clustered together forming voids below surface level. This type of surface is characterised as having ‘non-positive’ texture to distinguish from the designed ‘negative’ texture provided by thin surface course materials.
When the HRA surface was tested using the specified volumetric patch method, it was possible to achieve the specified texture depth even though the pavement surface material had non-positive texture. Utilising the sand patch test would offer no clear indicator and thus a more objective test was needed.
The CRP technique developed at Ulster University is called the Ulster University Photogrammetric Method of Highway Surface Recovery and 3D Modelling or UUTex3D (6). The technique uses a bundle of 12 images to generate the 3D model, taken from 12 different orientations. Eight images are taken at a nominal 60-degree angle to the pavement surface, with the remaining four images taken normal to the surface.
While technically three images are sufficient to create a 3D model, acquiring 12 images offers redundancy enabling poor quality images to be rejected from the study while still enabling a 3D model to be extracted.
Six sites were selected for the evaluation of macrotexture type. An initial walking survey of each site was carried out by experienced pavement engineers to characterise the sites as being either positive or non-positive. Four sites were of relatively new HRA and characterised as displaying non-positive texture when visually assessed. The two remaining sites were characterised as having good positive texture based on visual assessment.
Following the initial visual assessment, a detailed CRP data collection exercise was carried out on each site. The sites were typically 500 metres in length. The exception was the N04 where, due to traffic management constraints, data collection was limited to a 225 metre length. On each site, data was collected at discrete points at 50 metre intervals, or at 25 metre intervals on the N04. Two point locations were selected at each interval – one in the left-hand wheel path and one in the centre of the lane – totalling 20 discrete point locations per site. For each point location, a set of 12 images was collected in accordance with the UUTex3D methodology. A Canon EOS 6D using a focal length of 70 mm was used to collect the images. Additionally, a volumetric patch test was carried out at each point location. Each volumetric patch test was performed at the identical point location used in the CRP data collection.
The Islands Module: Effectively a virtual horizontal slice is taken through the surface at a user defined threshold. The islands are the portions of the model which project above this slice. This is analogous to discrete stone chips protruding above the surface of the mastic in a positive-textured HRA. Accordingly, the Islands module was the primary analytical tool used in this study. The number of islands identified at each location, together with the horizontal surface area of each island, was calculated and reported.
This study has provided an objective analytical method of distinguishing between positive and non-positive textured HRA surface courses using 3D modelling. Islands analysis has been shown to be effective in this regard. A clear delineation is apparent between positive and non-positive sites in the Islands analysis.
The Count50 parameter has been proposed for adoption in the latest version of the TII Publications CC-SPW-00900, Specification for Road Works Series 900 and CC-GSW-00900, Notes for Guidance on the Specification for Road Works Series NG 900.