Portable Image Analysis System for Characterizing Aggregate Morphology

Abstract

In the last decade, the application of image-based evaluation of particle shape, angularity and texture has been widely researched to characterize aggregate morphology. These efforts have been driven by the knowledge that the morphologic characteristics affect the properties and ultimate performance of aggregate mixtures in hot-mixed asphalt, hydraulic cement concrete and bound and unbound pavement layers, yet the lack of rapid, objective, and quantitative methods for assessment have inhibited their application in the engineering process. Developed systems for computer-based imaging and image analysis can cost up to $30-40,000 and are usually not portable to the field. However, recent advances in technology have produced pocket computers having as much processing power as was available in some desktop computers. This project takes advantage of these advances to develop an inexpensive portable image analysis system for characterizing aggregate morphology. The system was developed with an integral pocket computer-high resolution camera but can also use individual components consisting of a digital camera and lap-or desk-top computer.Digital images of aggregate particles are captured with the camera. These images are analyzed within the Matlab software program environment with a macro developed and written for this project that uses Fast Fourier Transform to characterize the particle morphology with respect to three parameters: shape, angularity and texture, based on the particle perimeter (outline or edge). By analyzing a number of particles from a source, it can be characterized with respect to these three parameters.Following development of the analysis program, 10 coarse aggregates from various Virginia sources were analyzed. Particles of each aggregate were randomly chosen so that each group contained the various shapes and textures representative of the source. Three images of each particle were obtained at distances of 2, 3, and 10 in to evaluate the resolution needed for adequate analysis. The reliability of the image processing was assessed by statistically analyzing the shape, angularity, and texture values to determine how the threshold parameter affects the particle edge acquisition. Asymptotic analysis was performed to determine the number of images needed to obtain a statistically stable value for each aggregate parameter.It was determined that images acquired at close range (2 or 3 in) were needed to provide sufficient resolution to adequately characterize the aggregate. It was also found that statistically valid values for aggregate shape, angularity, and texture can be obtained from fifteen particle images of random orientation. It can be concluded that the system can be successfully used to characterize coarse aggregate morphology.It is recommended that the Virginia Department of Transportation's Materials Division begin collecting images of aggregates used statewide and collaborate with the VTRC to perform the characterizations and build the database of aggregate morphologic characteristi...

Evaluation of image analysis methods used for quantification of particle angularity

“…Currently there are several image analysis methods available to quantify particle angularity, for example: the gradient angularity (GRAD) method, the angularity using outline slope (AI) method, the angularity factor using a two‐dimensional Fourier transform index (AF) method, the smoothing angularity index (SAI) method, the surface erosion‐dilation approach, the radius angularity method, the one‐dimensional Fourier transform index and the fractal dimension method (Masad et al ., ; Rao & Tutumluer, ; Masad et al ., ; Rao et al ., ; Maerz, ; Wang et al ., ; Al‐Rousan et al ., ; Masad et al ., ; Tafesse et al ., ; Wang et al ., ). This article evaluates four of these methods: GRAD and AI, as well as two methods that were not previously tested; namely the AF and SAI.…”

Stress State of Asphalt Mortar Based on Meso-Scopic Finite Element Simulation and Verification

“…In traditional research, the stress state of asphalt pavement is generally analyzed from the macro perspective, but it cannot reflect the internal stress state of asphalt mixture. This increases the risk of asphalt pavement damage [5][6][7][8][9]. On the other hand, the random internal structure of asphalt mixture will affect its macro mechanical properties.…”

Monitoring the performance of DMS circuits using RhoVol technology

“…Prior to generating any accurate density profiles by the RhoVol, an appropriate function for the volume correction factor for the 3D reconstructed volumes of particles needs to be established. Previous studies have proven that multi-view silhouette-based volume reconstruction methods will always overestimate the particle volume (Wang et al, 2009;Okonta and Magagula, 2011;Mangera and Morrison, 2016). The volume correction factor is primarily shape-dependent, but it is also material-dependent, since ultimately the nature of the material defines its final shape after comminution.…”