Aggregate Shape Characterization Using Digital Image Processing

Aggregates must meet certain specifications to be acceptable in asphalt and concrete applications. Among these are specifications that deal with the various aspects of aggregate shape, including flatness and elongation and aggregate angularity. The current practice for ensuring those specifications is the use of manual-mechanical tests. These tests are time consuming, labor intensive and subjective. In addition, tests such as the compacted or uncompacted voids tests (Figure 1), are taken to be a shape (angularity) indicator, even thought the measured quantify is the void ratio. This report describes a prototype of an automated digital video image analysis system that measures both the flat and elongation (F&E) ratio, and the angularity of aggregate (Figure 2). This report also compares the results of manual testing with the measurements of the imaging system. The concept is that by using a digital imaging system, quick, inexpensive and objective measurements can be made. Because the measurements are so quick, faster adjustments to manufacturing processes can be made, to decrease the cost of producing off-specification materials. Because the incremental cost of more measurements is negligible, more tests can be performed, improving statistically reliability. Because the measurements are more objective, test results will be less affected by inexperienced or inattentive operators. The innovation is in using state of the art video imaging hardware and software to make a real time measuring system to measure flat and elongation and angularity. New improvements include the use of backlighting to improve the imaging of the aggregate pieces and the measurement of the curve radius of the corners of the aggregate as a measure of angularity. In addition it was demonstrated the there is a potential to use this technology for sand-sized aggregate pieces. Research results show that image measured f&E ratios are fairly close to matching caliper results, although some differences were found. Repeatability of the imaged measurements was found to be better than with manual tests. Research results shows that the image measured angularity measurements can correlate well to voids tests. The repeatability of the imaged measurement is not quite as good as that of the voids test. Analysis of the flat and elongation measurements as a function of crusher type showed that impact type crushers tend to produce more cubical particles, even when rock type is not accounted for