Technical and Computational Aspects of the Measurement of Aggregate Shape by Digital Image Analysis

Abstract

Aggregates need to pass numerous tests to ensure the performance of asphalt and concrete structures and pavements. Some of these tests are fairly onerous, requiring manual, labor intensive, cost ineffective measurements that do not provide significant statistical validity, and are prone to errors through ignorance, negligence, or even in some cases through deliberate misrepresentation. This paper presents a vision based alternative to measure the shape of aggregate particles. The system, although requiring increased capital investment, will result in objective, cost effective, and timely testing of aggregate shape. The system uses dual, synchronized, double speed progressive scan cameras to image the aggregate piece from two directions. A dual image acquisition card simultaneously digitizes both images and does real-time thresholding to create a binary image, which is ported to the host computer. A software trigger determines the presence of an aggregate piece in the image, and the boundaries of the piece are delineated by a perimeter-walking routine. Measurements of aspect ratio and minimum curve radius are made on the perimeter array, and are compared to flat and elongated tests, coarse aggregate angularity (uncompacted voids), compacted voids, and fractured face counts.