Gradation Analysis Of Asphalt Mixture Based On Digital Image Analysis And Stereology

Aggregates, bonder and voids consist of asphalt mixture. In the image, the colors of aggregates, bonder and voids are repectively close to white, plumbeous and black. The image of asphalt mixture can show its microstructure greatly because of good color contrast. In order to obtain component of aggregate, digital image analysis technique was applied to process secion image of asphalt mixture.Voids and bonder are treated as "background", because that the main research of this paper is gradation of aggregate as well as it is difficult to distinguish void and bonder for their close color and the characteristic of void, such as small proportion, random distributing, tiny volume and etc. In this paper, we compare and analysis the existing theory and technique of digital image analysis, and then choose some effective method from them. A program of digital image analys of asphalt mixture was developed using Visual Basic exclusively, with which some two-dimensional parameters of aggregate particle, such as area, perimeter, equivalent diameter and etc can be obtained.Research on three-dimensional size distribution of aggregates through 2-D image is very important. It is the key to practice from theory. There are three dominating methods. One is direct measurement of 3-D image, which is costly. Another is 3-D space reconstruction by sieries of 2-D images, which is multifarious and time-consuming. The last one is stereology, which is the right method used in this paper. According to geometry facilitation, such as shape of particles being spherical approximately, estimation of three-dimensional size distribution from 2-D by stereology technique can be achieved, and meanwhile shape correction coefficient is applied to revise the former result. The result obtained at this stage is volume gradation. On account of the same source and material of aggregates, density of aggregates is treated as equality and shape of that is thought as analogical, from which we can know that mass gradation is equal to volume gradation. We make 15 Marshall Specimen with the same gradation, and obtain 100 section image of asphalt mixture by cutting specimen and CCD camera. We can have the mass gradarion by proceeding with technique of digital image analysis and stereology. And then this mass grading curves is compared to that obtained by mechanical sieving, from which we can know the two methods are consistent by and large when the shape correction coefficient is 0.87. This conclusion is consistent to that obtained by other researches (coefficient is 0.81～0.89), such as Mora. The study demonstrates this new method is a convenient and accurate for gradation analysis of asphalt mixture.