Determination of the particle size distribution and shape charectiristics of crushed stone aggregates by using image analysis

The main problems geotechnical engineering has to solve are can be summarized as keeping permeability and deformations within certain limits. At this point, the characteristics of the natural ground and the filling material determine the soil behavior. Unlike other engineering branches, geotechnical engineering should primarily determine the ground properties. In this way, the type and characteristics of the ground with heterogeneous structure can be determined. It is important to know the classification of soil. Commonly used soil classification systems are unified soil classification system and the AASHTO classification system. Which use grain size distribution and plastic properties as the basis for determining soil type. The grain size distribution indicates that the soil is coarse or fine. Again, it can be determined from the grain size distribution if it is sand or gravel. However, it is disadvantageous that the essential properties of the granular soil are expressed only by the grain size distribution. The shape characteristics of the aggregates as well as the grain size distribution have an effect on strength. For example, under external loads, angular and elongated aggregates are more likely to crash than rounded aggregates. The grain size and shape characteristics of the granular soils are so important that in the past the shape characteristics have to be found in very laborious manual measurements. However, in parallel with the developments in technology, size distribution and shape characteristics of large number of aggregates with complicated shapes can be found by image processing techniques. In this thesis, grain size distribution and shape characteristics of crushed aggregates were obtained by image processing techniques. Within the scope of this thesis, crushed stone samples of 4.00-12.00 and 12.00-22.00 mm diameter from the quarries located in the Zonguldak province borders were divided into subgroups using standard sieves. Three different grading curves were formed by using these groups of crushed stone aggregates and photos of these samples were taken in the laboratory. In this thesis study, a lighting system has been developed to reduce shadow formation when photographs of samples are taken. These images were digitally analyzed in open source software called Image J. Aggregate dimensions were obtained from the image analyzes and shape characteristics such as roundness and sphericity of the crushed aggregates were obtained in the Matlab program where image analysis and geometric calculations are used together. As a result, the grading curves of mechanical sieving and image processing were compared.