Fly ash is one of the residues generated in combustion, and comprises the fine particles that rise with flue gases. In the US about 43% is recycled and is often used to supplement Portland cement in concrete production. Fly ash can improve the concretes mechanical properties and decrease cost. Depending upon the source and makeup of the coal being burned, the components of fly ash vary considerably. These variations affect the quality of the final product. Accordingly, it is important for cement manufacturers to know the amount and type of the components in these particles. The objective of this project is segmentation of images of fly ash particles acquired using a micro computed tomography (μCT) imaging device. A set of grayscale images is produced, with each image representing a particular slice of the particle. The desired segmentation operation should identify particles and label regions of a given image based on similarity, as perceived by human observers. In this paper, two techniques are proposed for segmenting different phases of material in these images. The first technique uses Contrast Stretching and Histogram Matching and is based solely on the gray scale value of the pixels in the image slices. In the second proposed technique, Circular Gabor Filters (CGF) are used to segment the regions with porous textures in the cross section of the particle. We have also proposed a technique for designing the CGF such that when applied to the gray scale images, the filter passes the porous regions of components accurately, while blocking non-porous regions. By combining these techniques, we have developed a program that is able to segment different types and regions of impurities in the cross sections of a fly ash particle and create 3D models of these particles, presenting the locations and sizes of different phases of material.

This content is only available via PDF.
You do not currently have access to this content.