Starr White

A rapid and quantitative image analysis method has been developed for analyzing reflected cross-polarized light micrographs of coke samples, enabling the determination of the coke’s morphology. The method uses k-means clustering to segment the image into four discrete domain types and quantifies their sizes and anisotropy. From this analysis, an optical texture index (OTI) is computed following a procedure similar to the manual approaches used in the past to rank needle coke samples.