Texture feature analysis of coke slice image

The micro optical texture structure is one of the most relevant characteristics of coke. In this paper, the texture features, especially the fractal feature, of coke slice image is analysed, and a box-counting algorithm is implemented to calculate the fractal dimension of coke slice image. At the same time, some co-occurrence matrix-based statistical texture features are also analysed. From the experimental results, the relationships between fractal dimension statistical texture parameters and the porosity of coke are developed, which provide an important basis for the auto analysis of coke quality.