Application of Textural Features in the Analysis of Breakstone Grading
A. I. Makarov1, V. A. Ermakov1, D. A. Ekimov1,2
1Institute of Physics and Technology, Petrozavodsk, 185910 Russia
2Department of Multidisciplinary Scientific Research, Karelian Research Center, Russian Academy of Sciences, Petrozavodsk, 185910 Russia
Keywords: Гранулометрический состав, автокорреляционная функция, текстурный подход, Grain size composition, autocorrelation function, texture approach
Abstract
Accuracy of breakstone grain-size analysis using digital images in the initial method and its modification based on algorithm proposed by Rubin is compared. A modification with averaging of features over all directions and the method with a classification feature represented by difference of intensity distribution functions of fragment projections are described. The results obtained using these methods in a series of tests on grading of five breakstone fractions measured in a certified laboratory. It is shown that the modified method by Rubin with averaging over all directions provides the highest accuracy.
|
