Some Methods Of Processing And Feature Analysis Of Texture Images

Zbigniew Rudnicki

AGH University of Science and Technology.

DOI:

https://doi.org/10.7494/cmms.2002.1.0014

Abstract:

Pictures of texture have significant meaning in a number of disciplines in science and engineering, such as metallography and crystallography, tribology (pictures of slipping surfaces and particles produced by wear), geology (surfaces of rocks), medicine (fragments of tomographs and X-ray pictures) etc. The present paper characterizes various computer methods used for primary analysis of pictures and, further, analysis and extraction of features of these pictures. The work concentrates on discussion of selected methods and tools, which are used now (determination of statistical parameters, Haralick’s parameters based on matrices of events etc.), as well as on methods proposed by the Author and validated for the case of analysis of slipping surfaces. Some abilities of the software Matlab and Mazda-Convert-B11 (developed for the analysis of tomographs) are also discusses. Among numerous features those were selected, which reflect essential differences in textures and allow their classification and drawing conclusions regarding the features of materials, objects and processes represented by pictures of texture.

Cite as:

Rudnicki, Z. (2002). Some Methods Of Processing And Feature Analysis Of Texture Images. Computer Methods in Materials Science, 2(1), 1 – 18. https://doi.org/10.7494/cmms.2002.1.0014

Article (PDF):

Keywords:

References: