Automation of pattern recognition and fractal-geometry-based pattern quantification, exemplified by mineral-phase distribution patterns in igneous rocks

Mark Peternell, Jörn H. Kruhl

Research output: Contribution to journalArticlepeer-review

20 Scopus citations

Abstract

This study analyses the possibility of accurate quantification of automatically digitized mineral-phase distribution patterns in igneous rocks. Based on their colour contrast, different minerals were manually and automatically digitized on micro to macro scales. Depending on the digitized mineral phase, the accuracy of the automated digitizing procedure varies. Quantification of mineral distribution patterns was performed by box-counting. The results do not depend greatly on the pattern's pixel density and are similar to each other, even if automatic recording is performed at reduced precision in comparison to manual recording. Consequently, box-counting measurement of automatically recorded mineral distribution patterns (i) leads to fast and accurate pattern quantification, (ii) allows analysis of various phase distribution patterns from micro to macro scales, and (iii) forms an excellent basis for receiving information on pattern-forming processes, which is not available otherwise.

Original languageEnglish
Pages (from-to)1415-1426
Number of pages12
JournalComputers and Geosciences
Volume35
Issue number7
DOIs
StatePublished - Jul 2009

Keywords

  • Box-counting
  • Fabric inhomogeneity
  • Fabric quantification
  • Fractal geometry
  • Granite

Fingerprint

Dive into the research topics of 'Automation of pattern recognition and fractal-geometry-based pattern quantification, exemplified by mineral-phase distribution patterns in igneous rocks'. Together they form a unique fingerprint.

Cite this