ON THE IMPORTANCE OF ADVANCED IMAGE PROCESSING METHODS IN MATERIAL RESEARCH

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60461373%3A22310%2F16%3A43902021" target="_blank" >RIV/60461373:22310/16:43902021 - isvavai.cz</a>

Výsledek na webu

<a href="http://www.icct.cz/getattachment/PREDCHOZI-KONFERENCE/2016/Proceedings_ICCT2016.pdf.aspx" target="_blank" >http://www.icct.cz/getattachment/PREDCHOZI-KONFERENCE/2016/Proceedings_ICCT2016.pdf.aspx</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

ON THE IMPORTANCE OF ADVANCED IMAGE PROCESSING METHODS IN MATERIAL RESEARCH

Popis výsledku v původním jazyce

Scanning electron microscopy and focused ion beam nanotomography belong to the common tools for investigation of microstructural features of numerous materials, e.g., porous media, ceramics, composites and alloys. Secondary electron (SE) and back-scattered electron (BSE) images obtained using these techniques often suffer from noise, non-uniform illumination or lack of contrast. Therefore, there is a continuous need for advanced image processing methods that are capable of remedying these shortcomings. For the purpose of microstructural analysis, the image processing usually includes histogram equalization, filtering in a spatial or frequency domain, and segmentation. The later operation is crucial for unique determination individual phases forming the material. However, without a careful preprocessing step, particularly involving filtering, results of segmentation can be questionable, i.e., locations of phase interfaces can be ambiguous. In an extreme case, segmentation can completely fail. In this contribution, we deal with the application of PDE-based image analysis of SE and BSE images of porous media and composite materials and its combination with other preprocessing steps, particularly histogram equalization. Using public domain software and our parallel computer code, we evaluate the performance of various edge-preserving diffusion filters and other classic blurring filters for porous media in two- and three-dimensional spatial domains. To do so, we segment the images using Otsu's method [4] of global thresholding and compare resulting macroscopic measures, e.g., volume fractions or a specific surface area, with their counterparts obtained independently of image analysis. In all cases, the edge-preserving diffusion filters beat the classic ones.

Název v anglickém jazyce

ON THE IMPORTANCE OF ADVANCED IMAGE PROCESSING METHODS IN MATERIAL RESEARCH

Popis výsledku anglicky

Scanning electron microscopy and focused ion beam nanotomography belong to the common tools for investigation of microstructural features of numerous materials, e.g., porous media, ceramics, composites and alloys. Secondary electron (SE) and back-scattered electron (BSE) images obtained using these techniques often suffer from noise, non-uniform illumination or lack of contrast. Therefore, there is a continuous need for advanced image processing methods that are capable of remedying these shortcomings. For the purpose of microstructural analysis, the image processing usually includes histogram equalization, filtering in a spatial or frequency domain, and segmentation. The later operation is crucial for unique determination individual phases forming the material. However, without a careful preprocessing step, particularly involving filtering, results of segmentation can be questionable, i.e., locations of phase interfaces can be ambiguous. In an extreme case, segmentation can completely fail. In this contribution, we deal with the application of PDE-based image analysis of SE and BSE images of porous media and composite materials and its combination with other preprocessing steps, particularly histogram equalization. Using public domain software and our parallel computer code, we evaluate the performance of various edge-preserving diffusion filters and other classic blurring filters for porous media in two- and three-dimensional spatial domains. To do so, we segment the images using Otsu's method [4] of global thresholding and compare resulting macroscopic measures, e.g., volume fractions or a specific surface area, with their counterparts obtained independently of image analysis. In all cases, the edge-preserving diffusion filters beat the classic ones.

Druh

D - Stať ve sborníku

CEP obor

JI - Kompositní materiály

OECD FORD obor

—

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2016

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název statě ve sborníku

PROCEEDINGS of the 4th International Conference on Chemical Technology

ISBN

978-80-86238-94-4

ISSN

2336-811X

e-ISSN

—

Počet stran výsledku

6

Strana od-do

268-273

Název nakladatele

Česká společnost průmyslové chemie (ČSPCH)

Místo vydání

Praha

Místo konání akce

Mikulov, Czech Republic

Datum konání akce

25. 4. 2016

Typ akce podle státní příslušnosti

CST - Celostátní akce

Kód UT WoS článku

—