Automatic Fiber Length Measurements with a Multi-Stencil Fast Marching Method on Microscopy Images

• Published in: Microscopy and microanalysis Vol. 26; no. 3; pp. 387 - 396
• Main Authors: Liu, Chanjuan, Bergmeijer, Menno, Pierrat, Sébastien, Guise, Olivier
• Format: Journal Article
• Language: English
• Published: New York, USA Cambridge University Press 01.06.2020; Oxford University Press
• Subjects: Algorithms; Automation; Bias; Composite materials; Fibers; Human error; Image acquisition; Image analysis; Image processing; Measurement techniques; Mechanical properties; Microscopy; Morphology; Polymers; Sample preparation; Software; Software and Instrumentation; Germany; automated fiber length measurement; medial axis transformation (MAT); multi-stencil fast marching (MSFM); composites materials; digital microscopy images
• ISSN: 1431-9276; 1435-8115
• DOI: 10.1017/S1431927620001336

Fiber length has a strong impact on the mechanical properties of composite materials. It is one of the most important quantitative features in characterizing microstructures for understanding the material performance. Studies conducted to determine fiber length distribution have primarily focused on sample preparation and fiber dispersion. However, the subsequent image analysis is frequently performed manually or semi-automatically, which either requires careful sample preparation or manual intervention in the image analysis and processing. In this article, an image processing and analysis method has been developed based on medial axis transformation via the multi-stencil fast marching method for fiber length measurements on acquired microscopy images. The developed method can be implemented fully automatically and without any user induced delays. This method offers high efficiency, sub-pixel accuracy, and excellent statistical representativity.