Skip to main content
Doctoral thesis, 2009

Image analysis for volumetric characterisation of microstructure

Axelsson, Maria


Digital image analysis provides methods for automatic, fast, and reproducible analysis of images. The main contribution of this thesis is new image analysis methods for volumetric characterisation of microstructure with application in the field of material science. The methods can be used as tools to characterise material microstructure, in particular the structure of fibre-based materials, such as paper, wood fibre composites, and press felts. More information about the material microstructure enables design of new materials with more specialised properties. Volume images have recently become available to characterise material microstructure. Manual inspection of material properties using volume images is both non-reproducible and expensive. The methods presented in this thesis are developed to meet the growing need for automated analysis. The focus has been on 3D methods for high-resolution volume images, such as X-ray microtomography images. New methods for characterisation of both the fibre structure and pore structure in fibre-based materials are presented. The fibre structure can be characterised by measuring either individual fibres or the local structure of the material. A method for tracking individual fibres in volume images is presented. The method is designed for wood fibres, but can also be applied to other types or fibres or in other areas where tubular or elongated structures are analysed in volume images. A method for estimating 3D fibre orientation of both tubular and solid fibres is also presented. Both methods have been evaluated on real volume images acquired using X-ray microtomography with good results. Two new pore structure representations and corresponding measurements are introduced. The usefulness of the methods is illustrated on real data. A method for estimating the pore volume at the interface between press felt and fibre web is presented. It has been applied in a case study of press felts under load using confocal microscopy images. In addition to the methods for fibre-based materials, a general method for reducing ring artifacts in X-ray microtomography images is presented. The method is evaluated on real data with good results. It is also applied as a preprocessing step before further analysis of the X-ray microtomography images.


image analysis; image processing; fibres; paper; wood; ultrastructure

Published in

Acta Universitatis Agriculturae Sueciae
2009, number: 2009:19
ISBN: 9789186195663
Publisher: Centre for Image Analysis, Swedish University of Agricultural Sciences

Authors' information

Axelsson, Maria
Swedish University of Agricultural Sciences, Centre for Image Analysis

UKÄ Subject classification

Paper, Pulp and Fiber Technology

URI (permanent link to this page)