Image analysis for characterization and quantification of the steel-concrete interface microstructure

Authors: A. Kenny, A. Katz
Paper category: conference
Book title: Concrete Durability and Service Life Planning – ConcreteLife’09
Editor(s): K. Kovler
Print ISBN: 978-2-35158-074-5
e-ISBN: 978-2-35158-085-1
Publisher: RILEM Publications SARL
Publication year: 2009
Pages: 100 – 107
Total Pages: 8
Language: English

Abstract: The microstructure of steel-concrete interface is an important parameter affecting the rebar susceptibility to corrosion. Even though it was indicated as the main factor in some publications, no standard method or procedure exist for its characterization and quantification.
This work presents a method for characterization and quantification of the microstructure of steel-concrete interface based on SEM-BSE image analysis. The image processing is divided into two parts: classification and characterization.
Classification is the process of assigning physical class (concrete, pore or steel) to every pixel. It was done by using the mean-shift algorithm, which is a statistical unsupervised method.
Characterization is the quantification of the geometrical properties of the microstructure, based on the classified image.
Some of the properties obtained by this method showed high correlation with concrete mix characteristics, physical properties of the interface, and durability aspects. The results improve the confidence in the image analysis method as a technique for steel –concrete interface research.

Online publication: 2012-05-16
Publication Type: full_text
Public price (Euros): 0.00