Quantitative Image Analysis of Masonry Mortar Microstructure

Deep learning-based automated image segmentation for concrete petrographic analysis

“…Under a higher magnification as shown in Fig. 1c, some fine aggregates are colored to pink, which is related to the semi-transparent nature of sand grains (primarily quartz) [30,31]. Based on our observation, the color dye may infiltrate into the substrate through the interfacial transitional zone (ITZ) around sands, leaving the pinkish visual effect.…”

“…These limitations have long been recognized in concrete studies, and some efforts are seen to improve the segmentation accuracy using additional graphic features. In a study by Werner and Lange, for example, a convolution kernel algorithm extracting the texture contrast was used to differentiate the aggregate and cement paste in SEM images [31].…”

Compression triaxial behavior of concrete: the role of the mesostructure by analysis of X-ray tomographic images

“…This image processing technique has been developed by Eric Landis and his team (Landis et al, 2007) to study the behaviour of mortar, and was adapted for the purposes of this study. The segmentation process is not a new technique and has already attracted the interest of many authors (Werner & Lange, 1999;Yang & Buenfeld, 2001). One feature characterising such a processing approach is the lack of a single true result, but instead a number of results that depend on parameter choices.…”

Effect of Internal Curing as Mitigation to Minimize Alkali-Silica Reaction Damage

“…The quantification of the porosity and unhydrated cement particles was performed in the BSE maps containing only the area of the paste through wellestablished image analysis. [20][21][22][23][24][25][26] As a summary, the process required a binarization followed by a quantification of the area occupied by pores. The threshold level was chosen based on the tangent method proposed by Scrivener et al 21 Finally, the pixel size in the images (one pixel representing 0.2 x 0.2 µm [7.874 × 10 -6 x 7.874 × 10 -6 in.]…”