Fractal analysis of pore roughness in images of soil using the slit island method
journal contribution
posted on 2023-10-19, 11:16 authored by Unknown Author<p>A description and quantification of the surface of soil pores is important to promote our understanding of soil function because these surfaces provide sites for biological, chemical, and physical processes within the soil. Previous attempts to characterize the geometry of pore surfaces have included the slit island method, which is used to extract a fractal dimension to describe the roughness of the pore perimeters in two dimensions. Our objective in this study was to assess how this approach might be robustly applied to images of soil pore structure. Critical to the success of the technique is constancy of shape in the set of pores analyzed. We applied additional analysis to test whether this criterion is met, using the convex cover and the convexity of a two-dimensional object. This additional analysis coupled with the slit island method was applied to a soil image. In this instance, we were unable to detect a set of pores of common shape and concluded that the fractal dimension of value 1.4 derived from the image reflects not a fractal pore perimeter but rather an increasing complexity of pore shape with increasing pore size. We concluded that the analysis of pore shape proposed here, in addition to scaling of pore perimeter, is critical to sensibly describe the geometry of soil pore perimeters. © Soil Science Society of America. All rights reserved.</p>
History
School affiliated with
- University of Lincoln (Historic Research Outputs)
Publication Title
Vadose Zone JournalVolume
7Issue
2Pages/Article Number
456-460Publisher
Digital LibraryExternal DOI
ISSN
15391663Date Submitted
2018-07-25Date Accepted
2018-07-25Date of First Publication
2018-07-25Date of Final Publication
2018-07-25ePrints ID
27894Usage metrics
Categories
No categories selectedKeywords
Licence
Exports
RefWorksRefWorks
BibTeXBibTeX
Ref. managerRef. manager
EndnoteEndnote
DataCiteDataCite
NLMNLM
DCDC


