Premium
Fuzzy‐c‐Means and Kriging for Mapping Soil as a Continuous System
Author(s) -
Odeh Inakwu O. A.,
Chittleborough David J.,
McBratney Alex B.
Publication year - 1992
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1992.03615995005600060033x
Subject(s) - fuzzy logic , mathematics , soil map , kriging , fuzzy set , sampling (signal processing) , context (archaeology) , membership function , fuzzy classification , data mining , algorithm , soil science , statistics , soil water , computer science , geology , artificial intelligence , paleontology , filter (signal processing) , computer vision
Abstract Even though soil has long been recognized as a continous system in both the geographical and taxonomic context, there has been no practical approach to mapping soil as a continuum. Fuzzy set theory and the regionalized variable theory provide for such an approach. A predetermined optimal sampling scheme was used for sampling of pedons in an area of 627 by 375 m. The fuzzy‐ c ‐means (FCM) algorithm was first used to quantify the pedons into intragrade and extragrade classes by minimization of the fuzzy objective function. The resulting matrix of membership coefficients was used to determine membership semivariograms for the c + 1 continuous soil classes. The resulting semivariogram parameters were further used for kriging of the membership coefficients, yielding an isarithmic map for each of the c + 1 fuzzy soil classes. A composite map that presents the core areas of the continuous classification units was produced by overlaying of the isarithmic maps of the c + 1 classes. The significance of this approach to geographical (or soil) information systems is the associated minimum loss of information if fuzzy logic and operations are integrated into the systems.