Automated statistical method to align 2D chemical maps with 3D X-ray micro-tomography images of soils

Abstract

Recent 2-dimensional measurements reveal that soils are chemically very heterogeneous at nanometric and micrometric scales. Direct measurement techniques are still lacking to extend these observations to 3 dimensions. Sequential sectioning of soils, followed by 2-dimensional mapping of chemical elements and geometric interpolation to 3D, appears to be the only available alternative. Unfortunately, sectioning of soil samples suffers from geometric distortions that are difficult to avoid in practise. In this regard, the objective of the research described in this article was to develop a procedure enabling one to locate, in a 3D X-ray microtomographic image of a soil sample, a physical surface that is obtained by sectioning and for which a number of chemical maps are available. This procedure involves three steps: (1) the reconstitution of the physical structure of the soil layer surface, (2) the alignment of the chemical maps with the reconstituted soil surface image, and (3) the 3D alignment of the 2D chemical maps with the internal structure of the soil cube. Visual comparison of the C and Si maps and of the reconstituted CT images of the layer surfaces suggests a good correspondence between them, which is supported by Pearson correlation coefficients of - 0.57, - 0.58, 0.45, and 0.43 for the different surfaces and elements considered. Relative to the original 3D X-ray CT image of the soil sample, the planes associated with the C and Si maps, respectively, are nearly superposed, which further confirms the validity of the alignment procedure.