|
Descripción
|
|
|---|---|
| Many of the processes occurring in soil, like moisture retention or solute transport, are strongly influenced by the geometric structure of soil porosity [7]. In order to make prediction, models to simulate these processes need input parameters such as total porosity, pore size distribution, tortuosity, shape or orientation of soil voids, the conduits for liquid, particles and nutrients transport in soil. Classic techniques for soil structure characterization like mercury porosimetry or visualization methods using photographs of sections of soil blocks hardened with intruded resins provided indirect and partial two?dimensional information apart from being destructive techniques. X-ray computed tomography is a nondestructive technique that generate a three-dimensional image of the interior of a soil sample that facilitates the direct measurement of the geometrical attributes of the pore space structure. However, new sources of potential errors need to be handled as image filtering to reduce noise and artifacts or segmentation [4]. This work presents different tools and mathematical methods that have been developed in order to quantify the geometry of porous space of intact soil columns to characterize two soils with different managements practices. Samples of soil columns were taken at the experimental farm ?El Enc ?n?, Alcal ?a de Henares (Madrid) in November 2008 from two nearby soils subject to different managements: a natural soil (zone A) and a cultivated soil (zone B). The samples were taken in cylindrical containers of methacrylate 3 cm height and 2.6 cm or 2.1 cm diameter. Soil samples depths were 0 cm (level 1), 15 cm (level 2) and 30 cm (level 3) with four repetitions for each zone, depths and diameter, 48 samples in total. The samples were scanned at an energy level of 50 keV and resulting 16 bit images had resolution of 0,0302 mm per voxel but it was 0,0605 mm per voxel for samples with the lowest diameter. Image processing and analysis was performed using ImageJ public domain program [6]. The binarization of filtered images was carry out with isodata algorithm [10] (fig. 1 and 2) . BoneJ plugin [3], a program developed for ImageJ, provided geometrical measurements. The following geometrical attributes of the segmented images were considered: total porosity, surface area of pore space, shape and orientation [2] of connected components, number of branches of skeletons and the total length of skeleton branches, and fractal dimensions of pore space and pore-solid interface. In this study, we found statistically significant differences between the two managements practices for parameters ?area of pore?, ?orientation of the pores? and ?fractal dimension interface?. Not cultivated soil samples were also different in depth. The table 1 show a summary of the statistical differences found in the analysis. We can conclude that tomography is a good technique for obtaining information to characterize soil pore space geometric structure. It has the advantage of being non-destructive and fast. The filtering and segmentation are key steps when obtaining geometrical measurements of the pore space. Most estimates of the considered geometric attributes discriminate between samples drawn from different zones. In addition, we found differences in depth in the natural area samples while samples of the cultivated area have been uniform in depth. Fractal dimensions behaved differently, only fractal dimension of solid-void interface surface discriminates between the two groups of analyzed samples. These fractals dimensions were positively correlated with total porosity and total surface area of the solid?void interface. | |
|
Internacional
|
No |
|
Nombre congreso
|
Interpore: 3rd International Conference on Porous Media & annual meeting |
|
Tipo de participación
|
960 |
|
Lugar del congreso
|
Bordeaux (France) |
|
Revisores
|
Si |
|
ISBN o ISSN
|
|
|
DOI
|
|
|
Fecha inicio congreso
|
29/03/2011 |
|
Fecha fin congreso
|
31/03/2011 |
|
Desde la página
|
|
|
Hasta la página
|
|
|
Título de las actas
|
|