Ew. Tollner, MEASUREMENT OF DENSITY AND WATER-CONTENT IN SOILS WITH X-RAY LINESCANAND X-RAY COMPUTED-TOMOGRAPHY, Transactions of the ASAE, 37(6), 1994, pp. 1741-1748
Potentials for using X-ray computed tomography for determining soil bu
lk density and water content were investigated using Cecil sandy clay
and Lakeland sand soils. Soils were moistened and samples were compact
ed to several levels. These samples were sequentially dried and scanne
d using tomography reconstructions as well as the linescan mode availa
ble on a CT scanner. Mean X-ray absorption and pixel standard deviatio
n data were collected for each soil sample at scanning in both modes.
For each soil-mode combination, relationships between mean X-ray absor
ption and soil bulk density and water content were determined as were
relations between pixel standard deviation and soil density and water
content. The experiment was then repeated to provide validation data.
In the tomographic reconstruction mode with validation data, optimal w
ater content and bulk density values for samples were determined for e
ach corresponding mean absorption and pixel standard deviation measure
ments which minimized the error sum of squares. Density was predicted
in both soils to within 2%. Predicted water content was within 50% of
observed values. Using known values of density, water content war pred
icted to within 5%. In the linescan mode, density was usually predicte
d to within 5%. Water content was not significant (P < 0.15%) thus a p
rediction for water was not attempted in the linescan mode.