3494
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
Table 1. Smallest resolved particle sizes by Sedimaging and TST for
different camera resolutions.
Camera Model
Nikon
D100
Nikon
D7000
Nikon
D800
Leaf
Credo80
Year Introduced
2002
2010
2012
2012
Resolution (MP)
6.1
16.2
36.3
80.0
Resolution
(pixels × pixels)
3008
×2000
4928
×3264
7360
×4912
10320
×7752
Sedimaging
Soil Column Height (mm)
135
135
135
135
Required Magnification
(pixels/mm)
22.3
36.5
54.5
76.4
Minimum PPD for
Wavelet Analysis (pixels)
3
3
3
3
Smallest Particle
Resolved (mm)
0.134
0.082
0.055
0.039
TST
Longer TST Dimension
(mm)
910
910
910
910
Required Magnification
(pixels/mm)
3.3
5.4
8.1
11.3
Minimum PPD for
Watershed Segmentation
(pixels)
9
9
9
9
Smallest Particle
Resolved (mm)
2.7
1.7
1.1
0.8
Continuing advances in image sensor technology will yield
ever-increasing camera resolutions. This will gradually increase
the range of particle sizes that could be analyzed from a single
image. At the same time, improvements in optics will gradually
increase image magnifications. Common current methods for
increasing magnification include macro lenses, diopter rings and
extension tubes.
Table 2. Smallest resolved particles by various magnifying systems.
1)
Macro
Low Mag.
Macro
High Mag.
2)
Macro&
Diopters
3)
Macro&
Extension
Magnification
(pix/mm)
36.7
209.8
238.1
254.2
Minimum
PPD
(pixel)
3
3
3
3
Minimum
Particle Size
(mm)
0.082
0.014
0.013
0.012
1) AF-S Micro Nikkor 60 mm f/2.8G ED
2) Tiffen 62 mm close-up lens +1, +2 and +4
3) Kenko extension tube 12 mm, 20 mm and 36 mm
Table 2 lists the smallest soil particle sizes that could
theoretically be detected by wavelet analysis using various
combinations of these magnifying systems. A 60 mm macro lens
provides magnifications approaching 210 pix/mm. At this
magnification the field of view will be too small for Sedimaging
but it demonstrates that particles as small as 0.014 mm can be
detected. It is also worth noting that a magnification of 1500
pix/mm would be able to detect 0.002 mm particles, the
commonly cited silt/clay threshold. Smaller, clay-sized particles
would not be detected. The use of diopter rings and extension
tubes adds very little to the magnification achieved by the macro
lens alone. Furthermore, the authors found that diopter rings and
extension tubes decrease the image quality to the point that
measures of particle size are noticeably affected. As such, the
use of diopter rings and extension tubes is not recommended.
Higher magnifications can also be achieved with
photomicroscopy at the expense of having a very limited field of
view.
7 CONCLUSIONS
Image based systems for determining soil particle size
distributions have hitherto required that particles be detached
prior to imaging. Two new systems have been developed that
eliminate this prerequisite. The Sedimaging system determines
size distributions for soils containing particles from 2.0 mm to
0.075 mm while the Translucent Segregation Table (TST)
system is used for particles larger than 2.0 mm. Sedimaging uses
mathematical wavelets
to determine particle sizes and requires a
camera magnification that provides at least 3 pixels per particle
diameter (
PPD=3
). The TST uses
watershed analysis
to
digitally detach particles and requires
PPD
=9. Recent DSLR
camera advances have provided the requisite camera resolutions
and magnifications that have made Sedimaging and the TST
practical and cost-competitive with sieving. A much more
expensive medium format camera could simultaneously resolve
particles spanning nearly two orders of magnitude in diameter,
from 2.0 mm to 0.039 mm. Particles as small as 0.014 could be
resolved with less expensive DSLR cameras and macro lenses
but at the expense of limiting the field of view. With soon
available DSLR camera resolutions over 40 MP, the TST system
will size particles as small as 1.0 mm. As a result, Sedimaging
would be relieved of evaluating particles in the 2.0 mm to 1.0
mm range and thus its physical size could be reduced by 50%.
8 ACKNOWLEDGEMENTS
This material is based upon work supported by the U.S. National
Science Foundation under Grant No. CMMI 0900105. ConeTec
Investigations Ltd. and the ConeTec Education Foundation are
acknowledged for their support to the Geotechnical Engineering
Laboratories at the University of Michigan..
9 REFERENCES
Brown D.J., Vickers G.T., Collier A.P. and Reynolds G.K. 2005.
Measurement of the Size, Shape and Orientation of Convex Bodies.
Chemical Engineering Science
60, 289-292.
Fletcher T., Chandan C., Masad E. and Sivakumar K. 2003. Aggregate
Imaging System for Characterizing the Shape of Fine and Coarse
Aggregates.
Transportation Research Record
1832, 67-77.
Ghalib A.M. and Hryciw R.D. 1999. Soil Particle Size Distribution by
Mosaic Imaging and Watershed Analysis.
Journal of Computing in
Civil Engineering
13(2), 80-87.
Hryciw R.D. and Ohm H.S. 2012. Feasibility of Digital Imaging to
Characterize Earth Materials. A research report submitted to
Michigan Department of Transportation, ORBP Number ORE0908.
Jung Y. 2010. Determination of Soil Grain Size Distribution by Soil
Sedimentation and Image Processing. Ph.D. Dissertation, The
University of Michigan, Ann Arbor.
Masad E. and Tutumluer E. 2007. Test Methods for Characterizing
Aggregate Shape, Texture, and Angularity. NCHRP Report 555,
Transportation Reserach Board.
Ohm H.S., Sahadewa A., Hryciw R.D., Zekkos D. and Brant N. 2012.
Sustainable Soil Particle Size Characterization through Image
Analysis.
Proceedings of the 17th Great Lakes Geotechnical and
Geoenvironmental Conference
, Cleveland, Ohio, 26-33.
Rao C. and Tutumluer E. 2000. Determination of Volume of Aggregates.
Transportation Research Record
1721, 73-80.
Shin S. and Hryciw R.D. 2004. Wavelet Analysis of Soil Mass Images
for Particle Size Determination.
Journal of Computing in Civil
Engineering
18(1), 19-27.
