Turkish Journal of Engineering – 2023, 7(2), 108-115

Turkish Journal of Engineering

https://dergipark.org.tr/en/pub/tuje
e-ISSN 2587-1366

An alternative method for the particle size distribution: Image processing

Mert Aydın *1 , Talas Fikret Kurnaz 2
1 Mersin University, Department of Civil Engineering, Türkiye
2 Mersin University, Department of Transportation Services, Türkiye

Keywords Abstract
İmage processing technique Granular soils are used in different areas of civil engineering due to their easy accessibility and
Particle distribution low cost as a material. The sieve analysis method, which has been conducted in rock physics
Particle size measurements for many years, has been practicing determining the particle size distributions
Sieve analysis for those materials. In this study, a method based on image analysis technique has been
developed as an alternative to traditional sieve analysis method for determining the particle
Research Article size distribution in granular soils. This technique based on image processing to determine
DOI: 10.31127/tuje.1053462 particle distributions via the experimental setup that consist of a camera, tripod, light box and
mechanical shaking apparatus. In order to assess the reliability of this technique, each sample
Received: 04.01.2022 was subjected to traditional sieve analysis and the results of both analysis methods were
Accepted: 11.02.2022 compared. In conclusion, it was observed that the results obtained with the image processing
Published: 26.04.2022 technique had a minimum 95.22% closeness with the sieve analysis experiment data.

1. Introduction non-destructively, correlating the performance of

coatings with the materials used. Obaidat [4], aimed to
Digital image processing method is simply the measure void ratios of mineral aggregates in bituminous
process of transferring the sample that willing to be mixtures using image processing mechanism and
analysed with an image capture instrument in a suitable planimeter. Gaydecki et al. [5], proposed to find the
format for digital media and by processing the raw image location of the reinforcements in the reinforced concrete
making it suitable for further analysis, and lastly samples, in a non-destructive way through image
collecting and storing the desired data. This method, analysis. Soroushian et al. [6], used digital image
whose foundations were laid with the examination of processing techniques to detect voids and micro cracks
microscope images in the 1960s, has increased the areas in the concrete sample. Soroushian and Elzafraney [7],
of use day by day with the development of both studied the microstructure of concrete samples with the
computers and microscopes technologies. Digital image images they obtained with scanning electron microscopy
processing method has been used in many fields such as and fluorescence microscopy. Felekoğlu and Güllü [8],
the analysis of cells, organelles, organisms in medical- scanned clinker samples using an optical microscope and
based research, weather forecast in meteorology, the performed areal phase and porosity distribution analysis
derivation of technologies such as fingerprint, retina with image processing techniques. Sinha and Fieguth [9],
scanning, night vision in the defence industry, tracking investigated the usability of image processing techniques
the habitats in environmental science, observing and to detect deformations in concrete pipes used as
controlling in industrial production lines [1-3]. infrastructure elements and developed algorithms to
Digital image processing technique within the scope detect deformations in underground concrete pipes that
of civil engineering has been used for various purposes can work with image processing techniques. Park et al.
since the 90's such as strength, permeability, [10], aimed to develop an algorithm that allows the
compressibility of soils, shape parameters and detection of air gaps with image processing technique by
deformations of materials, examining cross sections of examining the colored images of the polished sections of
concrete samples, measuring the maximum strength concrete samples.
* Corresponding Author Cite this article
*([email protected]) ORCID ID 0000-0003-0620-0954 Aydın, M. & Kurnaz, T. F. (2023). An alternative method for the particle size distribution:
([email protected]) ORCID ID 0000-0003-2079-8315 Image processing. Turkish Journal of Engineering, 7(2), 108-115
 Turkish Journal of Engineering – 2023, 7(2), 108-115

Granular materials are used in many areas due to behavior of clay fillings under dynamic compaction using
their easy accessibility and low costs within the scope of image processing techniques. Edizer [24], explained the
civil engineering. They are being used as the main image processing technique comprehensively and
materials in both reinforced concrete and masonry performed grain size analysis with this technique by
structures, in addition to that they can be used for load using an open-source software. Sezer [2], used image
transfer in steel and wooden structures. Concrete, one of analysis techniques to determine the microstructural
the most common products, consists of aggregates properties of granular soils. Önal (2008), developed
between 60% -80% by volume, depending on the algorithms to directly use image processing technique in
purpose of use. As result of that when evaluating the final the scope of geotechnical engineering. Vangla et al. [25],
strength of concrete, the importance of both physical and claimed that digital image processing technique for size
chemical parameters of the aggregate used reveals its and shape analysis of sand particles would be more
importance. The wide area of usage of granular materials consistent than traditional sieve analysis method. Ehsan
-in the scope of excavation / filling in various et al. [26], used image processing techniques for grain
infrastructure and superstructure works, dam size analysis of granular soils. Dipova [13], arranged a
foundations and bodies, the leveling and filling area on new experimental setup to examine the grain
bituminous hot-coated roads, applications such as distribution of grained soils with the image processing
jetcrete and grout in slope stabilization or tunnel technique.
applications- require the appropriate material selection. In this study, digital image processing technique has
This selection happens with the help of a wide range of used as an alternative to sieve analysis method for the
parameters such as specific gravity, chemical structure, determination of the particle distribution in granular
mineralogical properties, permeability, cohesion values, soils. In order to capture the images, an experimental
internal friction angles [11-13]. The grain diameter size setup has developed and images have taken to determine
is the most common filter in making this kind of material the particle distribution of samples of different volumes
selection. Sieve analysis method is used to determine the and different origins. The same samples have examined
grain diameter distribution of grained materials. The by sieve analysis method and the results of the two
sieve analysis, which has taken its final form today, is a methods have compared. Inspired by the extensive
universal method of classifying granular materials literature study and its predecessors, the experimental
according to their size and weight. Testing procedures, setup was arranged in the light of the antecedent theories
that accepted by organizations such as ASTM, BS, and it was aimed to conduct an experimental study. By
AASHTO are being carried out with ISO standard using a simple mechanism and user-friendly software, a
equipment and instruments. The information which system that is compatible with automation and does not
obtained from the sieve analysis results are the most require the intervention and interpretation of the user
important parameters in processing and classifying the has been designed.
grained material [14]. Some researchers have used image
processing technique to determine grain diameter 2. Image Processing Technique
distributions as an alternative to the sieve analysis
method, which can be described as relatively difficult and
inconvenient. Yue et al. [15] studied the distribution, Digital image processing is the examination of the
orientation and grain shape concepts of coarse particles image captured within the scope of the intended analysis
in asphalt-added samples and concrete samples with the by using a processor. The sample to be analysed has
help of image processing technique. NG [16], studied the transferred to the digital environment under suitable
post-compaction microscopic structures of grained soils conditions, with a suitable tool that allows analysis, in the
using the split element method. Masad and Button [17], desired file extension, with the desired visual properties
examined the possibility of analyzing the grains' shape (brightness, resolution, colour, contrast, etc.). The tool
parameters with two different methods with using image should be capable of capture images such as digital or
processing technique by photographing the grained analog cameras, microscopes, telescopes, and video
samples at different resolutions. Mora and Kwan [18], recorders and transfer the captured image to the
developed a method that analyzes the sphericity, analysing environment. The captured image could be
convexity and shape parameters of the coarse aggregate filtered by parameters such as colour, brightness,
using image processing technique. Yue et al. [19], used contrast, sharpness, size, if necessary. The purpose of this
image processing techniques and the finite element process is to eliminate noises on image that occurs
method together to perform comprehensive analysis of during or post-image capturing phase and to prepare
two-dimensional shape of grained materials. Al Rousan optimized input for analysis. Analysis has performed by
[20], aimed to classify the grained materials according to existing software or algorithms developed by the user. In
their shape parameters with a computer-aided subjects such as classification, measurement, statistical
automation system. Alshibli and Alsalah [21], examined studies, database creation; It is essential to use a
the sand soil samples under a microscope and purposeful analysis algorithm in this method, which can
statistically analyzed by their surface roughness, be used for several fields such as medicine, genetics,
sphericity and roundness values. Yang [22], claimed that industry, electronics, textile and space research. For
image processing researches in scope of soil mechanics example, since cell organelles were willing to be
were two-dimensional and aimed to make a three- examined, it was necessary to capturing images in
dimensional analysis by filling the voids of the soil microscopic dimensions, in black and white or colour
samples with epoxy. Hu et al. [23], researched the according to the purpose to examinate the organelle

109
 Turkish Journal of Engineering – 2023, 7(2), 108-115

shapes or textures and making comprehensive and Even though particle size distributions are
appropriate choices for correct analysis results [24,27- determined by sieve analysis method, there were still
29]. some uncertainties about the shape of the grains. The
Image processing technology is the conversion of data size of a spherical particle may be described in one
into various readable format and processing after dimension; however, the grains' shapes in the soil sample
capturing, measuring and evaluating. The concept of are generally irregular, not uniform. In a sieve analysis
image is defined as a two-dimensional function in the test, the size of a particle is related to the size of the
format "f (x, y)". Here, x and y values are representing square apertures at mesh which the particle passes
coordinates, while f function represents the intensity of through. Since all axial sizes of a particle cannot be
the parameter [30]. measured and the shape of the particle is not taken into
It could be said that the different stages of the digital account, the method also has some limitations [32-36].
image processing process take place together or
independently from each other, that different targeted 4. Materials and Methods
methods may be created and associated with algorithms.
The method, which allows working multidisciplinary, 4.1. Soil Samples
also allows to include desired available parameters to the
analysis. Six different test samples from four different origins
have created for the study (Figure 1). The samples have
3. Particle Size Distribution obtained from different regions of Mersin province
(Turkey). The first three samples have consisted of
Particles that have evaluated within the scope of the crushed stones taken from Mersin/Silifke region. Three
soil mechanics could be examined with the help of a different samples containing different particle diameter
naked eye or a magnifying glass over a diameter of 0.075 ranges have prepared by separating them manually.
mm, while particles above this diameter are considered Thus, soil batches with the same surface and shape
as coarse materials. Particles smaller than 0.075 mm in parameters have examined in different particle diameter
diameter have classified as fine grains. It is possible to ranges. Sample No.4 has prepared with crushed stone
examine the particles between 0.075 mm and 2 µm in taken from Mersin/Toroslar region, and this sample's
diameter with the help of a microscope. Particles with a shape parameters and surface texture distinguish itself
diameter range of 2 µm to 0.1 µm could also be examined from the first three samples. Sample No.5 is a white-
with a microscope, but for accurate analysis of the colored, grained material have taken from marble
texture and shape parameters of the grains, it is quarries in Mersin/Erdemli region. By using this
recommended to examine the particles under 1 µm material, also known as mosaic powder, the suitability of
particle diameter by electron microscope. Examination image processing techniques for analysis of different
of the molecular structure of the particles is possible with colored materials has been investigated. The sixth
X-ray analysis [2]. sample contains particles that could be regarded as more
The part of the soil that could be classified as coarse round, taken from the Mersin/Adanalıoglu region and
consists of sand and aggregates. While gravels, which has has consisted of particles with lower unit weight and
defined as rock particles, may consist of one or more higher porosity than other samples. After the samples
minerals, sand particles generally consist of a single have grouped manually, they were washed and dried in a
mineral, quartz. When the particles of sand and gravel 110±5 degree celsius stove for 24 hours. After the drying
have examined in terms of shape details, they were process each sample has weighted and recorded as batch
divided into five sub-groups; angular, semi-angular, weight. The index properties of the samples are given in
semi-round, round, fairly round. Silt and clay particles Table 1.
may be angular, flat or needle shaped.
The sieve analysis method is a widely used technique
for determining the particle size distribution of granular
materials. By the particle size distribution information of Table 1. The index properties of the samples
the soil sample, it is able to gauge properties of soil such
Soil Classes
Sample No

D10 (mm)

D30 (mm)

D60 (mm)

as the water permeability, strength, compressibility of

the soil, determination of capillary water movements,
Cu

Cr

frost effect, compression speed under load.

The sieves used in the sieve analysis consist of square
grids of equally spaced wires. The aperture size varies 1 6.60 9.93 15.20 2.30 2.82 GP
inversely with the sieve number [31]. Sieve analysis 2 0.57 1.45 2.08 3.65 0.28 SP
starts with stacking of sieves one on top of the other in
decreasing aperture order. The sieve set has shaken 3 2.01 5.44 7.81 3.89 0.98 GP
uniformly for 15 minutes, preferably with the shaking
4 4.31 5.72 7.94 1.84 2.24 GP
device. The sample remaining on each sieve has weighed
separately after shaking process. Precise shaking time is 5 1.82 3.03 4.68 2.57 0.76 GP
needed, the short shaking process does not allow small
6 1.29 1.86 3.02 2.34 0.49 GP
particles to pass into the lower sieves, on the other hand
the long shaking process may cause deformation of the
particles and deflects the results [31].

110
 Turkish Journal of Engineering – 2023, 7(2), 108-115

reduced to a format that contains two values in the range

selection that the program automatically assigns. Pixels
above the entered threshold value and pixels below the
threshold were separated and particles and the
background has reduced to binary code. From the
program interface, the values requested by the user were
selected and outputs have obtained at the end of the
analysis. MS Excel has used for further analysis in the
study.

Table 2. Numerical information of the samples

Fmin (mm)
Fmax (mm)
Number of
Number of
Sample No

Weight (g)

Grain

Shot
1 1768.00 1459 2 41.418 2.580

2 158.42 121162 5 8.393 0.166

Figure 1. Samples prepared for the study
3 646.07 31507 4 22.050 0.500
4.2. Particle Size Distribution with Image Processing
4 1440.96 7987 5 27.404 0.478
An experimental setup has developed within the 5 326.71 17766 7 14.119 0.302
scope of determining the particle distribution in granular
6 104.25 24217 5 18.823 0.401
soils by image processing technique. In the method, each
sample whose weights have recorded was laid on a 60 *
60 cm lightbox. Images have captured with a Nikon D90
body, 18-55mm Vr Nikon lens with 4288*2848
resolution in 24 bits. Images have taken in dark room
conditions with the help of a tripod by adjusting the
height of the camera for proper framing (Figure 2). In
order to prevent the contact and overlap of the particles
in the sample laid on the lightbox, a shaking device has
placed under the lightbox during image capturing, the
photos were taken simultaneously by shaking the
lightbox with a spring mechanism's impact. The image
capturing setup has shown in Figure 3. All photos have Figure 2. Image acquisition environment
taken at f/5.6 aperture, 1/1250 shutter speed, 105mm
focal length and without flash. Denser samples have
needed more than single shot for more accurate analysis.
The number of shots taken for the samples and the
number of particles analyzed in the samples have given
in Table 2. In cases where more than one photo needed
to be analyzed, the algorithm that used for single analysis
has exported and rearranged to recall and analyze
different photos sequentially. With a simple macro
arrangement, it has ensured that more than one
photograph was analyzed in a single sample, and results Figure 3. The setup for image capturing
were obtained as a single document.
Image Pro Plus version 5.1.0.20 software has used for Since the analysis was related to the particle size, the
image processing. The first step in the program was parameters have examined in the analysis are related to
opening the photograph to be analyzed from the program the size. The area covered by the particle in the
interface. A pixel to millimeter conversion has required photograph has examined in square millimeter. In
in order to obtain the results in the metric system in the addition, Fmax and Fmin values, which are the side
analysis made on the photograph. This requirement was lengths of the smallest size surrounding quadrilateral,
met by an extra shot after a length measuring tool (ruler, were the parameters used in the analysis (Figure 4). The
tape measure, etc.) has placed on the lightbox. Unit fact that the image was transferred to the digital
distance-pixel conversion has performed using the environment in two dimensions and the reference
measurement tool from the program interface. The experiment being a three-dimensional mechanical
opened photo was in 24-bit RGB format, but since 8-bit experiment have posed a problem that must be overcome
grayscale would be sufficient for analysis, reduction for further comparisons. The data obtained by image
process has performed through the software. processing were information covering two dimensions as
Afterwards, this grayscale photo has masked and shown in Figure 4. Researchers who have previously

111
 Turkish Journal of Engineering – 2023, 7(2), 108-115

conducted similar studies have derived different curves have drawn for all sample batches. In the
correction coefficients using various parameters to make equation, the projection area of the particle has
the method consistent [1,18]. Even if it was accepted that multiplied by the value of Fmin and the possibility of
the particles of the sieves are limited by two axes, one adding a third dimension to the equation is examined and
more parameter has included in the analysis with the the results have evaluated.
weight measurements made after the sieving process.
Within the scope of this information, it was predicted 4.3. Sieve Analysis
that the data obtained from image processing, including
two dimensions, may not be sufficient for an accurate The samples have examined by image processing
dimensional analysis, and a new parameter has technique were subjected to the sieve analysis test with
investigated [1,15]. the traditional sieve analysis test procedure without any
loss. The sieve analysis results have recorded with
standard procedures and forms, and the gradation
curves of the samples have drawn. Since the particle size
distribution of the samples as different, the experiments
have carried out by selecting the most suitable sieve set
for each batch. The selection was made by the operator.
The use of different sieve sets made it possible to
compare and interpret the results of the image
processing method in various diameters and grain
distributions. Applications using sieve set, washing
Figure 4. Image processing analysis parameters and container, oven, scales and shaking devices have carried
display of three dimensions of particle [1,37] out in accordance with the descriptions and procedure
given in TS 3530 [31].
Mora et al. [1] encountered the necessity of shifting
the gradation curves for each sample when comparing 5. Results
the results of mechanical analysis and image processing.
In this context, they ensured that the curves have Both sieve and image processing results of 6 samples
approximated to the sieve analysis curves by assigning a have obtained at the end of the study and the results have
"C" correction coefficient. The fact that the correction analyzed comparatively. While the gradation curves of
coefficient assigned here was at the user's initiative for the samples subjected to traditional sieve analysis have
each sample could lead to subjective errors. For drawn following the standard procedure, the data related
consistent analysis this variable, which was not based on to the visuals examined with the Image Pro Plus software
a constant and allows interpretation and abuse, have had have transferred to the MS Excel program for further
to be excluded from the analysis [1,18]. analysis. Grading curves were drawn by using MS Excel
The results were obtained from the image processing after filtering and basic calculations. The sieve analysis
performed within the scope of the experiments in this results have taken as a reference, and the results
study have included in the further analysis by using the obtained by image analysis techniques have displayed on
raw data without using any correction coefficient. By a single graphic to compare both curves (Figure 5-10).
following this method, the possibility of data
interpretation and intervention has eliminated during 6. Conclusion
processing. A value related to the shape (sphericity,
angularity) and surface parameters of the particles has In the scope of the study, an experimental setup has
not been included in further investigations. This has led been designed to apply the image processing technique
to the elimination of user intervention. as an alternative to the relatively complicated and
The obtained image processing data were filtered by challenging sieve analysis method. As a result, the image
accepting the hypotenuses of the meshes in the sieves as processing technique, which could analyze without the
the limit, by using the standard sieve sizes. The aim here need for sieve analysis equipment and laboratory, has
was to classify the particle diameters by using the sieve yielded similar results to traditional sieve analysis.
diameters and Fmax parameters, as in the process of When the results have evaluated on the basis of sieves, a
weighing the materials remaining on the sieve in the maximum absolute difference of 4.78% was encountered
sieve analysis. Theoretically this approach has made the and the validity of the applied image processing
comparison possible between image processing and technique has shown.
sieve analysis in scale of sieve by sieve. By commenting When the materials of different origins used in the
on comparison results, evaluation of rate of success has study have examined in different diameters and volumes
been made. and compared with the traditional method, it has
observed that curves consistent with the reference
∑𝑝İ=1(𝐴𝑟𝑒𝑎 ∗ 𝐹𝑚𝑖𝑛 ) curves were drawn in all six samples. This proves that the
% Sieve Passing = (1) image processing technique is valid for all the particles in
∑𝑛İ=1(𝐴𝑟𝑒𝑎 ∗ 𝐹𝑚𝑖𝑛 )
different texture, color, unit weight and shape used
Calculations has made for each particle individually in within the scope of the research.
all particle samples by using Equation 1, and then grading While the image processing method benefits the
researcher from time and labor, the algorithms used

112
 Turkish Journal of Engineering – 2023, 7(2), 108-115

were not valid for every sample and a repeat loop have It is possible to develop an automation for particle
has to be created for each sample. Nevertheless, this was analysis with image processing techniques with a
not preventing the creation of a database according to database created by statistical analysis including
sample types and transition to automation. In addition to previous studies. With the developed automation, an
the necessity of the presence of computers, cameras and experimental method which excludes user intervention
lighting equipment, the researcher's has to be familiar and manipulation may be derived.
and able to use of these equipment was another
requirement.

Figure 5. Comparison of results for Sample No.1

Figure 6. Comparison of results for Sample No.2

Figure 7. Comparison of results for Sample No.3

113
 Turkish Journal of Engineering – 2023, 7(2), 108-115

Figure 8. Comparison of results for Sample No.4

Figure 9. Comparison of results for Sample No.5

Figure 10. Comparison of results for Sample No.6

Acknowledgement Digital Image Processing. Cement and Concrete

Research, 28 (6), 921-932.
The authors are thankful to Bünyamin Demir and 2. Sezer, A. (2008). Determination of Microapical
Bahadir Sayinci for the contributions for the study. Properties of Different Types of Soils Using Image
Processing Techniques, PhD Thesis, İzmir: Ege
Author contributions University Graduate School of Sciences, Department
of Civil Engineering (in Turkish).
Mert Aydin: Methodology, Software Talas Fikret 3. Murthy, V. N. S. (2002). Geotechnical Engineering,
Kurnaz: Writing-Original draft preparation, Principles and Practices of Soil Mechanics and
Investigation, Writing-Reviewing and Editing. Foundation Engineering, CRC Press.
4. Obaidat, M. T., Al-Masaeid, H. R., Gharaybeh, F., &
Conflicts of interest Khedaywi, T. S. (1998). An Innovative Digital Image
Analysis Approach to Quantify the Percentage of
The authors declare no conflicts of interest. Voids in Mineral Aggreagates of Bituminous Mixtures.
References Canadian Journal of Civil Engineering, 25:1041-1049.
5. Gaydecki, P., Silva, I., Fernandes, B. T., Yu, Z. Z. (2000).
A Portable Inductive Scanning System for Imaging
1. Mora, C., Kwan, A., Chan, H. (1998). Particle Size Steel-Reinforcing Bars Embedded Within Concrete.
Distribution Analysis of Coarse Aggregate Using Sensors and Actuators, 84:25-32.

114
 Turkish Journal of Engineering – 2023, 7(2), 108-115

6. Soroushian, P., Elzafraney, M., Nossoni, A. (2003). microscopy. ASCE Journal of Computing in Civil
Specimen preparation and image processing and Engineering, 18 (1), 36-45.
analysis techniques for automated quantification of 22. Yang, X. (2005). Three Dimensional Characterization
concrete microcracks and voids. Cement and of Inheret and Induced Sand Microstructure, Phd
Concrete Research, 33:1949-1962. Thesis, Georgia Institute of Technology.
7. Soroushian, P., Elzafraney. M. (2005). Morphological 23. Hu, R., Yue, Z., Tham, L., & Wang, L. (2005). Digital
operations, planar mathematical formulations, and image analysis of dynamic compaction effects on clay
stereological interpretations for automated image fills. ASCE Journal of Geotechnical and
analysis of concrete microstructure. Cement & Geoenvironmental Engineering, 131 (11), 1411-
Concrete Composites, 27:823-833. 1422.
8. Felekoglu, B., Güllü, D. (2006). Use of optical 24. Edizer, E. (2006). Particle Size Distribution with
microscope and image processing techniques in Digital Image Processing Method, Master Thesis,
clinker analysis. IMO Technical Journal, 247, 3761- Adana: Çukurova University, Institute of Science and
3770. Technology.
9. Sinha, S. K., & Fieguth, P. W. (2006). Automated 25. Vangla, P., Roy, N., Mendu, K., & Latha, G. M. (2014).
detection of cracks in buried concrete pipe images. Digital Image Analysis for the Determination of Size
Automation in Construction, 15, 58-72. and Shape Parameters of Sand Grains. Golden Jubilee
10. Park, S. H., Kim, H. K., Morales, A., & Ko, S. J. (2007). Conference of the IGS Bangalore Chapter, Geo
Air-void analysis system of polished concrete using Innovations, Bangalore India.
image processing. Journal of Applied Computer 26. Ehsan, K. M., Hossain, M. R., Manzur, T., Shohag, A., &
Science, 15 (2), 19-25. Tabassum, N. (2016). Particle Size Analysis by Image
11. Ozen, M. (2007). Investigation of Relationship Processing Technique. 1st Bangladesh Civil
Between Aggregate Shape Parameters and Concrete Engineering Summit, Dhaka, Bangladesh.
Strength Using Imaging Techniques, Master's Thesis, 27. Jahne, B. (2005). Digital Image Processing, Germany:
Ankara: Middle East Technical University, The Springer.
Graduate School of Natural and Applied Sciences. 28. Gonzales, R. C., & Woods, R. E. (2001). Digital Image
12. Comak, B., Beycioğlu, A., Başyigit, C., Kılınçarslan, S. Processing, USA: Prentice Hall.
(2011). Use of Image Processing Techniques in 29. McAndrew, A. (2004). An Introduction to Digital
Concrete Technology. 6th International Advanced Image Processing with MATLAB, Melbourne: Victoria
Technologies Symposium (IATS'11), Elazığ. University of Technology.
13. Dipova, N. (2017). Determining The Grain Size 30. Gonzales, R. C., & Woods, R. E. (2008). Digital Image
Distribution of Granular Soils Using Image Analysis. Processing, Upper Saddle River, NJ: Pearson Prentice
Acta Geotechnica Slovenica, 1, 29-37. Hall, 1-3.
14. German, R. M. (1994). Powder Metallurgy Science, 31. Turkish Standards Institute TS 3530 EN 933-1/April
New Jersey: Metal Powder Industries Federation. (1999). Experiments for Geometric Properties of
15. Yue, Z. Q., Bekking, W., Morin, I. (1995). Application of Aggregates Part 1: Determination of Particle Size
digital image processing to quantitative study of Distribution-Sieving Method
asphalt concrete microstructure. Transportation 32. Mertens, G, & Elsen, J. (2006). Use of computer
Research Record, 1492:53-60. assisted image analysis for the determination of the
16. NG, T. T. (1999). Fabric Study of Granular Materials grain- size distribution of sands used in mortars.
After Compaction. ASCE Journal of Engineering Cement and Concrete Research 36:1453-1459 DOI:
Mechanics, 125 (12), 1390-1394. 10.1016/j.cemconres.2006.03.004.
17. Masad, E., & Button, J. (2000). Unified Imaging 33. Fernlund, J. M. R., Zimmerman, R. W., & Kragic, D.
Approach for Measuring Aggregate Angularity and (2007). Influence of volume/mass on grain-size
Texture. Computer-Aided Civil and Infrastructure curves and conversion of image-analysis size to sieve
Engineering, 15(4), 273-280. size. Engineering Geology, 90:124–137 DOI:
18. Mora, C. F., & Kwan, A. K. H. (2000). Sphericity, shape 10.1016/j.enggeo.2006.12.007.
factor, and convexity measurement of coarse 34. Allen, T. (1968). Particle size measurement, Chapman
aggregate for concrete using digital image and Hall, London.
processing. Cement and concrete research, 30(3), 351- 35. Pettijohn, F. J. (1984). Sedimentary Rocks. S.K. Jain,
358. India.
19. Yue, Z. Q., Chen, S., & Tham, L. G. (2003). Finite 36. Fernlund, J. M. R. (1998). The effect of particle form
element modeling of geomaterials using digital image on sieve analysis: a test by image analysis.
processing, Computers and Geotechnics, 30, 375-397. Engineering Geology 50:111-124 DOI:
20. Rousan, T. M. A. (2004). Characterization of Aggregate 10.1016/S0013- 7952(98)00004-0.
Shape Properties Using a Computer Automated 37. Erdogan, S., & Fowler, D. (2005). Determination of
System, Phd Thesis, USA: Texas A&M University. aggregate shape properties using x-ray topographic
21. Alshibli, K., & Alsaleh, M. (2004). Characterizing methods and the effect of shape on concrete rheology.
surface roughness and shape of sands using digital İnternational Center for Aggregate Research,
ICAR:106-1.

© Author(s) 2023. This work is distributed under https://creativecommons.org/licenses/by-sa/4.0/

115