Journal of Food Engineering 94 (2009) 366–371

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier.com/locate/jfoodeng

Bread crust thickness measurement using digital imaging and L a b colour system
Y.M. Mohd Jusoh a,b, N.L. Chin a,*, Y.A. Yusof a, R. Abdul Rahman a,c
a
Department of Process and Food Engineering, Faculty of Engineering, Universiti Putra Malaysia, 43400 UPM, Serdang, Selangor, Malaysia
b
Department of Bioprocess Engineering, Faculty of Chemical Engineering and Natural Resources Engineering, Universiti Teknologi Malaysia, 81310 UTM, Skudai, Johor, Malaysia
c
Department of Food Technology, Faculty of Science and Technology, Universiti Putra Malaysia, 43400 UPM, Serdang, Selangor, Malaysia

a r t i c l e i n f o a b s t r a c t

Article history: A simple and new method was developed for the evaluation of baking process on bread quality through
Received 19 September 2008 the measurement of bread crust thickness. By distinguishing the crust and crumb regions of bread, the
Received in revised form 20 March 2009 system which uses digital imaging and the L a b colour system can predict bread crust thickness from
Accepted 4 April 2009
the colour measurements of bread surface browning. Standard baking tests were conducted at different
Available online 17 April 2009
levels of temperature and time combinations to produce open breads with different crust thickness. The
results show that the crust thickness which ranged from 6.02 to 9.00 mm has a negative relationship with
Keywords:
each of the L, a, and b values and a positive correlation with the total colour difference (DE) of bread crust.
L a b values
Crust thickness
The data also demonstrated that crust thickness increases with the investigated baking temperatures of
Bread 185, 195, and 205 °C more significantly (p < 0.0001) than baking times of 25, 30 and 35 min (p < 0.001).
Digital imaging Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction
During the baking process, dough experiences major physical
and biochemical changes due to heat exposure which lead to the
transformation from raw dough to bread with two distinctive
structures i.e. the crust and the crumb. The crust is associated to
the brown surface of bread while crumb is the inner white spongy
structure beneath the crust. The crust, which is formed through
Maillard reaction and caramelisation during baking, has several
important functions on bread properties. The thickness and charac-
teristics of the crust to a large extent define the product and give
its name (Wiggins, 1999; Cauvain, 1999). In general, bread crust
is referred as a marketing tool that attracts customers through its
appearance, aroma, and flavour (Zenthenbaur and Grosh, 1998;
Purlis and Salvadori, 2007). Food products which their appearances
have been used as a guide to estimate its overall quality are fruits,
vegetables, grains, meat, seafood and bakery products (Pearson,
1996; Chao et al., 2002; Blasco et al., 2007; Zou et al., 2007).
Abdullah et al. (2000) inspected the quality of muffins by using
an external property, i.e. colour. In terms of actual crust functions,
the formation of bread crust is imperative as it contributes to its
aroma, flavour and texture while influencing its baked volume, cell
structure and density (Zhang et al., 2007). An early crust formation
limits bread expansion, causing formation of coarser bread struc-
ture due to cell rupture and coalesces, and also causes densification
within the crumb. Besides that, bread crust is also known to be clo-
sely related with moisture loss of bread during and post baking
* Corresponding author. Tel.: +60 3 89466353; fax: +60 3 86567123.
E-mail addresses: chinnl@eng.upm.edu.my, cnyukling@gmail.com (N.L. Chin).
periods. The crust formation affects the amount of moisture evap-
orating from wet dough during the baking process as a thicker
crust is produced with a higher moisture loss in bread (Wiggins,
1999). This is because crust formation develops simultaneously
as moisture evaporates during the baking process. Moisture loss
during baking translates to weight loss of bread and this is less
favourable for breads sold by its weight. For post baking periods,
crust functions as an insulator that prevents moisture from migrat-
ing to surrounding thus may give an impact towards reducing
crumb staling. Detailed studies by Wahlby and Skjoldebrand
(2002) showed that crust functions as weight loss barrier since
moisture loss of crustless bun was higher compared to crusted
bun during a reheating treatment. Their observation was strength-
ened by Primo-Martins et al. (2006), who showed that bread with
crust experienced lower moisture loss in comparison to crustless
bread during storage. In relating bread crust properties in terms
of its colour with moisture loss, Purlis and Salvadori (2007) pre-
sented a strong correlation between the moisture loss and the
crust colour formation in their study on browning kinetics of
bread. From these literatures, it seems viable to control moisture
loss from bread during baking and storage through the creation
of a desired crust with suitable properties.
There are various attempts to define and measure bread crust
although to date, there is not any widely accepted and approved
definition for bread crust and crumb. Attempts to define crust
and crumb based on bread’s physical structure, i.e. density and
behaviour are found. Jefferson et al. (2006) defined crust as a part
of bread near its surface where the density is highest in bread
structure. Lostie et al. (2004) interpreted crust and crumb based
on their behaviour where crumb behaves as viscous compressible

0260-8774/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2009.04.002
 Y.M. Mohd Jusoh et al. / Journal of Food Engineering 94 (2009) 366–371
mixture while crust acts like a porous shell. Gallagher et al. (2003)
and Zhang et al. (2007), using the scanning electron microscope
(SEM), defined crust and crumb based on the difference of their
physical structure such that crust has a denser structure compared
to the crumb. In measuring crust thickness, Zanoni et al. (1992)
used quick-freezing to separate the crust from crumb since both
have different structure that easily tears apart from each other dur-
ing thawing. Other methods include the visual techniques using
scanning electron microscope (SEM) by Gallagher et al. (2003)
and the confocal laser scanning microscopy (CLSM) by Primo-Mar-
tin et al. (2006). The disadvantage of quick-freezing method is that
it is prone to damage the bread structure while the SEM and CLSM
require high technology and expensive equipment.
The application of computer vision, digital imaging and colour
system to determine food qualities is becoming more popular
(Brosnan and Sun, 2004; Pedreschi et al., 2006). Images of food
can be conveniently captured and its colours are determined using
various systems, e.g. the L a b, RGB (red, green, blue), XYZ or the
CMYK (cyan, magenta, yellow, black) (Pedreschi et al., 2006) to
make correlations with food properties. Yam and Papadakis
(2004) designed a simple digital imaging to qualitatively and quan-
titatively analyze food surface and structure. The ability and reli-
ability of this method was tested by Larrain et al. (2008) who
used digital imaging to estimate colour coordinates of beef. Purlis
and Salvadori (2007) found a high correlation between bread crust
colour and moisture loss while Zhang et al., (2007) have estab-
lished relationships between the outer crust and the inner crumb
property. The aim of this research is to develop a non-destructive
method to measure crust thickness by integrating the digital imag-
ing and colour systems. Ultimately, bread crust thickness could be
obtained through measurement of surface browning of bread using
a chromameter as an indication of its inner properties or product
quality.
2. Materials and methods
2.1. Bread sample preparation
Open breads were produced using the straight dough method
following the standard baking tests using formulation as presented
in Table 1. All the ingredients were mixed in a vertical mixer
(SPN25053, Lian Huat, Malaysia) for a total of 16 min, i.e. 4 min
at low speed and 12 min at high speed. The whole dough was let
to rest for 5 min after mixing. After resting, the dough was weighed
and divided into 380 g dough balls, rounded and let to rest again
for another 5 min before moulded by an automatic moulding ma-
chine (CM750, Lian Huat, Malaysia). The moulded dough were
put into stainless steel baking tins with dimension of
10 cm  19 cm  10.5 cm and stored in the retarded proofer
(LRP36052, Lian Huat, Malaysia) for 90 min at 28 °C and 85% rela-
tive humidity. The doughs were baked in tins without lids at com-
binations of three temperatures, i.e. 185, 195 or 205 °C and three
times, i.e. 25, 30, or 35 min in a convective deck oven (EO3050C5,
Lian Huat, Malaysia) to obtain various crust thickness. All tests
were conducted in triplicates and statistical analysis was per-
Table 1
Open bread formulation (based on 3000 g flour loading).
Ingredients
Flour
Water
Sugar
Salt
Yeast
Shortening
367
formed using Microsoft Excel (XP Edition, Microsoft Corporation,
USA).
2.2. Outer crust and inner crumb colour determination and
measurement
The L a b colour system was used for determining bread crust and
crumb colours because it is the most commonly used colour system
in colorimeter, data acquisition and image processing systems
(Pedreschi et al., 2006). Besides that, it also gives uniformity in col-
our distribution and closeness to human perception (León et al.,
2006). The L value represents lightness component on surface that
the value ranges from 0 to 100 while a and b values are chromatic
components of redness to greenness and blueness to yellowness
that ranges from 120 to 120, respectively (Papadakis et al., 2000).
The colour of the outer crust and inner crumb from all 27 baked
samples were measured using a chromameter (CR410, Konica
Minolta, Japan) with xenon lamp as light source to determine its
colour of crust and crumb regions in terms L a b values. The scan-
ning of the outer crust colour was performed in a consistent manner
with sufficient lighting by placing the chromameter probe onto the
top surface of bread crust while for crumb, the probe was placed
onto the centre part of a central slice of bread (Fig. 1). The average
values of L a b colours describing the outer crust and inner crumb
regions were obtained from all 27 baked open loaf samples. The to-
tal colour difference, DE of the bread slices from the reference is:
h i12
DE ¼ ðLo  LÞ2 þ ðao  aÞ2 þ ðbo  bÞ2 ð1Þ
where Lo = 100, ao = 0 and bo = 0.
2.3. Crust thickness measurement
Crust thickness in this study is defined as the distance between
outer crust and the point of inner crust where its colour satisfies
the colours as crust pre-determined in Section 2.2. This means
the point where the inner crust colour such that the L value
is the minimal and the a and b values are the maximal of the crumb
regions is taken as the thickness. For determining this crust thick-
ness, a sample slice of bread was placed on a transparency grid on a
scanner machine with a cold cathode fluorescent lamp (Scanjet
2400, Hewlett–Packard, USA) to be scanned simultaneously. The
scanned bread slice image in Tiff format was transferred to the
Adobe Photoshop software (Photoshop CS2, Adobe, USA) to allow
the user to determine the crust thickness by counting the number
of boxes which meets the requirement of L a b values known as the
crust region (Fig. 2). The Info application in Photoshop CS2 reads
the original L a b values of the scanned bread slice. Fig. 2 illustrates
that by moving the cursor slowly from the central crumb region to-
wards the crust manually, the thickness is obtained at the point
Location of outer crust colour
measurement using chromameter
Baker’s %
Centre slice
100
63
6 Location of inner crumb colour
1.5 measurement using chromameter
1
5 Fig. 1. Locations for measuring surface browning of outer crust and inner crumb
colours using a chromameter.
368 Y.M. Mohd Jusoh et al. / Journal of Food Engineering 94 (2009) 366–371

Fig. 2. Measurement of crust thickness of a scanned bread slice with the aid of underlying 1 mm grid boxes using the Photoshop software, inset shows the toolbox for L a b
values.

where the L a b values meet the requirement of the crust region 100
pre-determined in Sections 2.2 and 3.1 which is L < 70, a > 0 and
b > 13. This method was adopted and improvised from the previ- L value a value b value
90 Crust Crumb Crust Crumb Crust Crumb
ous work of Yam and Papadakis (2004) and Collar et al. (2005).

80 79.0
3. Results and discussion

70 70.5
3.1. Outer crust and inner crumb colour regions

Table 2 shows the average colour ranges for outer crust and in- 60
ner crumb regions from all the 27 baked open loaf samples. In
identifying the colour region for determining crust thickness, the 50 48.7
inner crust region was identified as the region lying between the
extreme colours of external crust and internal crumb. Fig. 3 illus- 40
trates the distinguishable colour regions of the outer crust and in-
ner crumb in terms of L a b values for determination of colour 30 29.4
ranges for inner crust region as L < 70, a > 0 and b > 13.00. The in- 26.7
ner crust region is defined as the crust thickness where its colour
20
range identification is determined based on its location towards
14.7
the crust region and commences where at the point of minimum 13.0
L value, and maximum values of a and b. The lower L value, higher 10 8.8
5.8 5.9
a and b values at the outer crust compared to the crumb, respec-
0
tively indicate a darker shade, more reddish and yellowish pig- 0
- 0.8
mented region for the crust. The colour pigments formed on the
crust generally agree to the fact that crust experiences carameliza- -10
tion and Maillard reactions which are highly influenced by the
quality and quantity of the precursors, thermal processing param- Fig. 3. L a b values of the outer crust and inner crumb regions for determination of
inner crust regions and colour identification of crust thickness.

eters, pH and quantitative ratio of amino nitrogen to reducing su-

Table 2
gar during baking (Martins et al., 2001).
Colour ranges for outer crust, inner crust and inner crumb obtained from chromam-
eter scanning.
3.2. Effect of baking temperature and time on crust colour and
Bread regions L value a Value b Value
thickness
Outer crust 29.41–48.66 8.78–14.66 5.80–26.65
Inner crumb 70.50–79.00 0.81 to 0.00 5.90–13.00
Fig. 4 shows that the L a b values of outer crust and inner
Inner crust 48.66–70.50 >0.00 >13.00
crumb decreases with increasing baking temperature and time.
 Y.M. Mohd Jusoh et al. / Journal of Food Engineering 94 (2009) 366–371 369

100 16 30

80 12 24

18

b value
L value 60 8

a value
40 4 12

20 0 6

(a ) (b) (c)
0 -4 0
24 28 32 36 24 28 32 36 24 28 32 36
Baking time (minute) Baking time (minute) Baking time (minute)
Fig. 4. Colour trends for outer crust (—) and inner crumb (  ) in terms of (a) L, (b) a and (c) b values at three baking temperatures, 185 °C (), 195 °C (j), and 205 °C (N).

10
2 not really affected by baking conditions and suggests that it has
205C y205C = 0.108x + 5.15, R = 0.952 been insulated by the crust layer. Referring literatures, the L value
2
195C y195C = 0.083x + 4.84, R = 1 has been a reliable colour parameter to describe the crust and
2
185C y185C = 0.092x + 3.74, R = 0.978 crumb regions (Shittu et al., 2007; Purlis and Salvadori, 2007;
9
_
Ibanoğlu, 2002). The reported a and b values for crust are less con-
Crust Thickness (mm)

9.00 _
sistent (Ibanoğlu, 2002) and this could be due to the difference in
8.25 the type of baked products and ingredients used.
8 Besides creating difference in colours and tone of bread surface
browning, baking temperature and time also produced bread with
7.92
7.75 different crust thickness. Using bread crust thickness measurement
established in Section 2.3, the system showed capability to detect
7 7.33 small differences and gives reliable results. The bread crust thick-
ness from all the 27 baked samples ranged from 6.08 to
6.92 7.00
9.00 mm. The reliability of this crust thickness measurement
6.42 method is proven by Fig. 5 which shows that a higher baking tem-
6 perature and time produced bread with higher crust thickness.
6.08
These results are consistent with existing findings on crust thick-
ness as affected by its baking temperature and time (Zanoni
et al., 1992; Jefferson et al., 2006). Zanoni et al. (1992) found that
5
an extension of 5 min’ baking at a fixed baking temperature of
24 28 32 36
203 °C caused bread crust thickness to increase 50% from its origi-
Baking time (minute) nal thickness while Jefferson et al. (2006) discovered that a 14% in-
crease in baking temperature caused a 10% increase in crust
Fig. 5. Crust thickness trends as a result of three baking temperatures, 185 °C (),
thickness. A higher heat and mass transfer, and evaporation pro-
195 °C (j), and 205 °C (N) and three baking times, 25, 30, or 35 min.
cess occur at higher baking temperatures which cause a thicker
crust formation. Fig. 6 shows the image of crust thickening as af-
The similar decreasing trend of L value as affected by increasing fected by baking temperature. The highest rate of crust formation
baking parameters was observed by Shittu et al. (2007). However, took place at 205 °C with value of 0.108 mm/min. The ANOVA re-
the decrease of L a b values are less prominent in the inner crumb. sults showed that the baking temperature (p < 0.0001) causes more
This insignificant decreasing trend shows that the crumb colour is significant crust variation compared to the baking time (p < 0.001).

Fig. 6. Scanned bread crust images depicting thickness of 6, 7, and 9 mm obtained from respective baking temperatures, (a) 185 °C, (b) 195 °C, and (c) 205 °C for a 25 min
baking period.
370 Y.M. Mohd Jusoh et al. / Journal of Food Engineering 94 (2009) 366–371

10 10 10
(a ) (b) (c)

Crust Thickness (mm)

Crust Thickness (mm)

Crust Thickness (mm)

8 8 8

6 6 6

y = -0.147x + 13.115 y = -0.4229x + 12.723 y = -0.1371x + 9.7692

R 2 = 0.9549 R 2 = 0.8515 R 2 = 0.9416
4 4 4
20 30 40 50 60 6 10 14 18 3 12 21 30
L value a value b value
Fig. 7. Correlations between outer crust colour reported as (a) L, (b) a, and (c) b values with crust thickness for all 27 samples. ( = 185 °C, 25 min, j = 185 °C, 30 min,
N = 185 °C, 35 min,  = 195 °C, 25 min, + = 195 °C, 30 min,  = 195 °C, 35 min, } = 205 °C, 25 min, h = 205 °C, 30 min, and D = 205 °C, 35 min).

3.3. Correlations between surface browning of crust and its thickness or bakery products for the purpose of prediction of its crust prop-
erties. The crust properties are known to be useful in terms of
Fig. 7 shows clear negative correlations between each of the L a moisture loss and control hence may contribute a great deal in
b values with crust thickness. The strength of the relationships are lowering its staling rate.
strong with high coefficients of correlations, R2, of 0.9549, 0.8515
and 0.9416 for L a b values, respectively. These high R2 values infer 4. Conclusions
that the colour components obtained from the outer crust could be
used to determine its crust thickness. In terms of L value, a lower L The thickness of bread crust was successfully determined by
value obtained from a darker crust surface infers a thicker crust. As distinguishing the crust and crumb colour regions via the applica-
the crust thickens, the redness implied by the a value and the yel- tion of digital imaging and the L a b colour system. The system is
lowness implied by the b value, both decrease. However, as the able to pick up small differences from all 27 samples which thick-
crust thickens, the L value still dominates as the a and b values loss ness ranged from 6.08 to 9.00 mm. The strong positive correlation
their visibility on the surface, and replaced by the L value solely. between crust thickness and the total crust colour difference (DE)
With the total colour difference, DE, calculated using Eq. (1), a po- was found and this allows prediction of crust thickness from the
sitive correlation with the thickness was found (Fig. 8). With a high brown surface colour of baked bread for bread quality evaluation.
R2, of 0.9467, this relationship could be used to predict bread crust The baking trials showed that the crust colour and thickness in-
thickness. Similar linear relationships between the outer crust col- crease with baking temperature and time. The temperature
our and its thickness could be established for other types of loaves (p < 0.0001) has more significant impact on both crust properties,
the colour and thickness as compared with time (p > 0.001). This
crust thickness measurement method can be established for var-
10 ious bakery products and provides an alternative of cheap and
fast technique for more studies towards understanding crust
properties as crust has significant influences on bread quality
and shelf-life.
9
Acknowledgements
Crust Thickness (mm)

The authors would like to acknowledge the financial support

8
from Ministry of Higher Education of Malaysia for financial support
of this project (Vot. No. 5523214) and thank the Research, Develop-
ment and Commercialization Centre (RDCC) of Interflour Sdn. Bhd.
7 for providing baking facilities.
y = 0.1724x
2
R = 0.9467 References

6 Abdullah, M.Z., Abdul Aziz, S., Dos Mohamed, A.M., 2000. Quality inspection of
bakery products using a colour based machine vision system. Journal of Food
Quality 23, 39–50.
Blasco, J., Aleixos, N., Gomez, J., Molto, E., 2007. Citrus sorting by identification of
most common defects using multispectral computer vision. Journal of Food
5 Engineering 83, 384–393.
Brosnan, T., Sun, D-W., 2004. Improving quality inspection of food products by
30 35 40 45 50 55 computer vision – a review. Journal of Food Engineering 61, 3–16.
Chao, K., Chen, R., Hrushka, W.R., Gwozdz, F.B., 2002. On-line Inspection of poultry
ΔE carcasses by a dual-camera system. Journal of Food Engineering 51, 185–192.
Collar, C., Bollain, C., Angioloni, A., 2005. Significance of microbial transglutaminase
Fig. 8. Prediction of crust thickness from its outer crust colour as total colour on the sensory, mechanical and crumb grain pattern of enzyme supplemented
difference (DE). ( = 185 °C, 25 min, j = 185 °C, 30 min, N = 185 °C, 35 min, fresh pan breads. Journal of Food Engineering 70, 479–488.
 = 195 °C, 25 min, + = 195 °C, 30 min,  = 195 °C, 35 min, } = 205 °C, 25 min, Gallagher, E., Gromley, T.R., Arendt, E.K., 2003. Crust and crumb characteristics of
h = 205 °C, 30 min, and D = 205 °C, 35 min). gluten free breads. Journal of Food Engineering 56, 153–161.
 Y.M. Mohd Jusoh et al. / Journal of Food Engineering 94 (2009) 366–371
_
Ibanoğlu, E., 2002. Kinetics study on colour changes in wheat germ due to heat.
Journal of Food Engineering 51, 209–213.
Jefferson, D.R., Lacey, A.A., Sadd, P.A., 2006. Understanding crust formation during
baking. Journal of Food Engineering 75, 515–521.
León, K., Mery, D., Pedreschi, F., León, J., 2006. Colour measurement in L*a*b* units
from RGB digital images. Food Research International 39, 1084–1091.
Lostie, M., Peczalski, R., Andrieu, J., 2004. Lumped model for sponge cake baking
during the ‘‘crust and crumb” period. Journal of Food Engineering 65, 281–286.
Martins, S.I.F.S., Jongen, W.M.F., Van Boekel, M.A.J.S., 2001. A review of Maillard
Reaction in food and implications to kinetic modeling. Food Science and
Technology 11, 364–373.
Papadakis, S.E., Abdul-Malek, S., Kamdem, R.E., Yam, K.L., 2000. A versatile and
inexpensive technique for measuring colour of foods. Food Technology 54, 48–
51.
Pearson, T., 1996. Machine vision system for automated detection of stained
pistachio nuts. Lebensmittel-Wissenschaft and Technolgie 29, 203–209.
Pedreschi, F., León, J., Mery, D., Moyano, P., 2006. Development of a computer vision
system to measure the colour of potato chips. Food Research International 39,
1092–1098.
Primo-Martin, C., Van de Pijpekamp, A., Van Vliet, T., De Jongh, H.H.J., Plitjter, J.J.,
Hamer, R.J., 2006. The role of gluten network in crispness of bread crust. Journal
of Cereal Science 43, 342–352.
371
Purlis, E., Salvadori, V.O., 2007. Bread browning kinetics during baking. Journal of
Food Engineering 80, 1107–1115.
Shittu, T.A., Raji, A.O., Sanni, L.O., 2007. Bread from composite cassava-wheat flour:
I. Effect of baking time and temperature on some physical properties of bread
loaf. Food Research International 40, 280–290.
Wahlby, U., Skjoldebrand, C., 2002. Reheating characteristics on crust formed on
buns and crust formation. Journal of Food Engineering 53, 177–184.
Wiggins, C., 1999. Proving, baking, cooling. In: Cauvain, S.P., Young, L.S. (Eds.),
Technology of Breadmaking. Aspen Publishers, Maryland, pp. 120–148.
Yam, K.L., Papadakis, S.E., 2004. A simple digital imaging method for measuring and
analysing color of food surfaces. Journal of Food Engineering 61, 137–142.
Zanoni, B., Peri, C., Pierucci, S., 1992. A study of the bread baking process I: a
phenomenological model. Journal of Food Engineering 19, 389–398.
Zenthenbaur, G., Grosh, W., 1998. Crust aroma of baguettes. I. Key odorants of
baguettes prepared in two different ways. Journal of Cereal Science 28, 81–92.
Zhang, L., Lucas, T., Doursat, C., Flick, D., Wagner, M., 2007. Effects of crust
constraints on bread expansion and CO2 release. Journal of Food Engineering 80,
1302–1311.
Zou, X., Zhao, J., Li, Y., 2007. Apple colour grading based on organization feature
parameters. Pattern Recognition Letters 28, 2046–2053.