Content uploaded by Moslem Namjoo
Author content
All content in this area was uploaded by Moslem Namjoo on May 05, 2016
Content may be subject to copyright.
197
Agriculturae Conspectus Scienti cus . Vol. () No. (-)
ORIGINAL SCIENTIFIC PAPER
Summary
Non-destructive impulse response technique was tested on two melon varieties
(‘Zard-Eyvanekey’ and ‘Sousky-Sabz’) in ve di erent stages of ripening. Resonance
frequency and sound pressure level (SPL) from the impulse response were compared
with mass, elastic modulus, soluble solids contents (TSS) and sensory evaluation.
Among acoustic and destructive parameters, resonance frequency is an indicator
to distinguish of di erent maturity stage of the melon in both varieties. e sound
pressure level (SPL) was not a reliable parameter to evaluate melon ripeness. It was
established that the impulse response technique is useful to decide fruit maturity
stage and appropriate harvest time.
Key words
acoustic, impulse, response, non-destructive, melon, ripeness
Acoustic Testing for Melon Fruit
Ripeness Evaluation during Di erent
Stages of Ripening
Farhad KHOSHNAM 1
Moslem NAMJOO 1( )
Hossein GOLBAKHSHI 2
1 Department of Mechanical Engineering of Biosystems, Faculty of Agricultural,
University of Jirof t, 78671-61167, Jiroft, Iran
e-mail: m.namjoo@ujiroft.ac.ir
2 Department of Mechanical Engineering, University of Jiroft, 78671-61167, Jiroft, Iran
Received: October 8, 2015 | Accepted: March 15, 2016
ACKNOWL EDGEMENTS
The authors wish to express their sincere thanks to University of Tehran for financial assis-
tance and staff members of the Faculty of Agricultural Engineering & Technology, particularly
the volunteers who participated in the sensory evaluation. They also acknowledge the Iranian
Agricultural Engineering Research Institute for providing the laboratory facilities.
Agric. conspec. sci. Vol. () No.
198 Farhad KHOSHNAM, Moslem NAMJOO, Hossein GOLBAKHSHI
Introduction
Eating quality of melon depends mainly on harvest of mature
melons at desired stage or ripeness. Generally, optimum eating
quality requires adequate sugar and avor development, and the
centre meat with a melting texture that progress to a crisp texture
toward the rind. Immature or underripe melons have less sugar
and avor development, and have a rmer texture than those
at optimum ripeness. Flavor and texture degrade dramatically
as melon progress from ripe to overripe. It is di cult to judge
ripeness by external characteristics such as size, external color,
stem condition or feel. Ripe melons should be rm, symmetri-
cal, and fresh looking. Determining optimum melon maturity
at harvest time, however, is a critical but di cult task, even for
experienced growers.
Melons (Cucumis melo L.) are commercially important fruits,
but their ripening has been relatively poorly studied compared
to other fruits such as tomatoes, avocadoes or apples. A large
number of diverse melon cultivars are available that exhibit
variation in ripening characteristics. Early and late harvesting
varieties are known for many fruits, but in melons a selection
can also be made according to fruit color, shape or sweetness.
ere is also variation in the respiratory climacteric, which is
probably a variety-dependent characteristic of melon (Nukaya
et al., 1986; Had eld et al., 1995).
Many people have claimed that the maturity and other qual-
ities of certain fruits, such as apples, melons, and pineapples,
can be determined by listening to the sound produced by strik-
ing them. Several researchers have tried to verify such claims by
studying the acoustic responses of fruits. Hernández Gómez et
al. (2006) evaluated the capacity of acoustic signal response to
monitoring the mandarin fruit rmness change during storage.
e results indicate that it might be possible to identify the ripe-
ness state of an individual mandarin by using the present method,
and that the nondestructive acoustic test could replace conven-
tional compression test in order to determine mandarin fruit
rmness and expected shelf- life (Gómez et al. 2006). Taniwaki
et al. (2009) investigated time-course changes in the elasticity
index (EI) and texture index (TI) of two persimmon cultivars
during the postharvest period. EI was determined using the for-
mula EI = f22 m2/3, where f2 is the second resonance frequency
of a sample, and m is the mass of the sample. ey found that
changes in the EI of both cultivars showed quasi-exponential
decays. An improved texture measurement device was used for
measuring the TI of the cultivars. ey determined the optimum
eating ripeness of persimmons along with the sensory test in
terms of their EI (Taniwaki et al., 2009). Zheng et al. (2014) ad-
dresses the problem of distinguishing between ripe and unripe
watermelons using mobile devices. ey found that through an-
alyzing ripeness-related features extracted by thumping water-
melons, collecting acoustic signals by microphones on mobile
devices, this method can automatically identify the ripeness of
watermelons. Experimental results show that this method is
currently able to correctly classify ripe and unripe watermelons
with an overall accuracy exceeding 89% (Zeng et al., 2014). To
evaluate the rmness of fruits Macrelli et al. (2013) compared
three novel sti ness indexes based on acoustic methods and in-
volving Young’s modulus and sound propagation velocity. e
e ectiveness of the considered indexes is tested by means of an
experimental setup built with two piezoelectric transducers
contacting several samples of kiwifruits during their ripening
process. ey found that sti ness indexes based on propagation
delays are more rapid and reliable than those based on fruit res-
onance in assessing the ripeness degree (Macrelli et al., 2013).
Hongwiangjan et al. (2015) investigated the maturity as-
sessment of pomelo using acoustic properties obtained from
an impact of fruit, optical properties of the peel and variables
related to oil glands from peel images. ey found that the clas-
si cation model based on the nondestructive variables showed
that fruits could be separated into immature, early-mature and
late-mature groups with an accuracy of 96.7% (Hongwiangjan
et al., 2015). Mao et al. (2016) developed an acoustic device a er
investigating the in uence of hitting ball and fruit tray on spec-
trum. ey proposed three rmness indices to correlate with
rmness of watermelon. ey found that signi cant correlation
was between rmness and these indices using linear regressive
model and nonlinear model of arti cial neural network (ANN)
(Mao et al., 2016). An experimental system for nondestructive
rmness evaluation, based on the exible piezoelectric sensors,
a microphone or and accelerometer was developed and tested
on several fruit and other products, such as: apple (Yamamoto
et al., 1980; Woensel et al., 1988; Armstrong et al., 1989; Chen et
al., 1992; Chen 1993); tomato (Duprat et al., 1997; Schotte et al.,
1999; Baltazar et al., 2008), avocado (Shmulevich et al., 2003);
muskmelon (Sugiyama, 1994); watermelon (Stone et al., 1996;
Diezma-Iglesias et al., 2004); pear (Jancsók et al., 2001; Wang et
al., 2004); peach (Goliáš et al., 2003; Gómez et al. 2005); potato
(Baritelle, 1997), and many others. Most of these researches used
an experimental system for non-destructive rmness evaluation
based on microphone techniques, which have successfully been
used for several fruits and other products. De Belie et al. (2000)
point out the advantage of the acoustic technique that is an
overall measure and a very reproducible and sensitive method.
With the acoustic impulse response technique the resonance
frequency is obtained by performing a Fourier transformation
on the recorded sound of an intact fruit. e sound is produced
by impacting the fruit, which then vibrates and causes pressure
waves in the air. For contact sensing typically sensors, such as
accelerometers and piezoelectric lms are used. For non-contact
sensing a microphone or laser vibrometer is needed. e prima-
ry objective of the present work was to develop a nondestruc-
tive method based on impulse response technology for quality
evaluation of two di erent melon varieties ‘Zard-Eyvanekey’
and ‘Sousky-Sabz’.
Materials and methods
Sampling of melon
This research was conducted on ‘Zard-Eyvanekey’ and
‘Sousky-Sabz’ varieties obtained from a plantation in Garmsar
township (35º13ʹ20˝ N, 52º20ʹ26 E). ey were carefully picked
by hand during the summer and autumn in the early morning
from the area of Davarabad, Garmsar, Iran. Fruits were selected
according to color, size and lack of blemishes in order to obtain
homogeneous samples. Before each test series, the melons were
transferred to department laboratory at 18 to 22oC temperature
Agric. conspec. sci. Vol. () No.
199
Acoustic Testing for Melon Fruit Ripeness Evaluation during Different Stages of Ripening
and kept for 24 hours. ey were selected at ve di erent stages
of ripening (Table 1). All physical parameters were studied for
65 fruits from each variety. e experiments were conducted at
Biophysical and Biological laboratory of University of Tehran,
Karaj, Iran.
Determination of physical and mechanical
properties and TSS
Mass (M) of the melons was measured with a precision bal-
ance with a sensitivity of 1 g and actual volume was determined
by the water displacement technique. e true density (ρ) is the
ratio of mass of melon to its actual volume. ree cylindrical
cores were cut near the equator from one half of each melon,
where the esh tends to have the largest thickness, using a cy-
lindrical borer through the esh along the radial direction. Each
14 mm diameter cylindrical core was then cut into 14 mm long
samples. e modulus of elasticity (E) of melon was evaluated
using 20 samples by Universal Testing Machine (Santam, SMT-
5). e machine was equipped with a load cell of 150 N at a com-
pressive rate of 25.4 mm/min
For assessment of the total soluble solids (TSS or the ‘Brix’
refractomet ric measurement) a core (22 mm d iameter or greater)
was sampled from a randomly selected equatorial position on
the fruit, and the rind and green inedible tissue (5 mm thick-
ness) and placenta and seed were removed. All remaining edible
mesocarp tissue was juiced.
Panel selection and training
e eating experience of a piece of fruit re ects physical
characters such as texture, and chemical characters such as
sweetness, acidity, and volatiles composition. Sensory analysis
was performed for each harvest by 30-35 untrained assessors
that are familiar with the product in question. e testers for
the sensory evaluation were graduate students of the faculty of
agricultural engineering and technology, university of Tehran,
Iran. e esh of fruit was cut into about 2x2 cm pieces and then
samples were coded with 3-digit random numbers.
Presentation of samples was randomized among the panelists
and sessions, and they were presented one at a time under mod-
erate incandescent lighting. Assessors were instructed to clean
their palates with a sip of room temperature water and it was
a small time lag before each sample. Taste analysis acceptabil-
ity (consumer-oriented testing method) scored on a scale of 1-5
where: 1, very bad; 2, poor; 3, fair; 4, good and 5, excellent were
used to assess ve attributes: avor, bitterness, sourness, rm
texture and juiciness. Finally the assessors determined the gen-
eral acceptability (GA) and later the average of it was calculated.
Experimental equipment
e laboratory recording system used to acquire the acoustic
impulse information is comprised of a melon-bed, an impactor
(pendulum), sensing device, a lap-top computer and so ware to
control the experimental setup and to analyze its results. During
the test, the fruit (melon) was placed on a so foam support in
order to create free support conditions and not to disturb the vi-
bration pattern. e frequency of all individual fruit was meas-
ured on the three positions along the equator approximately 120º
between them. e impact needs a su cient stroke, mass, veloc-
ity and right angle. e combination of these causes problems to
miniaturise the little impactor. e impactor consists of a steel
ball of diameter of 26 mm and a 256 mm long copper rod. e
weight of the impactor was 72 g; the impact angle was of 70º.
Chen (1993) postulates as i mpact requi rements that the impact
force is high enough to excite the expected frequency range. He
also mentions that the force should be low enough not to damage
the fruit. He suggests that the mass, the initial contact velocity,
the curvature in the contact area, and the elastic modulus of the
impact material should be controlled (Chen, 1993).
e acoustic signal was sensed by a sound level meter (SLM)
type 2270 B&K that have a prepolarised free- eld 1/2” micro-
phone type 4189 B&K with at frequency response in human
threshold of hearing range (20 to 20000 Hz) and sensitivity
50mVPa-1. e sound level meter is a measuring instrument
Stage Operation Date Description
1 First series of test Mid-Au
g
ust Immature
2 Second series of test Late-Au
g
ust Earl
y
ri
p
enin
g
3 Third series of test Mid-Se
p
tember Moderatel
y
ri
p
e
4 Forth series of test Late-Se
p
tember Ri
p
e
5 Fifth series of test Mid-October Overripe
Table 1. Date of harvesting and tests series
Figure 2. Setup of experiment for acoustic impact test
Figure 1. Wave form Di splay
Agric. conspec. sci. Vol. () No.
200 Farhad KHOSHNAM, Moslem NAMJOO, Hossein GOLBAKHSHI
used to measure sound pressure level. It was at a distance of 2-5
mm from the fruit surface when detecting the impulse acous-
tic response. User friendly Windows-based so ware, Cool Edit
Pro 2.0 was used for the control of the process and the regis-
ter of data, providing an easy output to be used with Microso
Excel. e so ware shows the acoustic signal ‘time versus sam-
pling rate’ for each test on the screen, and saves it in an ASCII
le (Fig.1). Cool Edit Pro 2.0 is combination of digital recording
and editing features. To start recording, we simply used /File/
New to open a new le, then selected the sample rate, bit resolu-
tion, and number of channels (stereo or mono) that we wanted to
use, pressed OK, and clicked on the Record button in the lower
le area of the main window to begin. When we were done re-
cording, we clicked on Stop and then saved our recording. We
selected the sample rate 192000, resolution 16-bit, and channel
mono. e Amplitude Ruler displayed the relative amplitude
of a waveform over time. e ruler’s display format can be set
to either Samples (exact sample value of the data), a percentage
(from -100% to 100%, where 100% is 0 dB) or as a normalized
value (-1 to 1) in Waveform View. In Spectral View, the vertical
ruler is always in frequency (Hz) format. e display format can
easily be changed, using the le and right and dragging. Fig. 2
shows setup of experiment for acoustic impact test. e acoustic
response of each melon was measured by hitting the fruit with
an impactor and detecting the output sound by a sound level
meter. e sound pressure level (SPL) was measured by sound
level meter. A fast Fourier transform (FFT) of the signal was
performed to determine the natural frequencies of the melons
(Fig. 3). Analysis of variance (ANOVA) was applied to the data.
Means corresponding to the di erent stages of evolution were
compared using Duncan’s multiple range tests (p 0.05).
Results and discussion
During ripening, a fruit passes through a series of changes in
color, texture and avor indicating that compositional changes
are taking place.
Table 2 shows changes in mass (M), true density (ρ), modu-
lus of elasticity (E), total soluble solids (TSS), frequency (f) and
sound pressure level (SPL) during the ripening of melon fruits.
As it can be seen in table, melon mass, total soluble solid and
sound pressure level increased during ripening period, while
true density, modulus of elasticity and frequency decreased from
immature fruits to over ripe fruits.
e Fig. 4 shows that the mass increased progressively in
both varieties over the period of development and ripening as
expected. However, the increasing rates were di erent during
the harvesting stages in both varieties: faster rate in the initial
stages and lower in the nal stages. e mass of samples increased
during the growing season rapidly; the mass of ‘Zard-Eyvanekey’
increased from 591.73 to 3913.15 grams (about 6.6 times) and
the mass of ‘Sousky-Sabz’ from 569.52 to 3904.56 grams (about
Figure 3. Typical acoustic signal of melon: (a) time domain, (b) frequency domain
Variet
y
Maturity Stage M (g) ρ (kg/m3) E (MPa) TSS (°Brix) f (Hz) SPL (dB)
Zard-Eyvaneke
y
First 591.8e 901.84
a
0.466
a
4.93
d
132.25 49.21c
Second 1490.2d 886.13
a
0.309b
6.53
d
128.91 49.59c
Third 2978.8c 863.16b 0.211c
9.43c 119.14 54.19b
Fourth 3663.9b 855.45bc 0.199bc 11.07b 113.34 56.93
a
Fifth 3913.1
a
842.33c 0.149
d
13.03
a
111.33 57.74
a
Sousk
y
-Sabz First 569.5e 908.11
a
0.417
a
4.83
d
128.91 49.66
d
Second 1581.5d 898.85
a
0.294b 6.43c 123.05 50.15
d
Third 3016.6c 878.34
a
0.226bc 9.30bc 119.14 53.23c
Fourth 3424.7b 865.41b 0.176c 10.80b 114.26 56.25b
Fifth 3904.6
a
860.03bc 0.154c
12.43
a
111.33 58.22
a
Means in the same column followed by different letters are significantly different according to Duncan’s test (p<0.05).
Table 2. Changes of physical and mechanical properties and acoustic parameters of melon during harvesting (average values of
65 melons from each variety)
Agric. conspec. sci. Vol. () No.
201
Acoustic Testing for Melon Fruit Ripeness Evaluation during Different Stages of Ripening
6.8 times). e average of mass in the fourth harvest (ripened),
‘Zard-Eyvanekey’ and ‘Sousky-Sabz’ was estimated to be 3663.92
and 3424.70 grams, respectively. e ‘Zard-Eyvanekey’ mass was
higher than the ‘Sousky-Sabz’ mass in full ripening, alt hough t he
rate of increase of the mass was higher in ‘Sousky-Sabz’ variety.
e true density of two varieties had shown a decreasing trend.
True density or ‘Zard-Eyvanekey’, was 901.84kgm–3 in rst stage
and 842.33kgm–3 in h stage (reduction 6.6%), whereas these
values for ‘Sousky-Sabz’ were 908.11kgm–3 and 860.03kgm–3
(reduction 5.3%). e true density of (865.41kgm–3) was a little
higher than of ‘Zard-Eyvanekey’ (855.45kgm–3) in full ripen-
ing (Fig. 5).
The elastic modulus values decreased in both varieties
throughout ripening (Fig. 6). Its values fall from 0.466 to 0.149
MPa and from 0.417 to 0.154 in ‘Zard-Eyvanekey’ and ‘Sousky-
Sabz’, respectively. Initial elastic modulus values of ‘Zard-
Eyvanekey’ were higher than those of ‘Sousky-Sabz’. is may
be related to di erences in physiological stage of ripeness, but
also to the higher amount of cell wa ll polysaccharides in ‘Zard-
Eyvanekey’ than in ‘Sousky-Sabz’ pulp. Both varieties have fruits
with high elastic modulus at the rst harvest date that shows
compact and high density fruit. e rate of elastic modulus loss
was also higher in ‘Zard-Eyvanekey’ than in ‘Sousky-Sabz’. e
elastic modulus of ‘Zard-Eyvanekey’ variety (0.199 MPa) esti-
mated to be higher than in ‘Sousky-Sabz’ variety (0.176 MPa)
in full ripening.
Melons are among the fruits with the highest sugar content.
In melons, sweetness is the dominant consumer acceptance
factor, and total soluble solids (TSS), as a measure of sweetness,
is the most useful chemical measurement for assessing melon
acceptability.
e total soluble solids (TSS) followed upward trends through-
out ripening in both varieties as expected, though values were
higher in ‘Zard-Eyvanekey’ variety than in ‘Sousky-Sabz’ vari-
ety. TSS of ‘Zard-Eyvanekey’ variety was initially at 4.93°Brix
and reached a value of 13.03°Brix, these values were 4.83°Brix
and 12.43°Brix in ‘Sousky-Sabz’ variety (Fig. 7).
e resonance frequency values of ‘Zard-Eyvanekey’ and
‘Sousky-Sabz’ varieties decreased during whole harvesting
period from 132.25 to 111.33 Hz and from 128.91 to 111.33 Hz
respectively. e average of resonance frequency in the fourth
harvest (ripened) for ‘Zard-Eyvanekey’ and ‘Sousky-Sabz’ was
estimated to be 113.34 and 114.26 Hz, respectively. Consequently,
the resonance frequency of ‘Zard-Eyvanekey’ variety was a little
Figure 4. Mass changes during the fruit ripening of melon
Figure 6. Modulus of elasticity changes during the fruit
ripening of melon
Figure 5. True density changes during the fruit ripening of
melon
Figure 7. TSS changes during the fruit ripening of melon
Agric. conspec. sci. Vol. () No.
202 Farhad KHOSHNAM, Moslem NAMJOO, Hossein GOLBAKHSHI
lower than the resonance frequency of ‘Sousky-Sabz’ variety in
full ripening (Fig. 8).
As mentioned, the resonance frequency decreased in both
varieties during ripening. e other researchers kept samples in
cold storage (optimum temperature and relative humidity) and
found that resonance frequency decreases with fruit ripening.
e samples have been provided with the uniform shape, size
and color at the same time. e researches were done on speci-
ed samples under controlled conditions and generalization of
the conclusions is di cult. In this work, the samples were pre-
pared at several times and had non-monotonous shape, size and
color. Furthermore, the control of many environmental and eld
factors is di cult and even impossible.
e response to vibrations in the fruits and vegetables de-
pends on their elasticity modulus, mass and shape. is result
means a diminution of 15.8% and 13.6% in the resonance fre-
quency values respective to their initial values. is diminution
is most likely caused by a di erent ripening rate of each vari-
ety, since it is assumed that the resonant frequency in fruits is
changed mainly by ripeness. High frequencies were found from
unripe fruit for both variet ies. In general, it means that t he more
the fruit is ripened, the lower is its frequency. Chen (1993) found
the measured frequencies to decrease with storage time and to
be correlated with fruit rmness and sensory measurements.
As it can be seen in Fig. 9 the sound pressure level (SPL) varied
and increased from 49.21 to 57.74 dB for ‘Zard-Eyvanekey’ vari-
ety and from 49.66 to 58.22 dB for ‘Sousky-Sabz’ variety during
ripening of fruit. e average of sound pressure level in the fourth
harvest (ripened) for ‘Zard-Eyvanekey’ and ‘Sousky-Sabz’ was
estimated 56.93 and 56.25 dB, respectively. Consequently, the
sound pressure level of ‘Zard-Eyvanekey’ variety was a little
higher than the sound pressure level of ‘Sousky-Sabz’ variety.
e variation in the distance between sound level meter (SLM)
and melon surface a ect the sound pressure level (SPL). is
method is not very precise because of distance change (2–5 mm)
of sound level meter from the melon surface. erefore we found
that it is not possible to measure ripeness accurately simply by
determining the sound pressure level. Probably due to the in-
crease of the mass of both varieties during the growing season,
the sound pressure level of the process increased.
We concluded that the melon ripening is characterized by a
progressive increase in the mass and TSS and a decrease in the
resonance frequency and elastic modulus. e higher mass and
TSS correspond to the sweeter melon and the higher resonance
frequency and elastic modulus correspond to the rmer melon
prior harvesting.
e sensory score for both varieties increased with the degree
of ripeness to reach a maximum general acceptability at fourth
stage, and therea er decreased with further ripening. Consumers
preferred the general acceptability of melon pieces from ‘Zard-
Eyvanekey’ variety compared to ‘Sousky-Sabz’ variety (Fig. 10).
Due to the similar the taste and rmness of melon to cu-
cumber at the rst stage of harvesting (Mid-August, immature
melon) to cucumber, the Garmsar farmers used the melons as a
salad, and this would a ect the assessors scoring. e assessors
scoring were high in this stage. To solve this problem, we asked
the assessors to compare the samples to ripe melon. e melons
are not consumed at second stage of harvesting (Late-August,
early ripening) because of rapid increase in the amount of TSS
in respect to the rst stage ( e TSS of ‘Zard-Eyvanekey’ vari-
ety were 4.93° Brix and 6.53° Brix and in ‘Sousky-Sabz’ variety,
these values were 4.83° Brix and 6.43° Brix) and water shortage.
In the nal stage of harvesting of both varieties assessors deter-
mined that although TSS increased, the rmness of tissue and
consequently general acceptability decreased.
One of the main objectives in scienti c research is the re-
lationship between these phenomena. Correlation analysis was
carried out to determine the strength of the relationship between
each acoustic parameter with destructively measured ripeness
indicators for both varieties (Table 3).
Figure 8. Frequency changes during the fruit ripening of
melon
Figure 9. Sound Pressure Level (SPL) changes during the
fruit ripening of melon
Figure 10. General acceptability changes in melon
(average va lu es)
Agric. conspec. sci. Vol. () No.
203
Acoustic Testing for Melon Fruit Ripeness Evaluation during Different Stages of Ripening
A very high correlation between resonance frequency and
sound pressure level parameters and mass, total soluble solids
(TSS) were obtained for both varieties.
e correlations between mass and TSS for both varieties are
excellent, with a determination coe cient of 0.965 and 0.977 for
‘Sousky-Sabz’ and ‘Zard-Eyvanekey’, respectively, which means
that more than 96% of the variation in mass is explained by TSS.
Levels of correlation coe cient of 0.522 for ‘Zard-Eyvanekey’
variety and 0.632 for ‘Sousky-Sabz’ variety were found between
the resonance frequency and general acceptability. e correla-
tion coe cients were low because the resonance frequency had
increase trend, while the general acceptability had increase trend
until forth stage of ripening and then decreased.
Also, the elastic modulus and resonance frequency showed
negative correlations with TSS, sound pressure level and general
acceptability parameters in bot h varieties. e good correlation
(R2=0.865) was obtained for ‘Zard-Eyvanekey’ variety and ex-
cellent correlation (R2=0.963) for ‘Sousky-Sabz’ variety between
elastic modulus and resonance frequency. ese ndings can be
explained by the high sensitivity of the acoustic method (reso-
nance frequency) to elastic modulus of fruit.
Also, a strong correlation between resonance frequency
and sound pressure level (R2=0.991 for ‘Zard-Eyvanekey’ and
R2=0.933 for ‘Sousky-Sabz’) was observed. Statistical analysis
of the data indicated high correlation between acoustic param-
eters and others ripeness indicators.
Our results showed that the value of resonance frequency
is a useful indicator of maturity stage, as it decreases linearly
throughout ripeness for both varieties.
e acoustic impulse response method is then an indirect
way for non-destructive sensing of the melon ripeness. On the
other hand, the data suggests that resonance frequency and elas-
tic modulus can both be used to distinguish among fruits with
diferent maturity and ripeness levels. During the all experiments
no sign of bruising was observed on the melon’s skin, con rm-
ing the nondestructiveness of this technique.
Conclusions
e major changes that occur during harvesting period of
both varieties are reduction in true density and elastic modulus.
e general acceptability of both melons determined by asses-
sors and correlation coe cients between them are low because
the resonance frequency had increase trend, while the general
acceptability had increase trend until forth stage of ripening
and then decreased. A very high correlations between resonance
frequency and sound pressure level parameters and mass and
total soluble solids (TSS) were obtained for both varieties. It is
not possible to measure ripeness accu rately simply by determin-
ing the sound pressure level (SPL). Experiments con rmed that
non-destructive acoustic impulse response can be successfully
used to distinguish di erent stage of ripeness.
References
Armstrong P., Zapp H. and Brown G. (1989). Impulsive excitation of
acoustic vibrations in apples for rmness determination. Trans
ASAE. 3(4):1353-1359.
Baltazar A., Aranda J. I. and González-Aguilar G. (2008). Bayesian
classi cation of ripening stages of tomato fruit using acous-
tic impact and colorimeter sensor data. Comput Electron Agr.
60(2): 113-121.
Baritelle A. (1997). Factors a ecting potato impact sensitiv-
ity. Proceedings, National Potato Council and Snack Food
Association, CHIP-IN Chipping Potato Seminar, Hilton Head,
S. Carolina, March.
Chen H. (1993 ). Analysis of the acoustic impulse resonance
of apples for nondestructive estimation of fruit quality.
Dissertationes de Agricultura (Belgium).
Chen P., Sun Z. and Huarng L. (1992). Factors a ecting acoustic
responses of apples. Trans ASAE. 35: 1915-1920.
De Belie N., et al. (2000). PH—Postharvest Technology: Firmness
Changes of Pear Fruit before and a er Harvest with the
Acoustic Impulse Response Technique. J Agr Eng Res. 77(2):
183-191.
Diezma-Iglesia s B., Ruiz-Altisent M. and Barreiro P. (2004).
Detection of internal quality in seedless watermelon by acoustic
impulse response. Biosyst Eng. 88(2): 221-230.
Duprat F., et al. (1997). e acoustic impulse response method for
measuring the overall rmness of fruit. J Agr Eng Res. 66(4):
251-259.
Goliáš J., Bej ček L., Graetz P. and Klusáček S. (2003). Mechanical
resonance method for evaluation of peach fruit rmness.
Horticultural Science 1(1): 1-6.
Gómez A. H., P ereira A. G., Jun W. and Yong H. (2005). Acoustic
testing for peach fruit ripeness evaluation during peach storage
stage. Rev Cie Téc Agr. 14(2): 28-34.
Gómez A. H., P ereira A. G. and Wang, J. (2006). Acoustic impulse
response potential to measure mandarin fruit ripeness during
storage. Rev Cie Téc Agr. 15(4): 24-30.
Had eld K. A. , Rose J. K. and Bennett A. B. (1995). e respira-
tory climacteric is present in Charentais (Cucumis melo cv.
Reticulatus F1 Alpha) melons ripened on or o the plant. J Exp
Bot. 46(12): 1923-1925.
Hongwiangjan J ., Terdwongworakul A. and Krisanapook K. (2015).
Evaluation of pomelo maturity based on acoustic response and
peel properties. Int J Food Sci Tech. 50(3): 782-789.
Jancsók P. T., Clijmans L., Nicolai B. M. and De Baerdemaeker
J. (2001). Investigation of the e ect of shape on the acous-
tic response of ‘conference’pears by nite element modelling.
Postharvest Biol Technol. 23(1): 1-12.
Variet
y
M E TSS f SPL GA
Zard-
Eyvanekey
M 1 0.929 0.965 0.986 0.957 0.522
E 1 0.883 0.865 0.795 0.399
TSS 1 0.978 0.961 0.400
f 1 0.991 0.517
SPL 1 0.508
GA 1
Sousk
y
-
Sabz
M 1 0.969 0.977 0.961 0.890 0.632
E 1 0.931 0.963 0.822 0.646
TSS 1 0.979 0.963 0.642
f 1 0.933 0.675
SPL 1 0.678
GA 1
M (Mass, kg), E (Elastic Modulus, MPa), TSS (Total Soluble Solids, °Brix),
f (Resonance Frequency, Hz), SPL (Sound Pressure Level, dB) and GA
(General Acceptability, without unit)
Table 3. Absolute value of correlation coe cients between
acoustic parameters and other ripeness indicators of ‘Sousky-
Sabz’ and ‘Zard-Eyvanekey’ melons
Agric. conspec. sci. Vol. () No.
204 Farhad KHOSHNAM, Moslem NAMJOO, Hossein GOLBAKHSHI
Macrelli E., Rom ani A., Paganelli R.P., Sangiorgi E., Tartagni M.,
(2013). Piezoelectric transducers for real-time evaluation of fruit
rmness. Part II: Statistical and sorting analysis. Sensors and
Actuators A: Physical 201: 497-503.
Mao J., Yu Y., R ao X. and Wang J. (2016). Firmness prediction and
modeling by optimizing acoustic device for watermelons. J Food
Eng. 168: 1-6.
Nukaya A., Ishid a A., Shigeoka H. and Ichikawa K. (1986). Varietal
di erence in respiration and ethylene production in muskmelon
fruits. HortScience 21: 853.
Schotte S., De B elie N. and De Baerdemaeker J. (1999). Acoustic
impulse-response technique for evaluation and modelling
of rmness of tomato fruit. Postharvest Biol Technol. 17(2):
105-115.
Shmulevich I., G alili N. and Howarth M. (2003). Nondestructive
dynamic testing of apples for rmness evaluation. Postharvest
Biol Technol. 29(3): 287-299.
Stone M., Armstr ong PR., Zhang X., Brusewitz GH., Chen D.D.,
(1996). Watermelon maturity determination in the eld using
acoustic impulse impedance techniques. Trans ASAE. 39(6):
2325-2330.
Sugiyama J. (199 4). Firmness measurement of muskmelons by
acoustic impulse transmission. Trans ASAE. 37(4): 1235-1241.
Taniwaki M., Han ada T. and Sakurai N. (2009). Postharvest quality
evaluation of “Fuyu” and “Taishuu” persimmons using a non-
destructive vibrational method and an acoustic vibration tech-
nique. Postharvest Biol Technol. 51(1): 80-85.
Wang J., Teng B. and Yu Y. (2004). Pear dynamic characteristics and
rmness detection. Eur Food Res Technol. 218(3): 289-294.
Woensel G., Verd onck E. and Baerdemaeker J. (1988). Measuring
the Mechanical Properties of Apple Tissue Using Modal
Analysis. J Food Process Eng. 10(3): 151-163.
Yamamoto H., Iwa moto M. and Haginuma S. (1980). Acoustic
impulse response method for measuring natural frequency
of intact fruits and preliminary applications to internal qual-
ity evaluation of apples and watermelons. J Texture Stud. 11(2):
117-136.
Zeng W., Huang X ., Arisona S. M. and McLoughlin I. V. (2014).
Classifying watermelon ripeness by analysing acoustic signals
using mobile devices. Pers Ubiquit Comput. 18(7): 1753-1762.
acs80_30
