POSTHARVEST SORTING OF FRUITS BY OPTICAL GRADERS

1
Prof. Mario ROSSI
Università degli Studi

• POST-HARVEST SORTING OF FRUITS BY OPTICAL GRADERS
  
• Efficiency tests on citrus fruits and table
  tomatoes

INTERNATIONAL SYMPOSIUM ON FRESH PRODUCE SUPPLY
CHAIN MANAGEMENT 6-10 December 2006, Lotus Pang
Suan Kaeo Hotel, Chiang Mai, Thailand
2
Italian fruits and vegetables have lost ground
in terms of competitiveness

• On both domestic and international markets

• Why
• Extreme fragmentation of the business fabric
• Extreme fragmentation of distribution channels
• Delayed enforcement of UE quality regulations
• High quality of imported produce

3
Therefore producers should put a greater emphasis
on post-harvest operations

• Correctly performed and optimized post- harvest
  operations result in a remarkable increase in
  produce quality

• What is more, post-harvest technology is
  continuously developing

4
Fuits and Vegetables Sorting LineGeneral Features
5
Fruits and Vegetables Ranking

• Ranking essentially relies on size, weight,
  colour (extrinsic features)
• Today sophisticated ranking systems are being
  developed to assess also fruit intrinsic features
  (pulp firmness, degree of maturation, chemical
  characteristics, etc..)

Few experimental data are available about the
actual operational parameters of the different
systems used to carry out the most important of
post-harvest operation grading
6
GRADING,Under EU Fruits and Vegetables
Commercialization Regulations grading is
determined by the measurement of maximum chord of
the normal section of the fruit polar axis.
In fruits and vegetables packhouses grading is
performed by continuosly evolving special systems.
7
Mechanical systems are still utilized which are
based on elements including

• Vibrating meshes with gauged holes

Properly distanced tilting rollers
8
However mechanical systems are poorly accurate.
Do not allow for a fast change in sorting schemes
according to the different fruits to be
sorted Need a lot of labour to make up for
grading errors Have a limited operation capacity
and a remarkable overall dimension. Are losing
ground to electronic systems.
Diverging Conveyors System
9
Electronic graders are in most cases based on
vision technology a digital image processor
analyzes fruit images acquired by one or more
videocameras to obtain the geometric sizes
required for correct grading.
10
Operation of Optical GradersCopiare

• I sistemi di movimentazione portano i frutti
  all'interno di una camera di visione, dove una
  telecamera acquisisce le immagini del prodotto
  che transita.
• Nel momento in cui i frutti entrano nella camera
  di visione i doppi coni che li supportano
  cominciano a ruotare, provocando anche la
  rotazione dei frutti stessi, indispensabile per
  permettere la visione di tutta la loro
  superficie. Avviene quindi l'acquisizione di più
  immagini per ogni frutto, che vengono inviate ad
  una scheda di acquisizione con convertitore
  analogico-digitale e riversate in un calcolatore.
• Identificati i parametri dimensionali del frutto,
  il calcolatore lo destina alla corrispondente
  uscita, seguendo il programma di selezione
  impostato dall'operatore. In corrispondenza
  dell'uscita impostata per la classe cui il frutto
  appartiene, parte l'impulso che aziona il sistema
  di espulsione ed il prodotto abbandona il nastro
  di movimentazione per confluire nei bancali di
  lavorazione, dove operatori qualificati
  provvedono alle operazioni di confezionamento

11

• Optical graders are charaterized by
• elevated accuracy
• the possibility of changing sorting schemes
• gentle handling of fruits.

12

• Many are the optical grader models produced today
  which are likely to meet the different needs of
  the most diverse businesses. Producers declare a
  grading accuracy within 1 mm also at a conveyor
  speed of 11 fruits/sec.
• But is this speed real for all fruits and at all
  speeds?
• Therefore it was decided to study the operation
  of two optical graders operating in two modern
  plants in order to obtain data on their
  efficiency on the basis of some preliminary tests
  performed with oranges, lemons and table tomatoes

13
The study focused on the operation features and
on the yield quality of two plants A and B
() the plant is made up of two identical and
independent sorting lines
14
TEST METHODOLOGIES

• The operation capacity of the two graders under
  study was measured by assessing the filling
  percentage of those sectors of the roller
  conveyors destined to held fruits and by applying
  the following relation
• C V x 3600 x R x n fruits/ h, where
• C operation capacity
• V speed, as expressed in the sectors/s (
  fruits/s, 0.1 m/s)
• R percentage of sectors filled
• n number of lanes.

15
TESTS METHODOLOGIES

• Tests were made to determine the accuracy of
  grading in the lines under consideration.

16
TESTS METHODOLOGIES

• Fruits were numbered with a manual gauge.
• They were then placed, one by one, on the
  conveyors of Graders A and B and advanced to pass
  thruogh a vision chamber.
• Ten repeats were made for each fruit.
• Tests were repeated at three speeds of the
  conveyor for A 0.6 m/s, 0.8 m/s and 1.1 m/s
  and at two speeds for B 0.8 m/s e 1.1 m/s.
• Data from the two graders were compared with the
  measurements made with the manual gauge.

17
RESULTS AND REMARKS
Graders A and B where found to have a filling
percentage of the conveyor sectors destined to
fruits of about 70 which remained constant at
different operation speeds. Operation capacity is
extremely elevated.
Operation capacity of the graders under study
Data of one of the two graders of Plant B
18
RESULTS AND REMARKS

• Grading accuracy tests
• Data from the two graders were compared with the
  measurements made with the manual gauge.
• Data were analyzed taking into account the
  absolute error (Ae), i.e. the difference between
  the fruit "equatorial" diameter calculated by the
  grader and that measured by the manual gauge.

• Ae Cm Cr mm
• where
• Cm maximum chord of the section normal to the
  friut longitudinal axis measued by the optical
  device
• Cr maximum chord of the section normal to the
  friut longitudinal axis measued by the manual
  gauge.

The influence on Ae of both fruits speed and
gauge was assessed.
19
RESULTS AND REMARKS

• Absolute error does not tend to increase at
  higher speeds
• - Underestimate error at least for smaller frits

Machine A - Absolute Error (average of 10
repeats) registered for sample fruits (oranges)
20
RESULTS AND REMARKS

• Modest absolute error it does not increase at
  higher speeds
• - It tends to increase in larger fruits

Machine B - Absolute Error (average of 10
repeats) registered for sample fruits (lemons)
21
RESULTS AND REMARKS

• A significant correlation is observed between
  increase in fruit size and increase in absolute
  error
• - Larger fruits are more irregularly shaped and,
  sometimes, do not rotate inside the vision
  chamber, thus preventing the acquisition of
  images of the entire suface of fruits.

Grader A, B - Correlation between fruit gauge and
the absolute error (Ae) of the machine
measurement (grader A- oranges grader B
tomatoes)
22
RESULTS AND REMARKS

• Test repeats highlighted a good level of
  homogeneity in the diameter readings" made on
  the same samples during successive passings
  before the videocamera, as can be inferred by the
  standard deviation average and the variation
  coefficient given in the Table below.

Average standard deviation and variation
coefficient (C.V.) for graders A and B
23
RESULTS AND REMARKS
Copiare Le norme comunitarie in materia di
commercializzazione degli agrumi e del pomodoro
precisano che, per tutte le categorie e per ogni
forma di presentazione, è ammesso solo il 10 in
peso o in numero di frutti che abbiano il calibro
immediatamente inferiore e/o superiore a quello
indicato sulla confezione. Pertanto si è ritenuto
opportuno prendere in considerazione l'errore
assoluto in valore assoluto (Ae) delle
rilevazioni effettuate sui frutti dalle due
calibratrici, rispetto al calibro reale. Sono
stati stabiliti tre intervalli di errore e si è
voluto indicare la percentuale di casi, rilevati
durante le prove alle differenti velocità, che
ricadono in essi.
24
RESULTS AND REMARKS
Copiare Per la calibratrice A, al lavoro su
arance, si è registrato un errore massimo, in
valore assoluto, di 7 mm. Considerando
l'intervallo di errore assoluto compreso tra 5 e
7 mm, si osserva che risultano affette da un
errore di tale entità l'uno per cento delle
misure alle velocità di 0.6 e 1.1 m/s. Alla
velocità di 0.8 m/s non si supera il tre per
cento di casi con tale errore.
Grader A - oranges percentage of cases observed
on three intervals of absolute error in terms of
absolute value (mm) Ae
25
RESULTS AND REMARKS
Copiare I risultati ottenuti con la calibratrice
B evidenziano una precisione più elevata, in
quanto l'errore assoluto massimo è stato di 5 mm
alla velocità di 0.8 m/s e di soli 4 mm alla
velocità di 1.1 m/s. Inoltre più dell'80 dei
frutti sono stati misurati con un errore assoluto
in valore assoluto inferiore ai 2 mm, mantenutosi
inferiore o uguale ad 1 mm nella maggior parte
dei casi.
Grader B lemons and tomatoes percentage of
cases observed on three intervals of absolute
error in terms of absolute value (mm) Ae
26
CONCLUSIONS

• Copiare
• Gli errori registrati sono superiori a quelli
  generalmente dichiarati dalle case costruttrici
  tuttavia, anche quest'ultima elaborazione
  evidenzia che il lavoro di calibratura viene
  svolto accuratamente a tutte le velocità di prova
  ed in particolar modo dalla calibratrice B.
• In entrambi i casi, i dati ottenuti attestano
  livelli di precisione soddisfacenti, soprattutto
  considerando che, volutamente, sono state
  condotte prove "difficili" per le macchine,
  scegliendo frutti-campione di dimensioni e forme
  differenti ed operando anche a velocità superiori
  a quelle comunemente adottate durante le
  lavorazioni e consigliate dalle case
  costruttrici.
• Lanalisi del lavoro condotta sulle
  selezionatrici ottiche ha permesso di evidenziare
  che tali macchine, pur lavorando a velocità
  elevate, che garantiscono una grande capacità di
  lavoro, mantengono unelevata precisione
  nelloperazione di selezione in base al calibro.

27
CONCLUSIONS
Copiare La precisione consente il risparmio di
manodopera alle uscite, in quanto gli addetti di
tale settore devono semplicemente trasferire i
frutti nei contenitori appositi, mentre, negli
impianti tradizionali, a volte, erano necessari
un ultimo controllo e comunque una maggiore
attenzione, alle uscite, per compensare gli
eventuali errori di calibratura.
28
CONCLUSIONS
The machines under study are in addition
characterized by the possibility of varying the
grading schemes selected by simply acting on the
control panel. They are also able to minimize
fruit damage thus responding to the demand of
quality on the part of consumers. Accuracy and
the possibility of processing different types of
fruits result in a prolonged use of the graders
throughout the year thus reducing both labour and
post-harvest operation costs. In order to obtain
better results optical graders need careful
upkeep and calibration as well as well skilled
workers trained to manage at best the control
systems of their optical/electronic
devices. The selection of graders must be
carefully weighed based on their own technical
specifications and on the needs of their users.
29
Prof. Gennaro GIAMETTA
Università degli Studi Mediterranea di Reggio
Calabria

• POST-HARVEST SORTING OF FRUITS BY OPTICAL GRADERS
  
• Efficiency tests on citrus fruits and table
  tomatoes

INTERNATIONAL SYMPOSIUM ON FRESH PRODUCE SUPPLY
CHAIN MANAGEMENT 6-10 December 2006, Lotus Pang
Suan Kaeo Hotel, Chiang Mai, Thailand