Research Article
Deep Learning Based on Residual Networks for Automatic
Sorting of Bananas
Abdulkader Helwan ,
1
Mohammad Khaleel Sallam Ma’aitah ,
2
RahibH.Abiyev ,
3
Selin Uzelaltinbulat ,
4
andBengiSonyel
5
1
School of Engineering, Lebanese American University, Byblos, Lebanon
2
Department of Management Information Systems, Near East University, Nicosia, TRNC, Mersin 10, Turkey
3
Department of Computer Engineering, Near East University, Nicosia, TRNC, Mersin 10, Turkey
4
Department of Computer Information Systems, Near East University, Nicosia, TRNC, Mersin 10, Turkey
5
Department of Educational Sciences, Eastern Mediterranean University, Famagusta, Mersin 10, Turkey
CorrespondenceshouldbeaddressedtoRahibH.Abiyev;rahib.abiyev@neu.edu.tr
Received 13 February 2021; Accepted 25 March 2021; Published 8 April 2021
AcademicEditor:RijwanKhan
Copyright©2021AbdulkaderHelwanetal.isisanopenaccessarticledistributedundertheCreativeCommonsAttribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
is study presents the design of an intelligent system based on deep learning for grading fruits. For this purpose, the recent
residuallearning-basednetwork“ResNet-50”isdesignedtosortoutfruits,particularlybananasintohealthyordefectiveclasses.
edesignofthesystemisimplementedbyusingtransferlearningthatusesthestoredknowledgeofthedeepstructure.Datasets
ofbananashavebeencollectedfortheimplementationofthedeepstructure.esimulationresultsofthedesignedsystemhave
shown a great generalization capability when tested on test (unseen) banana images and obtained high accuracy of 99%. e
simulationresultsofthedesignedresiduallearning-basedsystemarecomparedwiththeresultsofothersystemsusedforgrading
the bananas. Comparative results indicate the efficiency of the designed system. e developed system can be used in food
processing industry, in real-life applications where the accuracy, cost, and speed of the intelligent system will enhance the
productionrateandallowmeetingthedemandofconsumers.esystemcanreplaceorassisthumanoperatorswhocanexert
their energy on the selection of fruits.
1.Introduction
In the food industry, the quality of processed fruits is ex-
tremelyimportant.Meetingthedemandsoftheconsumers
andproducinghigh-qualityfruitsattheproductionlineata
very fast rate requires the implementation of high-perfor-
mancetechnologies[1].Moreover,thefoodindustryisone
of the few fields which have restricting conditions and
constraintsduetoitsdependencyonweatherconditionsand
the labor market [2]. For example, if the fruits were not
harvested at the most suitable time due to weather condi-
tions, the quality and quantity of the harvest may decrease
duetobadweatherconditionsandexcessiveripeningofthe
fruits. Over the years, the most technological processes in
this industry were mainly controlled by human operators.
Some delicate tasks such as postharvest and grading of
healthyanddefectiveproductswerebasedonhuman-made
decisions. Human operators are sometimes exposed to the
tirednessoftheeyesduetolackofsleepandfatiguecaused
by overworking that can affect their performances. Fruit
sorting is a decision-making task which is based on some
visual features of the fruit and decides whether a fruit is
healthyordefectivewhenitpassesthroughaconveyorbelt.
erefore, it is a computer vision problem which can be
perfectlysolvedusingmachinelearningthatcanpreventthe
errors caused by human operators [3].
Recently,differentresearchworkshavebeenperformed
forcontrollingandgradingoffruitsusingcomputervision
andmachinelearningtechniques.ecommonapplications
areclassificationandsortingoffruits[4,5],identificationof
Hindawi
Journal of Food Quality
Volume 2021, Article ID 5516368, 11 pages
https://doi.org/10.1155/2021/5516368