AAGF-CNN: An Android Application for Grading

Fruits using CNN

Epics Project Report submitted in partial fulfillment of the Requirements for the Award
of the Degree of

BACHELOR OF TECHNOLOGY
in
COMPUTER SCIENCE AND ENGINEERING
By

M.SAI MANVITHA (198W1A0532)

P.JAYA VINEELA (198W1A0544)

P.RISHITHA (198W1A0546)

Under the Guidance of

Mr. PRABU.U, M.Tech.,
Assistant Professor

DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

VELAGAPUDI RAMAKRISHNA
SIDDHARTHA ENGINEERING COLLEGE
(AUTONOMOUS)
Approved by AICTE, NAAC A+, NBA Accredited
Affiliated to Jawaharlal Nehru Technological University, Kakinada
Vijayawada-520 007
2022
 VELAGAPUDI RAMAKRISHNA
SIDDHARTHA ENGINEERING COLLEGE
(AUTONOMOUS)
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING

CERTIFICATE

This is to certify that the EPICS PROJECT Report entitled “AAGF-CNN: An An-
droid Application for Grading Fruits using CNN” being submitted by

M.SAI MANVITHA 198W1A0532

P.JAYA VINEELA 198W1A0544

P.RISHITHA 198W1A0546

in partial fulfilment for the award of the Degree of Bachelor of Technology in Computer
Science and Engineering to the Jawaharlal Nehru Technological University, Kakinada is
a record of bonafide work carried out under my guidance and supervision.

Mr.PRABU.U M.Tech Dr. D. Rajeswara Rao, M.Tech, Ph.D

Assistant Professor & Guide Professor & HOD,CSE

i
 DECLARATION

We hereby declare that the EPICS PROJECT entitled “AAGF-CNN: An Android

Application for Grading Fruits using CNN” submitted for the B.Tech Degree is our
original work and the dissertation has not formed the basis for the award of any degree,
associateship, fellowship or any other similar titles.

Place: Vijayawada M.SAI MANVITHA (198W1A0532)

Date: P.JAYA VINEELA (198W1A0544)

P.RISHITHA (198W1A0546)

ii
 ACKNOWLEDGEMENT

Behind every achievement lies an unfathomable sea of gratitude to those who

activated it, without whom it would ever have come into existence. To them we lay the
words of gratitude imprinted with us.

We would like to thank our respected Principal, Dr. A.V. Ratna Prasad and Dr.
D.Rajeswara Rao, Head of the Department, Computer Science and Engineering for their
support throughout our Project.

It is our sincere obligation to thank our guide, Mr. Prabu. U, Assistant Professor,
Department of Computer Science and Engineering, for his timely valuable guidance and
suggestions for this Project.

We owe our acknowledgements to an equally long list of people who helped us in

this Project. Finally, we wish to thank all the supporting staff who gave us facility in the
lab for the completion of this Project.

Place: Vijayawada M. SAI MANVITHA (198W1A0532)

Date: P.JAYA VINEELA (198W1A0544)

P.RISHITHA (198W1A0546)

iii
 ABSTRACT

Traditional manual visual grading of fruits has been one of the important challenges faced
by the agricultural industry due to its laborious nature as well as inconsistency in the
inspection and classification process. The task of fruit grading is vital because there is a
great demand for high-quality fruits in the market. Earlier they used to grade the fruits
through Machine Learning and Spectrophotometry. In this proposed system, the grading
of fruits is done by using a Convolutional Neural Network (ConvNet/CNN) which is a
Deep Learning technique. It is an efficient and effective machine vision system based on
the deep learning techniques and it offers a non-destructive and cost-effective solution for
automating the visual inspection of fruit freshness, ripeness and appearance. The pro-
posed system captures the image with a camera and converts it to pixels and preprocess
the image into a deep learning model which displays whether the fruit is healthy or de-
fected. The real time system efficiently segments multiple instances of the fruits from an
image then grades the individual objects (fruits) accurately. The system was trained and
tested on two data sets (apples and bananas). The test results show that the system can
sort 89% bananas and 86% apples accurately when tested with real time images.

Keywords: Fruit grading, Visual inspection, Convolutional Neural Networks, Image

Preprocessing, Ripeness assessment.

iv
 Table of Contents

1 INTRODUCTION 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Village Visited . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Photo with Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.5 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.6 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.7 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.8 Organization of the report . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 LITERATURE SURVEY 6
2.1 Date Fruit grading using Computer Vision . . . . . . . . . . . . . . . . . . 6
2.2 Determining the tomato and date fruit quality using Direct Color Mapping 6
2.3 Peach fruit grading using Spectroscopy . . . . . . . . . . . . . . . . . . . . 7
2.4 Mango grading using Machine Vision . . . . . . . . . . . . . . . . . . . . . 7
2.5 Rician k-Factors-Based Sensor for Fruit Classification . . . . . . . . . . . . 8
2.6 Evaluation indices of sour flavor for apple fruit . . . . . . . . . . . . . . . . 8
2.7 Fractional fuzzy 2DLDA approach for pomegranate fruit grade classifica-
tion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.8 Automatic Counting and Individual Size and Mass Estimation of Olive
Fruits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.9 Quality Assessment of Fresh Fruits using Non-Invasive Sensing . . . . . . 9
2.10 Mulberry grading using using artificial neural networks (ANNs) and sup-
port vector machine (SVM) . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.11 Apple Bruise Grading using Hyperspectral Imaging data . . . . . . . . . . 10
2.12 Fruit Grading with Spectrophotometry and ML Approaches . . . . . . . . 11
2.13 Findings of the survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 SOFTWARE REQUIREMENTS ANALYSIS 12

3.1 Functional Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.1 Software Requirements . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.2 Hardware Requirements . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Non-Functional Requirements . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2.1 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

4 PROPOSED SYSTEM 13
4.1 Proposed System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.2 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2.1 Data collection and pre-processing . . . . . . . . . . . . . . . . . . 14
4.2.2 Training the Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

v
 4.2.3 Fruit Grading Algorithm . . . . . . . . . . . . . . . . . . . . . . . . 14
4.2.4 Android Application Development . . . . . . . . . . . . . . . . . . . 15
4.2.5 Integration to Mobile Application algorithm . . . . . . . . . . . . . 15

5 Implementation 17
5.1 Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
5.2 Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
5.3 Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.4 Client Satisfaction Report . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

6 CONCLUSION AND FUTURE WORK 26

6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
6.2 Future work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

REFERENCES 27

vi
 List of Tables

4.1 List of Nomenclatures for Fruit Freshness detection algorithm . . . . . . . 15

4.2 List of Nomenclatures for Integration of CNN model to android application
algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

5.1 Banana and detected accuracy . . . . . . . . . . . . . . . . . . . . . . . . . 22

5.2 Apple and detected accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . 23

vii
 List of Figures

1.1 Kanuru . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Photo with client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Photo with client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

4.1 Proposed System Architecture . . . . . . . . . . . . . . . . . . . . . . . . . 13

4.2 Apple and Banana images from dataset . . . . . . . . . . . . . . . . . . . . 14

5.1 Grading fruit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

5.2 Real Time apple images used for testing . . . . . . . . . . . . . . . . . . . 18
5.3 Fresh Banana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.4 Fresh apple . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
5.5 Ripe banana . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.6 Rotten Apple . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
5.7 Fresh apple . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
5.8 Training and validation Accuracy Plot . . . . . . . . . . . . . . . . . . . . 24
5.9 Training and validation Loss Plot . . . . . . . . . . . . . . . . . . . . . . . 24
5.10 Client Satisfaction Report . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

viii
 Chapter 1
INTRODUCTION

1.1 Introduction
Fruits are key components of the human diet in terms of both nutrition and
economics [1]. In the agriculture sector, the demand for fruits increases day by day. Tra-
ditional grading of fruits is time-consuming, inconsistent, and tedious. Apple is often
regarded as the most renowned fruit globally, with strong market demand and an annual
output of 80 million tonnes [2]. Apple fruits are usually damaged during the harvesting
and shipping, which may produce visible bruising in apples, lowering their market value
and it affects the fruit quality [3],[4]. Sensory evaluation [5–7] and physicochemical anal-
ysis [8–10] are the most used approaches for detecting and assessing fruit bruises. Manual
grading of these fruits is a difficult task. In general, precise apple bruise grading and the
significance of early diagnosis and grading cannot be emphasized [11]. Classifying the
ripe fruits physically is a costly process in agriculture, and errors caused by humans can
result in inaccurate outcomes [12]. Maintaining accuracy and consistency in fruit grad-
ing is made more difficult by manual processing through a visual interpretation, which is
time-consuming [13].

In the past years, the different fruits considered for grading are Dates, Apple,
Mango, Peach, Pomegranate, Tomato, Olive, and Mulberry fruits [13-18]. The distin-
guishing features used to classify the fruits into different quality categories are Flabbiness,
size, shape, intensity, Maturity, Freshness, Ripeness, firmness, Color, Acidity, actual-days-
to-rot, and moisture content. These characteristics are used to determine the quality of
the fruit. In recent years, the techniques used for evaluating the fruit grade are Computer
Vision, Spectroscopy, Machine Vision, Direct Color Mapping, Back Propagation Neural
Networks, Support Vector Machine (SVM), Artificial Neural Networks (ANN), Hyper-
spectral Imaging, Non-Invasive Sensing, Spectroscopy, and Machine Learning [11-16]. For
classification, Linear discriminant analysis (LDA) and Quadratic Differential Analysis
(QDA) are used [19].

Wan, Peng, et al. [20] detected three stages of ripeness of tomatoes using com-
puter vision and the backpropagation neural network approach. Peng, Hongxing, et al.
[21] asserted that the SVM approach is suitable for estimation and classification. Pattern
recognition, image processing, and medical applications are all possible using the SVM
method. SVM is used to detect patterns and textures, estimate harvest times and de-
termine product quality. To distinguish between ripe and broken fruit, they employed
differentiation models. Mizushima, Akira, and Renfu Lu [22] used the SVM and the Otsu
thresholding approach to create an autonomously configurable system that assessed ap-
ple fruit based on its perceived color. Sun, Chengming, et al. [23] used the SVM and
image processing methods to evaluate the position and kind of whiteness in rice grains
with great accuracy in another investigation. Recently, researchers have concentrated on
non-destructive approaches for detecting and grading damaged fruits, particularly those
based on machine vision and spectral imaging techniques [24],[25].

1
 For agricultural grading operations, several monitoring systems have been devel-
oped. A machine vision technique called direct color mapping was presented to evaluate
the quality of tomatoes and dates [17]. It is easy to modify color choices or grading fac-
tors using these approaches. This color mapping approach calculates a unique coefficient
set intended to convert color space using predefined colors of relevance particular to a
specific application. This strategy used a second-order polynomial function to transform
a three-dimensional RGB color space of interest particular to a specific application to a
one-dimensional maturity function in the gradient domain.

Machine vision algorithms are extensively used to acquire optical images of un-
bruised and damaged fruits and recognize bruises on fruits using appearance descriptors
such as peel color, local form, and local texture features [26–28]. Pawar, Meenakshi M.,
and Meghana M. Deshpande [26] used spatial gray-level dependency matrices (SGDM)
and discrete wavelet treatment to retrieve the local color characteristics and textural as-
pects of these pomegranate fruits in the RGB color space (DWT). The support vector
machine (SVM) was utilized to detect defective pomegranate fruit areas.

Huang et al. [28] employed an 85 % accuracy combination of the SVM classi-

fier and the Gabor texture descriptor to identify damaged apples. Zhu, Bin, et al. [29]
used near-infrared images to generate high-dimensional Gabor properties that were subse-
quently dimensioned utilizing kernel principal component analysis (KPCA). The bruised
apples were spotted using SVM with SVM 90.6% accuracy.

The proposed system grades apple,banana and orange fruits using Deep learning
technique Convolutional Neural Networks(ConvNet/CNN). It is an automated android
application which is beneficial for the fruit vendors which provides the guarantee to cus-
tomers regarding the fruit quality. The dataset used for fruit grading is collected from
the public dataset “Fruits-360” referred from “Efficient fruit grading system using spec-
trophotometry and machine learning approaches” [12]. This dataset contains 1581 images
of fresh bananas, 2224 images of rotten bananas, 1693 images of fresh apples and 2342
images of rotten apples. All the images are in high resolution with white background
used for training. The test set of images have 381 images of fresh bananas, 530 images of
rotten bananas, 395 images of fresh apples and 601 images of rotten apples. Image data
generator function has been used to rotate images between 0 to 360 degrees.

The image is preprocessed into CNN Deep learning model. In order to classify
the fruits we train the CNN model.CNN model contains Convolutional layers with 20,20,3
as shape of layers and 3x3 as kernel size. It has 7 layers and ReLU activation function
is used in all layers. The output layer uses softmax activation function. The model is
trained with Fruits-360 dataset and it will predict the class label of the extracted image
from the trained CNN model. Then it is validated and tested using test dataset. Based
on the extracted features it will classify the fruit into particular grade.There are two class
labels fresh and rotten for both banana and apple fruits.

The model generated is converted to .tflite format and java files are included in
the android studio. The trained model is deployed and integrated into android applica-
tion.The interface is designed using XML files. The application uses mobile camera to

2
capture the real fruit images using Open CV techniques. It will update the probability
percentage based on the class label of the extracted image.The probability percentage of
fruit’s freshness and upto what extent it is rotten is displayed on the mobile screen.This
application works fine when tested with real-time images.

1.2 Motivation
The fiber and vitamins present in fruits/vegetables are crucial for a healthy body.
Phytochemicals in fruits and vegetables help to protect human beings from cancer and
other diseases. They are essential to maintain good health, but consumption of substan-
dard fruits and vegetables lead to adverse effects and poor health. Therefore, thorough
identification of edible and inedible fruits and vegetables are to be known. At present,
most existing fruit quality detecting and grading system have the disadvantage of high
cost and complexity. So, it is significant to develop high speed and low-cost fruit size
detecting and grading system.

1.3 Village Visited

We interacted with fruit vendor Paidipati Lakshmi garu who lives in Kanuru
village which is located at Kanuru in Krishna district to do our project. We considered
him as a primary client for our project. Not only that fruit vendor but also the other
fruit vendors in that village are also clients for our project. Figure 1.1 shows the map of
kanuru village.

Figure 1.1: Kanuru

1.4 Photo with Client

Figure 1.2,1.3 shows the photograph taken with our client Paidipati Lakshmi garu.

3
 Figure 1.2: Photo with client

Figure 1.3: Photo with client

1.5 Problem Statement

Fruits are rich in vitamins and minerals which reduces human diseases. Grading
of fruits is important because it helps people to have better quality fruits enhancing food
safety. Despite of advancement in computer vision, image processing, recognition and
advancement in computer technology, automatic fruit classification is a challenging task.
The primary parameters that play vital role while classifying a fruit include the machine
learning algorithm that is being used, quality of images in the fruit database, fruit’s
images’ shape and size and fruit’s color. Secondary parameters that affect the classification
are similar characters of fruits like color, shape, size, etc. If both primary and secondary
parameters are not analyzed properly in the beginning then it may cause problem during
classification and may lead to less accuracy and unexpected results. Many related works
have been conducted in fruit classification using different classification algorithms but
those approaches still lack in some aspects. The proposed system is trained with apple
and banana dataset that trains itself using color, size, texture parameters. The system is
proposed to capture image of fruit using mobile camera and predict grade of fruit in less
time and with more accuracy as possible.

1.6 Scope
• The scope is limited to fruits like apples and bananas only.
• Our mobile application is limited to only fruit vendors.

4
1.7 Objectives
• To grade the fruit based on the Ripeness, color, and size.

• To analyze the quality of fruit.

• To develop an efficient and affordable software which is beneficial for fruit vendors.

1.8 Organization of the report

The organization of the report is as follows. Chapter 2 describes about the survey
of the project related papers. It also comprise of merits and drawbacks of those papers.
Chapter 3 describes the software requirement analysis. Chapter 4 describes the architec-
ture diagram, modules, and algorithms of the proposed system. Chapter 5 describes the
implementation platform and experimental results. Chapter 6 concludes the report.

5
 Chapter 2
LITERATURE SURVEY

This chapter describes about the survey of the project related papers. It also
comprise of advantages and disadvantages of those papers.

2.1 Date Fruit grading using Computer Vision

Al Ohali, Yousef [30] proposed a system that applied computer vision technique to classify
dates based on flabbiness, defects, size, shape, and intensity. Through manual inspection
humans can’t efficiently grade the date fruits as it consumes more time and requires more
labour to perform the task. The proposed system is a sorting system which grades the
dates using Computer vision technique. The system captured RGB images and classified
the dates into 3 grades based on defects and quality of dates. The system uses Back
Propagation Neural Networks in sorting the dates. In order to test the accuracy of the
computer-mediated system they took samples of existing dates. The results of the test
proved that it can uniformly sort 80% dates.

Advantages

• This model helps the date growers in the harvesting season for grading and sorting
of dates.

• This model is consistent in grading and maintains uniformity in sorting.

Disadvantages

• This system cannot accurately estimate the bruises on date fruits.

• The date fruits were not graded properly, they were misclassified due to the limited
visibility of the defects.

2.2 Determining the tomato and date fruit quality

using Direct Color Mapping
Lee, Dah-Jye, James K. Archibald, and Guangming Xiong. [31] proposed a system that
applied Direct Color Mapping technique to classify tomato and date fruits based on color.
Through manual inspection humans can’t efficiently grade the fruits as it consumes more
time and requires more labor to perform the task. The proposed color grading system
grades the fruits using the Direct Color Mapping technique. The system captures RGB
images and it is easy for the user to set the color preferences and adjust the grading
parameters according to our convenience

6
Advantages
• Selection and adjustment of color preferences are much easier when compared to
other color grading techniques. Color mapping works efficiently with higher order
polynomials and non-linear functions.
Disadvantages
• Color index histogram analysis can be done using median instead of using average
value for color quality.

2.3 Peach fruit grading using Spectroscopy

Matteoli, Stefania, et al. [32] proposed an automated approach for peach fruit matu-
rity grading that, by exploiting fiber-optic spectroscopy-based sensors and multivariate
processing techniques, minimizes the operator intervention while reducing discharge and
waste. The use of a spectroscopic sensor complies with the so-called nondestructive mea-
surement method, Maturity grading is accomplished by retrieving estimates of the fruit
flesh firmness by means of multivariate retrieval techniques applied to the reflectance
spectra acquired with the spectrometer and by processing the retrieved values within the
framework of a maturity fuzzy classifier.

Advantages
• It is useful for industries and horticulture applications.

• It detects the fruit flesh’s firmness without causing any damage to the fruit.
Disadvantages
• Did not use Spectroscopic sensors, use of these sensors improves maturity grading
performance of peach fruits

2.4 Mango grading using Machine Vision

Nandi, Chandra Sekhar, Bipan Tudu, and Chiranjib Koley[33] proposed the work on in-
telligent machine vision-based techniques for grading mangoes in four different categories.
In this, the video image is captured by a charge-coupled device camera placed on the top
of a conveyer belt carrying mangoes. For maturity prediction in terms of actual-days-to-
rot, support vector regression has been employed and for the estimation of quality from
the quality attributes, multi attribute decision making system has been adopted. Finally,
fuzzy incremental learning algorithm has been used for grading based on maturity and
quality.

Advantages
• Proposed system predicts the actual-days-to-rot and estimates the mango quality
with considerable accuracy using machine vision.

• Average time to grade one mango is approximately 0.4 second

7
Disadvantages

• There is a problem in detecting the firmness from this vision-based measurement.

• They have considered only one side of mango image; a mechanical setup should be
introduced to rotate the mango within the conveyer belt to get the image of the
other side.

2.5 Rician k-Factors-Based Sensor for Fruit Classifi-

cation
Leekul, Prapan, et al. [34] proposed a non-destructive, non-invasive sensor system for
classifying durian fruits according to the maturity stages. The design of the sensor sys-
tem has adopted the principle of wireless communications in which the Rician k-factors
were utilized to determine the fruit maturity. The simulations were thus carried out using
a smooth surface so as to minimize the computational resources requirement.

Disadvantages

• Apart from the durians, the proposed sensor system could be applied to determining
the maturity stage of several other fruits.

2.6 Evaluation indices of sour flavor for apple fruit

Zhen, Y. A. N., et al. [35] proposed a novel method of Evaluation indices of sour flavour
and also grading standard. Apple is a typical fruit that is dominated by malic acid content
among all organic acids. The average content of malic acid in apples in the present study
accounted for 94.5% of ToA, and the content order of organic acids in most apples was
Mal>Oxa>Cit>Lac>Suc>Fum A large sample of apples (106 cultivars) and regression
analysis were used to discover significant linear or logarithmic relationships between five
apple sour flavor indices, namely Mal, ToA, AcV, TiA, and pH value. At present, the
evaluation indices of sour flavor for apple are generally expressed in the absolute value of
titratable acid content

Advantages

• This study provides a scientific basis for evaluating apple flavor and selecting apple
cultivars.

2.7 Fractional fuzzy 2DLDA approach for pomegranate

fruit grade classification
Gurubelli, Yogeswararao, Malmathanraj Ramanathan, and Palanisamy Ponnusamy [36]
proposed a novel methodology of pomegranate fruit grade classification. This work is
about classification of pomegranate by using fractional fuzzy. It also uses class scat-
ter matrix and also traditional 2D feature extraction techniques. In this method they
have used 4 types of linear discriminant analysis traditional 2DLDA, Fractional 2DLDA,

8
Fuzzy 2DLDA and FF2DLDA. In all these methods FF2DLDA (Fractional Fuzzy Two-
Dimensional Linear Discriminant Analysis) is the method it has produced the best results
in this classification.

Advantages

• Out of four algorithms the proposed fractional fuzzy 2DLDA gives best recognition
rate

Disadvantages

• Horizontal and Vertical image feature extractions using conventional and fuzzy
2DLDA approaches are not implemented

• Using another classifier viz. K-Nearest Neighbour, fuzzy C Means, neural networks
and all these techniques optimized with heuristic techniques can also be checked.

2.8 Automatic Counting and Individual Size and Mass

Estimation of Olive Fruits
Ponce, Juan Manuel, et al. [37] presents a new methodology aimed at supporting accurate
automatic olive-fruit grading by using computer vision techniques and feature modeling.
An image analysis algorithm, based on mathematical morphology, was designed to indi-
vidually segment olives and extract descriptive features. The image analysis algorithm
was conceived to deal with complex situations as fruit touching, or fruit cut by the image
borders.

Advantages

• The present paper proposes a new methodology for the automatic counting, and the
individual mass and size estimation, of olive-fruits in digital images.

• Regarding estimation models, and by relying on external validation results, variety-

dependent models outperformed the independent ones when estimating olive mass.

Disadvantages

• It argues for the methodology could operate under a real scenario

2.9 Quality Assessment of Fresh Fruits using Non-

Invasive Sensing
Ren, Aifeng, et al. [38] proposed viable, cost-effective, and feasible technique for recogniz-
ing defects based on Moisture Content in fruits. In this, the Moisture Content in fruits is
observed using THz(terahertz) waves in 0.75 to 1.1 THz frequency range. And then time,
frequency, and time-frequency domain features extracted and are applied ML algorithms
namely SVM, KNN, and Decision Tree. All three algorithms have shown significant re-
sults and SVM performance is more compared to others.

9
Advantages

• The proposed technique demonstrates the strong potential in the discipline of food
and science technology by integrating ML with THz waves to assess real-time infor-
mation of fruits on different days at the cellular level.

2.10 Mulberry grading using using artificial neural

networks (ANNs) and support vector machine
(SVM)
Azarmdel, Hossein, et al. [39] have applied Image processing and AI techniques to classify
mulberries as per ripeness. A total of 577 mulberries images were captured by imaging
system and using feature selection, their geometrical properties were extracted. ANN and
SVM are applied for classification of mulberry fruit. ANN is used with feature selection
methods like CFS and CONS and later resulted in more accuracy. Finally, ANN model
with features selected using CONS subset method with least no. of inputs and high clas-
sification accuracy with low MSE value is proposed as suitable model.

Advantages

• ANN model used with CONS subset method supports least number of inputs and
high accuracy and can be used for online applications.

• Machine vision can be used as an accurate, fast, and non-destructive method to

classify and grade mulberry fruit, reduce product loss and increase its marketability.

2.11 Apple Bruise Grading using Hyperspectral Imag-

ing data
Tang, Yu, et al. [40] proposed that Apple fruits gets bruises easily decreasing its quality.
This paper provides us an efficient bruise grading method for apples that depends pri-
marily on hyperspectral imaging(HSI). A real time spectral information is obtained from
apples using hyperspectral reflectance imaging device and these are bruised purposefully
to obtain spectral data. The data obtained is preprocessed using Procrustes analysis(PA)
and piecewise nonlinear curve fitting(PWCF) method is applied. Later error-correction
output coding-based support vector machine(ECOC-SVM) is used to grade apples with
bruises and the proposed system obtained an accuracy of 97.33%.

Advantages

• This model compares various algorithms with ECOC-SVM to grade apples and
proves it as best with 97.33% accuracy.

10
2.12 Fruit Grading with Spectrophotometry and ML
Approaches
Chopra, Hetarth, et al [41] proposed an efficient AI-based machine that uses spectropho-
tometry computer vision for grading of fruits. The proposed system identifies the fruit
fed to it and then using spectroscopy and ML approaches, grade is predicted. This model
uses dataset from Unitec’s Apple Sorting and Grading Machine and uses H2O’s Driver-
less.AI. With real-time data the accuracy of segregation peaked 72%.

Advantages

• This model segregates fruits efficiently and has been fully automated, AI based and
cost effective.

Disadvantages

• When the model is tested on unseen real-time data, it got only 72% accuracy which
can be improved along with some features.

2.13 Findings of the survey

The existing systems uses Heuristic Search technique for selecting the most impor-
tant features, watershed transform for feature extraction for apples. The mathematical
models used are Cartesian Equation for image segmentation, Singular Value Decomposi-
tion for grading tomatoes based on direct color mapping, Least Squares, tansig and logsig
for training Neural Network. Some of the researches grade the fruits by calculating acidity
and determining grade based on observed values. Spectroscopy, non-invasive sensing tech-
niques are applied at the initial stage for selecting features and then used deep learning
techniques. The techniques include Support Vector Machine, KNN for classifying apples,
mangoes into different classes. Decision Tree for evaluating options. The classification is
based on supervised learning. Multiple linear Regression is used for classification of peach
fruits. SVM is used for grading apples.

In the research papers, the features are confined to size, color and shape and
sometimes goes beyond that based on type of technique used, For example, Non invasive
sensing used the internal chemical composition of fruit for classifying. The major setback
is the complexity of machine and performance of proposed system on real time dataset.
Many other features can also be considered for classification. Also, the systems can be
made efficient using latest algorithms like Convolution Neural Network and AlexNet etc.

11
 Chapter 3
SOFTWARE REQUIREMENTS ANALYSIS

This chapter describes the functional and non-functional requirements of the

project.

3.1 Functional Requirements

Functional Requirements specify systems behavior about what the system does
for any action. Functional requirements define product features and focus on user require-
ments. The requirements are use cases, user stories, prototypes, etc. It determines if it
requires any computations, inputs, and outputs.

3.1.1 Software Requirements

The list of software requirements for the project includes the following
Programming Languages :Python 3.8, java jre 11
Environments :Anaconda, jupyter Notebook, Android Studio
Libraries used in our project are :numpy, pandas, keras, tensorflow, matplotlib

3.1.2 Hardware Requirements

The list of hardware requirements for the project include the following
Processor : i5-7200U
Hard Disk : 500 GB
RAM : 8GB
OS : Windows 10
GPU : NVIDIA GeForce MX350

3.2 Non-Functional Requirements

Non-functional requirements define system behavior, features, and general characteristics
that affect the user experience. How well non-functional requirements are defined and ex-
ecuted determines how easy the system is to use, and is used to judge system performance.
Non-functional requirements are product properties and focus on user expectations.

3.2.1 Performance
Performance is the measure of speed, responsiveness, and stability of a system when it
is under a workload. The performance of a system defines how capable a system is at
completing a task precisely without any issues at any point of time of its job. A system
performance defines whether it is a novel version of work or not and whether it has
considerable working capability in completing a typical task under any circumstance. So
performance is also an important measure that needs to be considered while developing
a project.

12
 Chapter 4
PROPOSED SYSTEM

This chapter describes about proposed system architecture, modules description

and algorithm.

4.1 Proposed System Architecture

The architecture Diagram gives the ideology of each step in the fruit grading. It
is mainly divided into three modules namely data collection and preprocessing, training
the model and integrating the model with android application.

Figure 4.1: Proposed System Architecture

The phase data collection and pre-processing is concerned with data and training the
model deals with developing proposed deep learning model. The model integration phase
deals with integrating the developed model with mobile application for real-time use.
Figure 4.1 describes about these modules.
The steps involved in data collection and pre-processing includes usage of Image data
generator for pre-processing the apple and banana images ready to be provided to train
the model. The model is then developed using the training dataset and the trained model
is converted to .tflite format useful for integrating the model to android application.

13
4.2 Methodology
The proposed system is divided into three modules

1. Data collection and pre-processing

2. Training the model.

3. Android application development

4.2.1 Data collection and pre-processing

The dataset used for fruit grading is “Fruits-360” Referred from “Efficient fruit
grading system using spectrophotometry and machine learning approaches” [18]. This
dataset contains 1581 images of fresh bananas, 2224 images of rotten bananas, 1693 images
of fresh apples and 2342 images of rotten apples. All the images are in high resolution
with white background used for training. The test set of images have 381 images of fresh
bananas, 530 images of rotten bananas, 395 images of fresh apples and 601 images of
rotten apples. Image data generator function has been used for image pre-processing for
rescaling. Figure 4.2 shows some samples from dataset.

Figure 4.2: Apple and Banana images from dataset

4.2.2 Training the Model

CNN model is developed for this system that contains Convolutional layers with
20,20,3 as shape of layers and 3x3 as kernel size. It has 7 layers and ReLU activation
function is used in all layers. The output layer uses softmax activation function. The
model is trained with Fruits-360 dataset and model returns class as fresh or rotten for
input image indicating if it is apple or banana also. Then it is validated and tested using
test dataset.

4.2.3 Fruit Grading Algorithm

This algorithm is used for developing a model that is able to classify fruits based
on its quality. There are two class labels fresh and rotten for each of banana and apple.
The Table 4.1 provides the information related to all the variables that are used in the
fruit grading algorithm.

14
 Table 4.1: List of Nomenclatures for Fruit Freshness detection algorithm

Notation Description
¥ List of training images
€ List of validation Images
£ List of testing images

Algorithm : Fruit Freshness Detection

Input: Apple, Banana images (Fresh and rotten)

Output: Class label representing grade fruit type and whether it is fresh or not
1: Begin
2: Call flow_from_directory() for training data
3: ¥:=Read(images)
4: ¥:=augment(¥)
5: Call flow_from_directory() for training data
6: €:=Read(images)
7: €:=augment(€)
8: Use sequential model from keras
9: Add conv2d layer with input shape(20,20,3)
10: Add MaxPool2D layer with pool size(2,2)
11: Add conv2d layer
12: Add MaxPool2D layer with pool size(2,2)
13: Flatten to reduce dimensions
14: Add Dense layer
15: Add Dense layer
16: Add Dense layer (softmax activation function)
17: Compile Model
18: Train the model with images in ¥
19: Call flow_from_directory() for test data
20: £ :=Read(images)
21: £ :=augment(£)
22: Validate and test the model with images in € and £
23: End

4.2.4 Android Application Development

The trained CNN model is saved as .h5 file and then converted to .tflite file so that
it becomes easy to integrate to android application. XML files for designing interface of
application are created and java code is written for integrating the model. The application
opens camera of mobile directly and predicts the class lo fruit.

4.2.5 Integration to Mobile Application algorithm

The model generated is converted to .tflite format and included in java files in the
android studio. The interface is designed using XML files. The application uses camera

15
from mobile and grades the fruits accordingly. The Table 4.2 provides the information
related to all the variables that are used in the integration algorithm.

Table 4.2: List of Nomenclatures for Integration of CNN model to android application
algorithm
Notation Description
¥ List of class labels
€ Model.tflite

Algorithm : Integration of CNN model to android application

Input: CNN model in .tflite format

Output: Android application for fruit grading
1: Begin
2: Save the CNN model in .tflite format.
3: ¥:=read(labels)
4: €:=read(model.tflite)
5: Create new project named fruit grading app
6: Load tflite packages
7: Create new activity
8: Create application layout
9: Add € and ¥ to assets
10: Link m and l to application using java
11: Run the application.
12: End

The model developed is validated with real-time dataset for better understanding
of functions of model. This helped to identify if retraining is needed to increase efficiency
of model.

16
 Chapter 5
Implementation

This chapter describes about the predicted outputs and the results obtained by
using the developed model.

5.1 Experimentation
The trained model integrated with android application is tested with 40 apples and
26 bananas. As shown in figure 5.1, opening the application automatically opens mobile
camera and detects the class of fruit placed in front of it. Each fruit is placed before
the application in different angles and the resultant class predicted with its accuracy is
observed.

Figure 5.1: Grading fruit

17
 Figure 5.2: Real Time apple images used for testing

5.2 Outputs
The output of our app which we developed are attached. Figure 5.2 shows that
the application detects the fruit as fresh banana with 99% probability because it is fully
yellow in color. Figure 5.3 shows the images of application detecting fruit as fresh apple
with 95% probability. Figure 5.4 shows the images of application detecting fruit as ripe
banana with 100% probability due to its features of having black colour. Figure 5.5 shows
the images of application detecting fruit as fresh apple. Figure 5.6 detects the fruit as
damaged apple.

18
Figure 5.3: Fresh Banana

Figure 5.4: Fresh apple

19
Figure 5.5: Ripe banana

Figure 5.6: Rotten Apple

20
 Figure 5.7: Fresh apple

5.3 Results and Analysis

The Fruit Grading application uses an CNN model trained on Apple and Banana
dataset and depicted an accuracy of 93.7%. The Figure 5.6 shows plot Accuracy over
Epochs for training and validation data over fruits dataset and figure 5.7 shows Plot Loss
over Epochs for training and validation data. Table 5.1 shows the banana and detected
accuracy for freshness. Table 5.2 shows apple and detected accuracy for freshness. The
overall accuracy for 26 bananas is 86%. The overall accuracy for 40 apples is 89.6%.

21
Table 5.1: Banana and detected accuracy

Banana Accuracy
Fruit 1 40
Fruit 2 98
Fruit 3 95
Fruit 4 62
Fruit 5 76
Fruit 6 99
Fruit 7 100
Fruit 8 85
Fruit 9 99
Fruit 10 98
Fruit 11 97
Fruit 12 96
Fruit 13 97
Fruit 14 70
Fruit 15 86
Fruit 16 75
Fruit 17 96
Fruit 18 100
Fruit 19 82
Fruit 20 92
Fruit 21 98
Fruit 22 97
Fruit 23 98
Fruit 24 52
Fruit 25 59
Fruit 26 89
AVERAGE 86

22
Table 5.2: Apple and detected accuracy

Apple Accuracy
Fruit 1 99
Fruit 2 99
Fruit 3 99
Fruit 4 96
Fruit 5 100
Fruit 6 100
Fruit 7 100
Fruit 8 100
Fruit 9 99
Fruit 10 100
Fruit 11 99
Fruit 12 92
Fruit 13 71
Fruit 14 92
Fruit 15 54
Fruit 16 84
Fruit 17 95
Fruit 18 50
Fruit 19 100
Fruit 20 85
Fruit 21 100
Fruit 22 100
Fruit 23 97
Fruit 24 93
Fruit 25 59
Fruit 26 55
Fruit 27 53
Fruit 28 100
Fruit 29 100
Fruit 30 100
Fruit 31 90
Fruit 32 100
Fruit 33 77
Fruit 34 95
Fruit 35 100
Fruit 36 97
Fruit 37 60
Fruit 38 100
Fruit 39 100
Fruit 40 92
AVERAGE 89.6

23
Figure 5.8: Training and validation Accuracy Plot

Figure 5.9: Training and validation Loss Plot

24
5.4 Client Satisfaction Report

Figure 5.10: Client Satisfaction Report

25
 Chapter 6
CONCLUSION AND FUTURE WORK

6.1 Conclusion
This project focusses on creating an automated application for grading of apple
and banana fruits using Convolutional Neural Networks which is an efficient deep learning
technique. It has been designed for fruit vendors who can assure the quality of fruits to
customers. This application uses a CNN model as backend support for classification
which is then converted to tflite model and integrated with android application. Based
on the features of the fruit, the trained deep learning model will classify the fruit into its
particular grade. The application worked fine on testing with real time dataset.

6.2 Future work

This application has ample of scope in future as Agriculture is one of the largest
economic sectors and due to high demand for quality food. Also, this project scope is
limited to apples and bananas only. One can develop methods for grading other fruits
and vegetables also.

26
 REFERENCES

[1] Matteoli, S., Diani, M., Massai, R., Corsini, G. and Remorini, D., 2015. A spectroscopy-
based approach for automated non-destructive maturity grading of peach fruits.
IEEE Sensors Journal, 15(10), pp.5455-5464.

[2] Keresztes, J.C., Diels, E., Goodarzi, M., Nguyen-Do-Trong, N., Goos, P., Nicolai,
B. and Saeys, W., 2017. Glare based apple sorting and iterative algorithm for
bruise region detection using shortwave infrared hyperspectral imaging. Postharvest
biology and technology, 130, pp.103-115.

[3] Van Zeebroeck, M., Ramon, H., De Baerdemaeker, J., Nicolaı̈, B.M. and Tijskens,
E., 2007. Impact damage of apples during transport and handling. Postharvest
biology and technology, 45(2), pp.157-167.

[4] Pan, X., Sun, L., Li, Y., Che, W., Ji, Y., Li, J., Li, J., Xie, X. and Xu, Y.,
2019. Non-destructive classification of apple bruising time based on visible and
near-infrared hyperspectral imaging. Journal of the Science of Food and Agriculture,
99(4), pp.1709-1718.

[5] Barrett, D.M., Beaulieu, J.C. and Shewfelt, R., 2010. Color, flavor, texture, and
nutritional quality of fresh-cut fruits and vegetables: desirable levels, instrumental
and sensory measurement, and the effects of processing. Critical reviews in food
science and nutrition, 50(5), pp.369-389.

[6] Bavay, C., Symoneaux, R., Maı̂tre, I., Kuznetsova, A., Brockhoff, P.B. and Mehi-
nagic, E., 2013. Importance of fruit variability in the assessment of apple quality
by sensory evaluation. Postharvest biology and technology, 77, pp6.67-74.

[7] Londhe, D.H., Nalawade, S.M., Pawar, G.S., Atkari, V.T. and Wandkar, S.V., 2013.
Grader: A review of different methods of grading for fruits and vegetables. Agri-
cultural Engineering International: CIGR Journal, 15(3), pp.217-230.

[8] Ahmadi, E., Ghassemzadeh, H.R., Sadeghi, M., Moghaddam, M. and Neshat, S.Z.,
2010. The effect of impact and fruit properties on the bruising of peach. Journal of
Food Engineering, 97(1), pp.110-117.

[9] Silva, V., Figueiredo, A.R., Costa, J.J. and Guiné, R.P.F., 2014. Experimental and
mathematical study of the discontinuous drying kinetics of pears. Journal of Food
Engineering, 134, pp.30-36.

[10] Magwaza, L.S. and Tesfay, S.Z., 2015. A review of destructive and non-destructive
methods for determining avocado fruit maturity. Food and bioprocess technology,
8(10), pp.1995-2011.

[11] Tang, Y., Gao, S., Zhuang, J., Hou, C., He, Y., Chu, X., Miao, A. and Luo, S.,
2020. Apple bruise grading using piecewise nonlinear curve fitting for hyperspectral
imaging data. IEEE Access, 8, pp.147494-147506.

27
[12] Chopra, H., Singh, H., Bamrah, M.S., Mahbubani, F., Verma, A., Hooda, N., Rana,
P.S., Singla, R.K. and Singh, A.K., 2021. Efficient fruit grading system using spec-
trophotometry and machine learning approaches. IEEE Sensors Journal, 21(14),
pp.16162-16169.

[13] Al Ohali, Y., 2011. Computer vision based date fruit grading system: Design and
implementation. Journal of King Saud University-Computer and Information Sci-
ences, 23(1), pp.29-36.

[14] Nandi, C.S., Tudu, B. and Koley, C., 2016. A machine vision technique for grading
of harvested mangoes based on maturity and quality. IEEE sensors Journal, 16(16),
pp.6387-6396.

[15] Ponce, J.M., Aquino, A., Millan, B. and Andújar, J.M., 2019. Automatic counting
and individual size and mass estimation of olive-fruits through computer vision
techniques. IEEE Access, 7, pp.59451-59465.

[16] Azarmdel, H., Jahanbakhshi, A., Mohtasebi, S.S. and Muñoz, A.R., 2020. Evalu-
ation of image processing technique as an expert system in mulberry fruit grading
based on ripeness level using artificial neural networks (ANNs) and support vector
machine (SVM). Postharvest Biology and Technology, 166, p.111201.

[17] Lee, D.J., Archibald, J.K. and Xiong, G., 2010. Rapid color grading for fruit quality
evaluation using direct color mapping. IEEE transactions on automation science and
engineering, 8(2), pp.292-302.

[18] Gurubelli, Y., Ramanathan, M. and Ponnusamy, P., 2019. Fractional fuzzy 2DLDA
approach for pomegranate fruit grade classification. Computers and Electronics in
Agriculture, 162, pp.95-105.

[19] Sucipto, A., Zyen, A.K., Wahono, B.B., Tamrin, T., Mulyo, H. and Ali, R.R., 2021,
September. Linear Discriminant Analysis for Apples Fruit Variety Based on Color
Feature Extraction. In 2021 International Seminar on Application for Technology
of Information and Communication (iSemantic) (pp. 184-189). IEEE.

[20] Wan, P., Toudeshki, A., Tan, H. and Ehsani, R., 2018. A methodology for fresh
tomato maturity detection using computer vision. Computers and electronics in
agriculture, 146, pp.43-50.

[21] Peng, H., Shao, Y., Chen, K., Deng, Y. and Xue, C., 2018. Research on multi-class
fruits recognition based on machine vision and SVM. IFAC-PapersOnLine, 51(17),
pp.817-821.

[22] Mizushima, A. and Lu, R., 2013. An image segmentation method for apple sorting
and grading using support vector machine and Otsu’s method. Computers and
electronics in agriculture, 94, pp.29-37.

[23] Sun, C., Liu, T., Ji, C., Jiang, M., Tian, T., Guo, D., Wang, L., Chen, Y. and Liang,
X., 2014. Evaluation and analysis the chalkiness of connected rice kernels based on
image processing technology and support vector machine. Journal of Cereal Science,
60(2), pp.426-432.

28
[24] Naik, S. and Patel, B., 2017. Machine vision based fruit classification and grading-a
review. International Journal of Computer Applications, 170(9), pp.22-34.

[25] Chandrasekaran, I., Panigrahi, S.S., Ravikanth, L. and Singh, C.B., 2019. Potential
of near-infrared (NIR) spectroscopy and hyperspectral imaging for quality and safety
assessment of fruits: An overview. Food Analytical Methods, 12(11), pp.2438-2458.

[26] Pawar, M.M. and Deshpande, M.M., 2012, November. Skin defect detection of
pomegranates using color texture features and DWT. In 2012 NATIONAL CON-
FERENCE ON COMPUTING AND COMMUNICATION SYSTEMS (pp. 1-5).
IEEE.

[27] Bai, G., Ge, Y., Hussain, W., Baenziger, P.S. and Graef, G., 2016. A multi-sensor
system for high throughput field phenotyping in soybean and wheat breeding. Com-
puters and Electronics in Agriculture, 128, pp.181-192.

[28] Huang, W., Zhang, C. and Zhang, B., 2011, October. Identifying apple surface
defects based on Gabor features and SVM using machine vision. In International
Conference on Computer and Computing Technologies in Agriculture (pp. 343-350).
Springer, Berlin, Heidelberg.

[29] Zhu, B., Jiang, L., Luo, Y. and Tao, Y., 2007. Gabor feature-based apple quality
inspection using kernel principal component analysis. Journal of Food Engineering,
81(4), pp.741-749.

[30] Al Ohali, Y., 2011. Computer vision based date fruit grading system: Design and
implementation. Journal of King Saud University-Computer and Information Sci-
ences, 23(1), pp.29-36.Lee,

[31] Lee, D.J., Archibald, J.K. and Xiong, G., 2010. Rapid color grading for fruit quality
evaluation using direct color mapping. IEEE transactions on automation science and
engineering, 8(2), pp.292-302..

[32] Matteoli, S., Diani, M., Massai, R., Corsini, G. and Remorini, D., 2015. A spectroscopy-
based approach for automated nondestructive maturity grading of peach fruits.
IEEE Sensors Journal, 15(10), pp.5455-5464.

[33] Nandi, C.S., Tudu, B. and Koley, C., 2016. A machine vision technique for grading
of harvested mangoes based on maturity and quality. IEEE sensors Journal, 16(16),
pp.6387-6396..

[34] Leekul, P., Chivapreecha, S., Phongcharoenpanich, C. and Krairiksh, M., 2016.
Rician k-factors-based sensor for fruit classification by maturity stage. IEEE Sensors
Journal, 16(17), pp.6559-6565.

[35] Zhen, Y.A.N., ZHENG, L.J., NIE, J.Y., LI, Z.X. and Cheng, Y., 2018. Evaluation
indices of sour flavor for apple fruit and grading standards. Journal of integrative
agriculture, 17(5), pp.994-1002.

[36] Gurubelli, Y., Ramanathan, M. and Ponnusamy, P., 2019. Fractional fuzzy 2DLDA
approach for pomegranate fruit grade classification. Computers and Electronics in
Agriculture, 162, pp.95-105.

29
[37] Ponce, J.M., Aquino, A., Millan, B. and Andújar, J.M., 2019. Automatic counting
and individual size and mass estimation of olive-fruits through computer vision
techniques. IEEE Access, 7, pp.59451-59465.

[38] Ren, A., Zahid, A., Zoha, A., Shah, S.A., Imran, M.A., Alomainy, A. and Abbasi,
Q.H., 2019. Machine learning driven approach towards the quality assessment of
fresh fruits using non-invasive sensing. IEEE Sensors Journal, 20(4), pp.2075-2083.

[39] Azarmdel, H., Jahanbakhshi, A., Mohtasebi, S.S. and Muñoz, A.R., 2020. Evalu-
ation of image processing technique as an expert system in mulberry fruit grading
based on ripeness level using artificial neural networks (ANNs) and support vector
machine (SVM). Postharvest Biology and Technology, 166, p.111201.

[40] Tang, Y., Gao, S., Zhuang, J., Hou, C., He, Y., Chu, X., Miao, A. and Luo, S.,
2020. Apple bruise grading using piecewise nonlinear curve fitting for hyperspectral
imaging data. IEEE Access, 8, pp.147494-147506.

[41] Chopra, H., Singh, H., Bamrah, M.S., Mahbubani, F., Verma, A., Hooda, N., Rana,
P.S., Singla, R.K. and Singh, A.K., 2021. Efficient fruit grading system using spec-
trophotometry and machine learning approaches. IEEE Sensors Journal, 21(14),
pp.16162-16169.

30