Nat. Volatiles & Essent.

Oils, 2021; 8(5): 11981-11992

Crop Yield Prediction Using Image Processing

S G Shiva Prasad Yadav1*, Rohan Reddy N2, Dharun D3, Niveditha A4, and Hema N5
1*,2,3,4,5
Electronics and Telecommunication Engineering MS Ramaiah Institute of Technology Bangalore, India

Abstract

Agriculture is a major Indian occupation. It Plays a vital role contributing to over 18% of India's GDP and provides jobs to 50% of the
population of India Population growth is a major food security challenge. Population growth is increasing the need for farmers to
produce more in the same agricultural country in terms of increasing supply. Computing the yield of any fruit or a vegetable takes a
lot of computation time and requires too much of processing work. Technology can help farmers to produce more with the help of
forecasting. The proposed model assists farmers in predicting crop yields. Fruit images are obtained with integrated charging
devices (CCD) and the background of each image will be removed using a different algorithm to separate the fruit area in the input
image. Fruits are calculated based on the centroid of the fruit. A machine-based approach is used to filter fruit by predicting the
degree of maturity and aims to change the staffing system. The program includes pre-image processing, feature extraction and fruit
classification using novel computational techniques such Machine Learning, Artificial intelligent, deep learning and so on. Using K
means the highest accuracy of the fruit number algorithm is available. The project presents a computerized concept and techniques
for machine learning of tree fruit acquisition, counting and sorting. With the introduction of IoT to use image processing the project
assists farmers in predicting the yield. The findings can be stored in a database that enables the farmer to move whenever needed
and keep track of the yield.

Keywords: Crop Yield Prediction, Image Processing, Predictive analysis

Introduction
The overall world's population is estimated to be reaching around 10 billion by 2050, which increases the
agricultural order in the context of local economic development by 50% compared to 2013. Currently,
about 37.7% of the total land is used for crop production. From job creation to National Revenue,
agriculture is important. It plays a vital role in the economic prosperity of developed countries and plays a
vital role in the economy of developing nations. Agriculture, and its affiliates, undoubtedly provide the
largest livelihoods in India, especially in large rural areas. It also contributes to a significant product in GDP.
The need of sustainable agriculture, is very important in terms of food security, employment opportunities
in rural areas, and environmental technologies such as soil conservation, sustainable natural resource
utilization and management and environmental protection, are essential to the overall development of
rural areas. Indian agriculture and co-operatives have seen many revolutions such as green, white
revolution, yellow and a blue revolution. Agriculture is also the backbone of the Indian economy. The local
environment of our country has been very productive in agricultural activities.
The physical characteristics that exist in India, namely, his climate, his soil and his ease have greatly helped
him to grow many crops here. From time immemorial, Indians have regarded agriculture as a means of
subsistence. Agriculture is the process of producing food, feed, fiber and many other products that you
want by growing certain crops and raising pets. The main function of Agriculture is to provide food for the
people, to provide for industrial inputs, to employ large numbers of people, to finance agricultural
development, the remnants of the country's economy if done well. India has been around for so long and is
still an agricultural country that plays a key role in holding it accountable for the overall performance of our
country. From the point of view of the contribution of agriculture to the Gross Domestic Product (GDP) of
our country, such as food industry, as the livelihood of the majority of our people, as a supplier of wages
and other agricultural value to our country's economy cannot be exaggerated. Apple having the scientific
name Malus pumila is one of the prominent fruit and stands in the fourth position in terms of most

11981
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

productive fruit in the world after bananas, oranges and grapes. China is the world's largest producer of
apples. In India, the area under cultivation of apples has increased by 24% from 1.95 lakh hectare in 1991-
92 to 2.42 lakh hectares in 2001-02 although production has increased by less than 1% (example from 11
tons to 12 lakh) [1]. This apple is widely grown in few states such as Jammu & Kashmir, Himachal Pradesh,
Uttaranchal, Arunachal Pradesh and Nagaland regions. An agricultural exporter has been setup in Himachal
Pradesh, which includes the Shimla, Siramour, Kullu, Mandi, Chamba and Kinnaur regions, with significant
potential to increase exports.
The initial objectives are to export apples to neighboring countries and western Asia as well as countries in
Southeast Asia. With the increase in the yield of apples in the State, the Government is taking useful and
efficient measures to market its products without providing packaging, transportation, to ensure
appropriate supply of apples under the Market intercession scheme, so that farmers get paid prices in the
market. Smart farming based on IoT technology will help the farmers and agriculturists to reduce waste and
improve yield by reducing the amount of fertilizer used, to the amount of travel produced by farm vehicles,
and allows for efficient use of resources such as water, electricity, etc. Farming solutions are a system
designed to monitor the field of crops with the help of sensors. IoT enabled applications for smart
agriculture focus on helping farmers by ensuring high yields, profitability and environmental protection.
The usage of IoT technology to ensure efficient use of all the resources to achieve higher crop yields by
reducing the operational and managerial costs is called precision Agriculture. IoT in agricultural technology
contains specialized devices, wireless communication devices, software and IT enabled services. There has
been a very significant contribution of agriculture in the Indian economy and it is being improving over
several years. According to the economic survey, the share of agriculture in gross domestic product (GDP)
reached about 20% for the first time in the last 17 years making a significant contribution to GDP
performance for the 2020-2021 financial year. Modern farm and agricultural activities have changed over
the years as a result of technological advances, including sensors, devices, machinery, and information
technology.

Block Diagram of the proposed system

The block diagram for the crop yield detection is shown in Figure 1 and the design is divided into two
phases – Front end and back end. In the front end, a GUI is built consisting of buttons such as refresh and
detect in it.
Figure 1. Block Diagram of the crop yield detection

At the backend on clicking the detect button, the software will run and the crops will be detected and an
estimation result will be given at the end. The Graphical User Interface for the proposed system of yeild
estimation is shown in figure 2.

11982

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

Figure 2. Block Diagram of GUI

In this GUI the frames of an input video are loaded. The feature buttons available are the refresh and the
detect buttons. The refresh button loads the video from beginning. On clicking the detect button,
The buttons can be created as per the wish of the developer. The GUI is developed using a Python’s GUI
module. The classic GUI library for Python used is Tkinter. Python when combined with Tkinter supplies a
rapid and simple way to make GUI applications. Tkinter supplies a considerable object-oriented interface to
the Tk GUI toolkit.
Tkinter is very user friendly and using Tkinter, creating a GUI application is very simple. Tkinter provides
various controls, such as buttons, labels and text boxes used in a
GUI application. These controls are commonly called widgets. All Tkinter widgets have access to specific
geometry management methods, which have the purpose of organizing widgets throughout the parent
widget area. Tkinter exposes the following geometry manager classes: pack, grid, and place.
The following steps are performed for the implementation:

• The Tkinter module should me imported.

• Next the GUI application should be created.
• We add one or more of the above-mentioned widgets to the
• GUI application.

• Enter the main event loop to hold action against each event triggered by the user.
The overall blocks needed for the implementation of yield estimation is shown in the Figure 3. The
individual blocks and their functionality are illustrated below.
Figure 3. Block Diagram of the Proposed Model

The image acquisition part for the proposed system is illustrated in the section below.

• Image Acquisition: The image is acquired by means of digital color charged coupled device (CCD)
camera. Images are captured in various weather natural lighting conditions such as full sun- front
lighting and back lighting, full sun- fruits in the shadow and in cloudy conditions. Same tree

11983

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

image is acquired at various angle of aperture. All the collected images transferred into
computer through USB port. By collection of image a big dataset is prepare and this dataset is
use as input to train the machine. Charge-coupled device (CCD) which is use to capture image of
tree fruits includes an image sensor, and it senses the amount of light rays strike on the area of
the CCD sensor and amount of light converted into an electric signal. CCD image sensor output is
in an analog format.
• Image Processing: In this is called image processing phase, once the image is acquired some
preprocessing has to be performed to modify the image into suitable form and set of rules is
applied to train and test image. Digital image is displayed on screen of computer as pictures,
which are understandable by the computer, and that images are changed into tiny dots.
Grayscale image processing is used to detect fruit from image. Color image processing is used to
obtain features of fruits.
• Image Segmentation: The process of cutting of background, adding and extract common feature
analysis of image focused at separating an input image into area that has strong relation with
objects or areas of interest has called as Image segmentation. In image segmentation, removed
data mostly depend on the exactness of the operation. If objects in image cannot be distributed
accurately, it is difficult to understand for object detection, measurement and classification.
Hence, it impacts on interpreting and understanding that image.
For segmentation part to be implemented, initially background leaves, branches are eliminated. This is
done by using a masking and overlying operations on image. K-Means algorithm, one of the most reliable
and efficient, is used to segment the input image.
The segmentation of individual section of the image involves the alignment together of alike pixels to form
areas signifying single objects within the image. Hue saturation value (HSV) color space transformation of
RGB input image provides better segmentation result with K-means clustering algorithm.
• Polygon Fitting: Overlying of circles has to be performing to get the required boundaries of fruit
region. Circle fitting algorithm is use to fit circle into scattered pixel data. First, the co-ordinates
of the edge pixels have to found. Then the mean of pixels has to calculate. Then the center of the
clusters by using the mean has computed. With the help of circle, overlying it is possible to get
perfectly segmented required fruit region.

Implementation of the Crop Yield Detection

The flowchart comprising of various steps for the implementation of the crop yield detection are shown in
figure 3. The acquired image is preprocessed for size adjustment and converted in HSV format. Based on
the threshold fixed, the acquired imaged is compared against the threshold and background subtraction is
carried out. The object detection is done based on the comparison against the threshold set. Then next part
of the algorithm would be to get the coordinates or position of the object so that the counting of the
objects can be done based on the contour of the object. Once the counting of the identified objects are
done, the sms message is sent to the user and a database will be updated. This database will have the
overall count of the identified objects and will provide the yield of the crop.
Figure 4. Implementation of the crop yield detection

11984

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

Tkinter is a very usefl and a standard GUI library for Programming in python. The combination of Python
with Tkinter provides a more efficient, faster and easy way to create GUI applications. The program is
mainly divided into two windows: The first window consists of:
Step1: Input image source is taken as video or from web camera.
Step2: The obtained input image is converted from RGB to HSV format.
Step3: After conversion the next step is to find the size and shape of the fruit.
Step4: The color of interest lower bound and higher bound are chosen Step5: With the help of lower bound
and higher bound filter is created
Step 6: The region of color of interest is counted.
Step 7: All the filter images and count of fruits image on laptop screen is plotted.
Step 8: Send SMS with all the counts of the fruits at the end of the run
Step 9: Upload all the counted fruits data to firebase cloud for future reference.
The second window consists of:
Tinker is the python library to create GUI
Step1: an interactive GUI for the project is made from tkinter
Step2: click on apple button to choose image for processing
Step3: backend code final.py is called.
Step 4: final. py gives the count, shape and size of the fruit s which is later show on the GUI

Literature Review
In this section, various methods of acquisition are discussed. This includes interviewing and testing of
various research projects that can be used to make our software or serve as a basis on which the proposed
software can be built.

11985

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

Nigam A et al. [5]: This survey focuses on enhacing chop making through several machine acquiring
techniques. Tests are performed on agricultural databases. Random Forest acquiring variants, XGBoost,
Logistic Regression, Linear Regression, and ANN have been utilized to gain a better detector to yield correct
prediction. These algorithmic programs are utilized in Python 3.7 utilizing the various libraries built: Numpy,
Sikik-read, Pandas and Keras. The needed hardware configuration is 8GB RAM with Intel i5 processor. This
money present several acquiring algorithmic program for machine crop prediction as attested by various
factors. Tests have been performed on the Indian authority database and is found that the Random Forest
Regressor supplies accuracy considered to be highest of the produce forecast. The without a break Simple
Recurrent Neural Network exemplar perform greater in predicting rainfall. LSTM is ready forecasting of
temperature.
Thomas van Klompenburga et al. [6]: Mechanical learning is an important decision-making tool for
predicting crop yields, including supporting decisions on where plants should be grown and what to do
during crop growth. Various algorithms are used to support prediction of crop yields. They have developed
a Systematic Literature Review (SLR) to extract and integrate algorithms and features used in yield
prediction studies. According to the search criteria, 567 relevant subjects from six electronic profiles, of
which 50 subjects were selected for further analysis using inclusion and exclusion criteria. They carefully
researched the selected studies, analyzed the methods and features used, and made suggestions for
further research. According to their analysis, the most commonly used features are heat, rain, and soil type,
and the most widely used algorithm is Artificial Neural Networks for these types. After this observation
based on the analysis of 50 machine-readable papers, they conducted additional searches in the electronic
information archives to identify in-depth reading lessons, reached up to 30 study-based papers, and
produced in-depth reading algorithms. According to this additional analysis, Convolutional Neural Networks
(CNN) is the most widely used algorithm for learning these subjects.
Mishra S et al. [7]: Performing the statistics mining strategies based on climate and crop fabrication
statistics various predictions can be composed as attested by knowledge collected where twist is assisted to
increase productivity of crops. DSS will have to be implemented to help farmers for preventing the
overheads of decisions regarding soil as well as crop to be cultivated. An application scheme DSS is used
which provides assistance for analysts to foretell and identify databases from a raw dataset or models of
commerce to analyze a plight and get results by fabrication decisions. Scheme used here will assist people
to make necessary changes which were earlier used in utilizing inefficient insignificant methods.
Akash Ashapure et al. [8]: Current duration have observed immense rise in not only Unmanned Aircraft
Entity (UAS) but also sensor tech which made it possible to derive tall temporal and spatial resolutions
information more than crops growing all season. Unbiased research was to design a machine acquiring
framework for tomato which is marketable and surrender abdicates estimation utilizing multiple source
and spatio-temporal secluded sensing databases. This proposed machine learning examples as said by ANN
and it takes UAS based multi-temporal options, excessive greenness Index besides weather databases that
is humidity, precipitation, temperature, etc. as input and predicts yield. Moreover, produce prediction
maps can be utilized to optimize field management practices.
Medar, R et al. [9]: Naive Bayes system and K-Nearest neighbour method. Accuracy can be obtained by
utilizing these two methods. To foretell the yield surrender abdicate rate a java application is created. This
application can add on three parts. 1st dealing with datasets, 2nd testing those datasets and thirdly to
analyze them. Two methods of ML have been considered as told above. During testing any part of the
process can be chosen and perform testing like by choosing specific plant, specific area or season results of

11986

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

yields can be obtained. During analyzing, we can give a complete dataset and accuracy of these two
methods can be obtained. This assists in predicting the best process.
Shiv Ram Dubey et al. [10]: Lately, commerce has doubled across the globe. Plenty of fruits are imported.
Manual identification involves consumption of time. This labor presents a segmentation of fruits with K-
means clustering unsupervised algorithm. The colours are utilized for deficiency segmentation. Deficiency
segmentation is conducted in two separate stages. The pixels are clustered by their color and spatial
features. Utilizing this procedure, it's possible to enhance the algorithmic effectiveness bypassing option
extraction for any pixel. But the color is not mainly utilized for imperfection segmentation, it generates a
lofty discriminating power for diverse regions of image. The simulation results explain that the proposed
advance is promising.
Inkyu Sa et al. [11]: This notes advances to detection of fruit utilizing deep CNN. It’s aim is to produce an
accurate, rapid and dependable fruit detection system, which is a crucial element of an autonomous
agricultural robotic platform; Current labor in deep neural networks has led to the development of a state-
of-the-art thing detector termed Faster Region-based CNN (Faster R-CNN). This model is adapted through
transfer learning for detection utilizing RGB and NIR. Before time and tardy fusion methods are explored for
combining the multi-modal (RGB and NIR) information. Moreover , this advance is additionally much
quicker as it needs bounding box annotation. The exemplar is retrained to compose the detect several
fruits.
Muhammad Ayaz et al. [12]: This article provides an overview of wireless sensors and IoT applications in
the field of agriculture. IoT devices and intercommunication strategies connected with various wireless
sensors came across in agriculture applications and have undergone detailed analysis. To enhance the
efficiency and make it more effective, the unmanned aerial vehicles are incorporated for crop surveillance
and other profitable applications. New emerging and state-of-the-art architectures with IOT enabled and
platforms are utilized in agriculture. Finally, they identify common and coming day’s patterns of IoT and
highlight potential research challenges in the field of agriculture.
D. Aggelopoulou et al. [13]: This study conducted in Central Greece recommends that apple flowering maps
are utilized for management decisions. In April 2007, when the plants are about to bloom, photos of these
plants were taken using random sampling method. In September 2007, surrender renounce mapping was
conducted measuring surrender renounce per ten trees and this situation was recorded. Utilizing these
statistics a vision based algorithm was designed to predict the tree surrender renounce by analyzing the
characteristics of the tree at rotund bloom. The rebases results were presented for the evaluation, where
the fault of the design predicted yield was set at 18%. The results of this work signifies that potential yield
can be predicted before time in any given season.

Experimental Results and Discussion

The software provides facilities like importing the video from a designated file location. Detect and refresh
are the operational buttons which are present on the GUI as shown in the figure a. Features like the shape,
colour and the size of the detected fruit will be will be given after analyzing the complete video input as
shown in the figure 5.
Figure 5. GUI Displaying the Output

11987

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

• Image is held as input from dataset of apple fruit which is accessible either online or offline video
by capturing an offline video of the apple fruit. The images acquired by digital camera can also be
considered as input. Daytime image is considered for the implementation and the pre-processing
of the acquired images is done by adopting a few computer vision based OpenCV libraries in
python which can be imported such as numpy, scipy, scikit, matplotlib and so on. Then the K-
means clustering algorithm is applied to identify the fruits which is as shown in the figure 6.
Additionally, a threshold is set up so as to avoid the detection of the same fruit multiple times as
shown in the figure 6.
Figure 6. Detection of Apples

• After analyzing the input in the form of the video using the algorithm and the mentioned libraries,
for every apple detected, the co-ordinates, the shape and the color will be mentioned individually
as shown in the figure 7. On completion of the given input, the total number of apples will be
given along with its shape, size and also the color as well as the date of the operation being
performed as shown in figure 7.
Figure 7. Yield of Apples as Output

11988

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

• As mentioned in the above discussion, the total count of the apples detected using the K-means
clustering algorithm will also be stored in the database along with the date and time of the
performed operation as shown in figure 8. The same data of the crop yield will also be sent as a
short message service (SMS) to the farmer or the person in charge of the field as shown in figure
9.
Figure 8. Data stored in Cloud

Figure 9. SMS Alert

11989

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

• Once the fruit is recognized, the algorithm continues to detect the other fruits of the particular
color as long the input video frames are running. Subsequently the counts of the number of fruits
present in these frames are displayed and the data is also sent to be stored in the cloud. The
model’s output is considered to be an error value if the difference between predicted center and
the actual centre of the player is more than 5 pixels. All the parameters of the software and its
performance are listed in table 1:

Table 1: Statistics of developed model and data

• For the prediction of the apple count using k means clustering, 10 test cases were considered with
a range of apples from 0-25. Each test case presents with accuracy according to the number of
apples in that particular test case. As the number of apples varies in each test case, the accuracy
varies accordingly. As mentioned in the above table 1 the accuracy obtained is 89% and the plot
for various test cases vs. the apple count is shown in the below figure 10.
Figure 10. Calibrated crop yield vs. the actual count of apple block

11990

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

Conclusion and Future Work

The module and usable yield prediction system is designed to use RGB and K means algorithm. This project
is basically a farmer can analyze agricultural data predictions to predict the production of crop yield
forecasts using image processing. The basics of computer vision are used to trace three colors red and blue
through this easy to see fruit. The Farmer can therefore keep track of the yield obtained. However, this task
can be very difficult and challenging as you need a lot of relevant data to produce the right result. Data
classification and data storage according to their properties and characteristics is done with k-means
clustering algorithm. This provides an exciting outcome that can help farmers improve their crop
production. The project concludes with the recent construction of computer vision using the OpenCv
platform and machine learning algorithms in the agricultural industry especially in the fruit industry. Image
imaging is used to accurately identify each fruit, in a plant using a standard RGB digital camera in
combination with machine learning capabilities. Acquisition of tree fruit image acquisition, image
processing including image enhancement, image extraction and image element separation followed by
image fragmentation, polygon equation, fruit detection and calculation is done within 2sec as soon as the
input is inserted. The advanced method does not require an increase and decrease in the threshold values
for each apple image to be identified because image classification is done based on computer visualization
OpenCv, numpy, scipy, scikit, matplotlib libraries use the adoption of this new technology in identification,
calculation and fruit planning will be of great benefit to the farmer.

Future Scope of The Work

Currently, the entire harvesting process is only performed on software with minimal hardware usage. The
project can be developed in a variety of ways using the methods discussed below:
• A project that uses a specific algorithm can be trained to get low quality video.
• The project can also be upgraded where the system should be able to get it right at high installation
speed.
• With the help of hardware, various types of data can be extracted and can be used for other analytical
purposes. • The project can be trained to find scattered apples.
• The software used is not limited to apples as shown above, but can be used on other fruits as well.
• The project can be used in real-time programs such as UAVs.
• The project can be done with the help of advanced sensors and GPS modules to locate the deviant and
rehabilitate the farmer.

REFERENCES

11991

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions
Nat. Volatiles & Essent. Oils, 2021; 8(5): 11981-11992

http://nhb.gov.in/report_files/apple/APPLE.htm

References on Wikipedia: https://en.wikipedia.org/wiki/Agriculture_in_India

https://en.wikipedia.org/wiki/List_of_countries_by_apple_production

https://vikaspedia.in/agriculture/crop-production/package-of-
practices/fruits1/apple1/apple#:~:text=In%20India%2C%20Apple%20is%20primarily,%3B%20Nagaland%3B%20Punjab
%20and%20Sikkim.

Article written by Economic Times on “Why India should start taking apples seriously”
https://economictimes.indiatimes.com/news/economy/agriculture/whyindia-should-start-taking-apples-
seriously/articleshow/71767025.cms?from=mdr

Nigam, A., Garg, S., Agrawal, A., and Agrawal, P. (2019). Crop Yield Prediction Using Machine Learning Algorithms.
2019 Fifth International Conference on Image Information Processing (ICIIP).

Thomas van Klompenburga , Ayalew Kassahuna and Cagatay Catal – “Crop yield prediction using machine learning”.

Mishra, S., Paygude, P., Chaudhary, S., and Idate, S. (2018). Use of data mining in crop yield prediction. 2018 2nd
International Conference on Inventive Systems and Control (ICISC).

Akash Ashapure, Sungchan Oh, Thiago G. Marconi, Anjin Chang, Jinha Jung, Juan Landivar and Juan Enciso-
“Unmanned aerial system based tomato yield estimation using machine learning”

Medar, R., Rajpurohit, V. S., and Shweta, S. (2019). Crop Yield Prediction using Machine Learning Techniques.2019 IEEE
5th International Conference for Convergence in Technology (I2CT). Department of Telecommunication Engineering,
RIT

Shiv Ram Dubey, Pushkar Dixit, Nishant Singh and Jay Prakash Gupta – “Infected Fruit Part Detection using K-Means
Clustering Segmentation Technique”

Inkyu Sa, Zongyuan Ge, Feras Dayoub, Ben Upcroft, Tristan Perez and Chris McCool –DeepFruits: “A Fruit Detection
System Using Deep Neural Networks”

Muhammad Ayaz, Mohammad Ammad-uddin, Zubair Sharif, Ali Mansour, and el-Hadi M. Aggoune – “Internet-of-
Things (IoT) based Smart Agriculture: Towards Making the Fields Talk”

A. D. Aggelopoulou , D. Bochtis , S. Fountas , K. C. Swain , T. A. Gemtos and G. D. Nanos- “Yield prediction in apple
orchards based on image processing”.

Petteri Nevavuorib, NathanielNarraa and TarmoLippinga- “Crop yield prediction with deep convolution neural
network”

Pinaki Mondal and Manisha Basu – “Adoption of precision agriculture technologies in India and in some developing
countries: Scope, present status and strategies”. Accepted 24 July 2008

Yemeserach Mekonnen, Srikanth Namuduri, Lamar Burton, Arif I. Sarwat and Shekhar Bhansali- “Machine Learning
Techniques in Wireless Sensor Network Based Precision Agriculture”. Published December 19, 2019.

11992

Downloaded from https://cabidigitallibrary.org by 2401:4900:2500:48c6:1:2:6ee5:6e41, on 11/14/24.

Subject to the CABI Digital Library Terms & Conditions, available at https://cabidigitallibrary.org/terms-and-conditions