IRJET- Automated Irrigation using IoT and Plant Disease Detection using Image Processing and Machine Learning

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 05 | May 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 5799
Automated Irrigation using IoT and Plant Disease Detection using
Image Processing and Machine Learning
Pavan Nataraj1, Praveen V Mugandamath2, Amit Vikram R3, Nithin Kumar A4
1,2,3,4Department of Computer Science and Engineering, Jyothy Institute of Technology, Bangalore, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract:- India is the largest freshwater user in the world.
86% of water is used for agriculture, 5% for industry and the
remaining 8% fordomesticpurpose. Waterplaysanimportant
role in plant lifecycle. India is mainly an agricultural country.
Irrigation is a vital component of agriculturalproduction. The
irrigation system can be classified as either manual or
automatic. Compared to manual irrigation, the automated
irrigation system can save water and maximize productivity.
This method may sometimes lead to over or under irrigation.
Manual irrigation takes a lot of time and effort. In automated
irrigation water is supplied only when it is required with
minimal or no human intervention. With the invent of plant
diseases, the yield is affected adversely. Hence it is important
to identify the disease at its earliest stages and find a cure to
eradicate the disease.
Key Words: – Internet Of Things(IoT), Digital Image
Processing, KNN algorithm, Otsu Method.
1. INTRODUCTION
The main aim of this project was to provide water to the
plants or gardening automatically using a microcontroller
(Arduino Uno). We can automatically water the plantswhen
we are going on vacation or don’t we have to bother my
neighbors, Sometimes the neighbors do too much watering
and the plants end up dying anyway. There are timer based
devices available in India which waters the soil onset
interval. Microcontrollers like Arduino are being used to
make the system to nearly completely automated. Arduino
like any other device is an open source platform which is
capable of interacting with theobjectsbysensingtheminthe
real world. The advantage of this little small electronic
device is that it is very easy to learn and Irrigation Station
with Supervised Learning using Artificial Intelligence
implement. Arduino can receive inputs from many sensors
and in turn can control the motors, solenoid valves. The
image processing techniques canbeusedintheplantdisease
detection. In most of the cases diseasesymptomsareseen on
the leaves, stem and fruit. The plant leaf for the detection of
disease is considered which shows the disease symptoms.
This paper gives the introduction to image processing
technique used for plant disease detection.
1.1 TYPES OF IRRIGATION
Surface Irrigation: Surface irrigation is the oldest form of
irrigation and has been in use for thousands of years. In
surface (furrow, flood, or level basin) irrigation systems,
water moves across the surface of anagricultural lands,inan
order to wet it and infiltrate into the soil. Surface irrigation
can be subdivided into furrow, border strip or basin
irrigation. It is often called flood irrigation when the
irrigation results in flooding or near flooding of the
cultivated land. Historically, this has been the most common
method of irrigating agricultural land and still used in most
parts of the world.
Micro Irrigation: Micro-irrigation, sometimes called
localized irrigation, low volume irrigation, or trickle
irrigation is a system where water is distributed under low
pressure through a piped network, in a pre-determined
pattern, and applied as a small discharge to each plant or
adjacent to it. Traditional drip irrigation using individual
emitters, subsurface drip irrigation (SDI), micro-spray or
micro-sprinkler irrigation, and mini-bubbler irrigation all
belong to this category of irrigation methods
Drip irrigation: Drip (or micro) irrigation, also known as
trickle irrigation, functions as its name suggests. In this
system waterfalls drop by drop just at the position of roots.
Water is delivered at or near the root zone of plants, drop by
drop. This method can be the most water-efficientmethod of
irrigation if managed properly, evaporation and runoff are
minimized. The field water efficiency of drip irrigation is
typically in the range of 80 to 90 percent when managed
correctly.
2. LITERATURE SURVEY
The author Ms.Sarika Rakshak et al., presents that the
Cultivation Management System mansion here is based on
cloud. The architecture of system allows user to achieve the
above mentioned activities in prearranged time so that
farmers can examine their farm field data details from
anywhere in between the range. Monitor system mainly
consist Hardware module that situated in farm or farm field
that has various sensors, devices,ICsfordata transformation
and transfer. Then Cloud implemented as Software as a
Services (SaaS) so that the Android smart phone used as a
remote control to make Arduino based automatedirrigation
system easy-to-use. The system design includes a soil
moisture sensor placed in different direction of farm field
that provides a voltage signal proportional to the moisture
content in the soil which is compared with a predefined
threshold value.[1]

The author Mr. Chandan Kumar sahu et al., presents a
prototype for automatic controllinga irrigationsystem.Here
prototypes includes sensor node and control node. The
sensor node is deployed in irrigation field for sensing soil
moisture value and the sensed data is sent to controller
node. On receiving sensor value the controllernodechecksit
with required soil moisture value. When soil moisture in
irrigation field is not up to the required level then the motor
is switched on to irrigate associated agriculture field and
alert message is send to registered mobile phone. The
experimental results show that the prototype is capable for
automatic controlling the experimental resultsshowthatthe
prototype is capable for automatic controlling of irrigation
motor based on the feedback of soil moisture sensor. This
system is used in a remote area and there are various
benefits for the farmers.[2]
The author Yunseop (James) Kim et al., presents An
automated closed-loop irrigation system requires three
major components: machine conversion, navigation, and
mission planning to support the solid communication
protocol. This paperdevelopedthe machineconversionfrom
a conventional irrigation system to an electronically
controllable system for individual control of irrigation
sprinklers and formulated the navigation of the irrigation
system that was continuously monitored by a differential
GPS and wirelessly transferred data toa basestationforsite-
specific irrigation control. This paper also provided
extensive details for the wirelesscommunicationinterfaceof
sensors from in-field sensor stations and for a
programmable logic controller from a control station to the
computer at a base station. Bluetooth wireless technology
used in this paper offered a plug-and-play communication
module and saved significant time and expense by using
commercially available sensors and controllers equipped
with serial communication ports.[3]
The author Sharada Prasanna Mohantyetal.,proposedUsing
Deep LearningforImage-BasedPlantDiseaseDetectionUsed
For In the following method, the training dataset containing
150 -170 images of mulberry plant leaves are analyzed and
tested against the test dataset having images of 20 to 30 of
the same leaf. The leaf is categorized into diseased and
normal leaves. The normal is categorized into 2 species of
the same plant namely red mulberry and white mulberry.
Using the deep convolutional neural network architecture,
the model is trained on images of plant leaves with the goal
of classifying both crop species and the presence and
identity of disease on images. Across all ourexperiments, we
use three different versions of the dataset. We start with the
dataset as it is in color. [4]
The author Sachin D. Khirade et al., proposed Plant Disease
Detection Using Image Processing system where the system
first acquires the image then image pre-processing is done.
Then image segmentation is performed, later feature
extraction in image is done which is then followed by
detection and classification of plant disease. The use of KNN
methods for classification of disease in plants such as self
organizing feature map, back propagation algorithm, SVMs
etc. can be efficiently used. From these methods, we can
accurately identify and classify various plant diseases using
image processing techniques. [5]
3. PROPOSED SYSTEM
A] Image Acquisition
The images of the plant leaf are captured through the
camera. This image is in RGB (Red, Green And Blue) form.
Color transformation structure for the RGB leaf image is
created, and then, a device-independent color space
transformation for
the color transformation structure is applied
B] Image Pre-processing
To remove noise in image or other object removal, different
pre-processing techniques is considered. Image clipping i.e.
cropping of the leaf image to get the interestedimageregion.
Image smoothing is done using the smoothing filter. Image
enhancement is carried out for increasing the contrast. the
RGB images into the grey images using color conversion
using equation.
f(x)=0.2989*R + 0.5870*G + 0.114.*B
Then the histogram equalization which distributes the
intensities of the images is applied on the image to enhance
the plant disease images. The cumulative distribution
function is used to distribute intensity values.
Fig 1 Basic Steps for Plant disease Detection and
Classification
C] Image Segmentation
Segmentation means partitioning of image into various part
of same features or having some similarity. The
segmentation can be done using various methods like otsu’
method, k-means clustering, converting RGB image into HIS
model etc. In Proposed system data is sensed through
sensors then KNN Machine learning algorithm is used for

data Prediction. Sensed data is compare with data set which
are stored on past experience, and resultisproduced. Asper
the predicted result farmer will take decision from this
system for profit gain. As shown in above scenario sensors
are deployed in farm which use are to sensed the data
related to humidity, temperature,sunlightandwindspeed.K
Nearest Neighbor algorithm is applied on sensed data to
classify and to form cluster. Clusters are then analyses with
predefined data set to generate the output. The predicted
result shows the whatever diseases can be cause due to
particular crop condition.
3.1 Technique
K Nearest Neighbor is one of the most powerful learning
models. They can have wide range of complex functions
whichrepresents multidimensional input-outputmaps. KNN
is also an information processing paradigmthatismotivated
by way biological nervous system, such as brain. KNN is
generally presented as system of interconnected "neurons"
which send message to each other. In proposed system the
MLP technique is used for data prediction. K Nearest
Neighbor are typically difficult to configure and slow to
train, but once prepared are very fast in application. With
image processing, first we need to pre-process the image
data i.e, the training data and then train the system using
KNN algorithm to predict whether the plant has a disease or
not.
Fig 2 Circuit Diagram
Initially the programmed code is fed into the micro
controller i.e. NodeMCU ESP 8266 and compiled to check if
there are any errors. If the code is successfully compiled
without error it can be uploaded into the ESP.TheNodeMCU
has an inbuilt in 4mb of flash memory and can store the
code. Once the code is uploaded into the ESP whenever it is
powered it runs in a loop. So, the code can be uploaded and
placed in the developed system and its runs with the power.
The power to all the sensors is given by the microcontroller
ESP, and the data that are obtained by the sensors from the
surrounding environment is sent to the cloud using Wi-Fi
and interface to the cloud is provided to the user through
which user can analyze and observe the data. The time inter
between the successive data that are sent to thecloudcanbe
set during the coding using the delay.
According to the requirement of the system to be designed
the threshold of the soil moisture is set. When the data sent
by the soil moisture sensorreachesthethresholdtheubidots
send the farmer with an email or message in local language.
Also, simultaneously when the data reaches the threshold
the water pump is triggered by the microcontroller using a
5v relay. The relay is always in normally connected state
where the circuit of the relay is always closed and works
when the power flows through it but it has to be triggered.
Ubidots is an IoT deployment platform where the data
received from the sensors are analyzed and viewed.
Different types of events like message and Email can be sent
for some threshold values of the sensor data. The admin or
the user uses ubidots via standalone system for viewing,
analyzing and also can control sensors. Statistics of these
sensor values can also be viewed and can also be monitored.
Advantages:-
1. A neural network can perform tasks whichlinearprogram
cannot.
2. It works even in the presence of noise with good quality
output.
3. Saves time and water.
4. water management and efficient use of water.
Disadvantages:-
1. Requires a lot of training and cases.
2. Often abused in cases where simpler solution like linear
regression would be best.
4. RESULT AND DISCUSSION
Fig 3 Contrast Enhanced

Contrast of the query image is enhanced to improve the
accuracy of the result. Contrast stretching is a simple image
enhancement technique that attempts to improve the
contrast in an image by `stretching' the range of intensity
values it contains to span a desired range of values
Fig 4 Segmented Images
Used K-Means clustering for segmentation and convert the
image from RGB Color Space to L*a*b* Color Space. The
L*a*b* space consists of a luminositylayer'L*',chromaticity-
layer 'a*' and 'b*'.All of the color informationisinthe'a*'and
'b*' layers. Classify the colors in a*b* color space using K
means clustering. Since the image has 3 colors create 3
clusters. Measure the distance using Euclidean Distance
Metric.
Fig 6 Black and White Image
Then the image the user selects after the segmentation
operation, is converted into black and white imageformatto
increase the accuracy.
Fig 6 Gray Scale Image
Then the user selected segmented image is then converted
into gray scale image to increase the accuracy.
Fig 7 Serial monitor connection to Wi-Fi and sensor values
display.
The readings obtained from soil moisture sensor and
ultrasonic sensor are shown here in the serial monitor.
Fig 8 Results in ubidots.
The readings shown in the serial port is being shown in the
ubidots server which can be viewed with an internet
connection.

Fig 9 Email notification to user
This the email that is sent to the farmer or the user from the
ubidots server about the status of the farm.
CONCLUSION
The proposed system provide agriculture solution using
Artificial Neural Network Machine learningalgorithm which
is used for performing data prediction on data sensed by
sensors. Due to use of IoT devicessystemprovideautomated
solution for data prediction. The accurately detection and
classification of the plant disease is very important for the
successful cultivation of crop and this can be done using
image processing.
ACKNOWLEDGEMENT
The authors express their sincere gratitude to the Principal
of Jyothy Institute of Technology, K. Gopalakrishna, Head of
the Computer Department, Dr. Prabhanjan, and our Project
guide, Mrs. Nikitha S, for giving constant encouragement and
support to complete the work.
REFERENCES
[1] Sarika Rakshak, Prof. R. W. Deshpande. Automated
Irrigation System Based on Arduino Controller Using
Sensors.//2017 IEEE.
[2] Chandan kumar sahu, Pramitee Behera, A Low Cost
Smart Irrigation Control System.//2015 IEEE.
[3] Yunseop (James) Kim, Robert G. Evansand William M.
Iversen, Remote Sensing and Control of an IrrigationSystem
Using a Distributed Wireless Sensor Network//IEEE 2008
[4] Sharada Prasanna Mohanty, David Hughes, Marcel
Salathé, Using Deep Learning for Image-Based Plant Disease
Detection //IEEE 2018
[5] Sachin D. Khirade, A. B. Patil, Plant Disease Setection
using Image processing//IEEE 2015

IRJET- Automated Irrigation using IoT and Plant Disease Detection using Image Processing and Machine Learning

More Related Content

What's hot (20)

Similar to IRJET- Automated Irrigation using IoT and Plant Disease Detection using Image Processing and Machine Learning (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Automated Irrigation using IoT and Plant Disease Detection using Image Processing and Machine Learning