Volume 8, Issue 8, August 2023 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Smart Agriculture using Automation

D.Sandeep1, B.Mahitha2, Ch.Prajwal3, K.Lokesh4, M.Balaji5
Department of Information Technology MLR Institute of Technology Hyderabad, India

Abstract:- Farming practices in India could be transformed by the suggested system for Smart Management
revolutionizedby integrating the Internet of Things (IoT) of Crop Cultivation utilizing IoT and Machine Learning. By
with agriculture. IoT sensors can be used to gather providing farmers with predictive insights, the system can
information on soil moisture, temperature, and humidity, help them make informed decisions about which crops to
which can then be usedto help farmers choose the right grow, optimize their crop production, and maximize their
crop and apply irrigation techniques. By using a profits. This will not only improve the livelihoods of farmers
microcontroller-based system, farmers can receive real- but also contribute to the overall growth of the Indian
time data on soil conditions and adjust irrigation economy.
schedules accordingly, without having to manually
monitor the crop. Moreover, the implementation of a II. LITERATURE SURVEY
recommendation engine that takes into account climatic
factors and crop quantity can assist farmers in selecting Both the economy and the existence of the Indian
the most appropriate crops for their specific conditions. people depend on agriculture. The technique is designed to
By usinga content-based recommendation technique, the assist farmers in boosting agricultural productivity. The
system can analyze the data collected by the sensors and instruments used to investigate the soil include a pH sensor,
provide farmerswith suggestions on which crops to grow, a temperature sensor, and a humidity sensor.[2] Creating an
based on their suitability for the given climate and soil efficient IoT-based smart irrigation system is anotheressential
requirement for farmers in the agricultural sector. The system
conditions This approach can be particularly useful in
also incorporates sensors for temperature, humidity, and
areas where water is scarce, as it can help farmers
optimize their water usage and ensure that crops are raindrops that have been updated to enable online remote
being grown most efficiently and sustainably possible. monitoring of these variables. [3] India is a nation where
With the help of the IoT and a recommendation engine, agriculture is extremely important. This websitehas a feature
farmers can make data-driven decisions that can improve that lets you manually or automatically regulate how often
their crop yields, reduce water usage, and ultimately plants are watered. Using a Raspberry Pi camera that provides
increase their profitability. live streaming to the webpage, the health of the plants is
tracked. [4] Tensor flow and deep learning neural networks
Keywords:- IOT-based system, Machine Learning, Sensors, are used by this intelligent system to identify animals
Recommendation system, Agriculture. according to their hazard level as well as unauthorised human
visitors to the farm and to promptly inform the farmer. The
I. INTRODUCTION gadget comes with an android application that enables remote
access and live video streaming surveillance. [5] For the
To assist Indian farmers in making educated decisions production of crops, agricultural irrigation is crucial. The
about which crops to plant, the suggested system for Smart author of this paper claims that soil constitution is related to
Management of Crop Cultivation utilizing IoT and Machine the availability of nutrients that plants need as well as the
Learning is a step in the right direction. By considering presence in soil of nutrients and chemical composition that
sensed parameters such as temperature and humidity, The exist in various ratios that are best for plants and soil
method can forecast the crop that would grow best in a organisms. The author also claims that providing the right
specific climate, taking into account other elements like soil amount of water to plants is the most important factor in
type, farm location, and rainfall. This will enable farmers to ensuring that all other nutrients work as effectively as
optimize their crop production and maximize their profits. possible.
Machine learning algorithms can further enhance the
performance of the proposed system by predicting the most A. EXISTING METHOD
suitable crop for a particular environment based on historical The existing automated drip irrigation system can irrigate
data. IoT applications in agriculture have a number of only a single crop land. Soil moisture sensors sense the
advantages. For example, it makes it possible for farmers to moisture level of a single crop and only that particular crop is
gather information on a variety of characteristics, including irrigated. Crops are supplied with water above their
soil moisture, temperature, and humidity, all of which are requirement. Automated drip irrigation is carried out
essential for crop growth. Farmers can decide when to irrespective of the available water resource for the crop. The
fertilize when to water, and when to harvest their crops by existing system doesn’thave the capability to make decisions
analyzing this data. IoT tools can also assist farmers in to irrigate multi- crop land. Irrigation is carried out even if the
spotting pests and diseases and taking action before they available water level is insufficient and leads to incomplete
seriously harm their crops. By analyzing data from past years, irrigation of land. Water is wasted if the crop is irrigatedmore
the system can identify which crops have performed well in than its requirement. They often find it difficult tonotice plant
a particular location and recommend them to farmers. This diseases that directly affect the production rate. Using
will help farmers avoid the risk of crop failure and maximize appropriate parameters like rain patterns, temperature
their profits. In conclusion, Indian agriculture could be patterns, soil structures, and other factors such as crop

IJISRT23AUG1284 www.ijisrt.com 1691

Volume 8, Issue 8, August 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
diseases makes it possible to yield accurate crop prediction C. GAP ANALYSIS
results. Our system is designed to automate tasks and reduce the
workload for farmers while also providing effective
B. DISADVANTAGES: monitoring. precision irrigation is a promising solution to the
 In the existing system, they considered only a particular issue of water wastage in agriculture, and it has the potential
State and not all the states and other parameters. to improve crop yields, conserve water resources, and reduce
 Relatively slower to build. the environmental impact of farming. Our system also
 Hard to Interpret incorporates a decision-making algorithm to assist farmers in
 Computationally expensive. selecting crops that yield better results. Unlike other
 Water is wasted if the crop is irrigated more than its irrigation systems that waste water resources by using
requirement. excessive amounts of water, our system utilizes
 Manual irrigation technique moisture/humidity sensors to detect soil moisture levels. This
allows for the automatic irrigation of crops with the
appropriate amount of moisture, ensuring optimal growth
conditions. Our approach is designed to overcome the
limitations of traditional agricultural methods and improve
crop yields while conserving water resources.

III. SYSTEM ARCHITECTURE

Fig. 1: Crop Recommendation Architecture

Fig. 2: IoT Architecture

Our project's structure and behaviour are defined using the internal procedures, techniques, and stages of a system.
a conceptualization-oriented model called a system "Fig. 3.1" and "Fig. 3.2" show the system architecture of our
architecture. A system's architecture is a clearly defined and recommendation model and the internet of things model,
representative structure that is arranged to support the logic respectively. The decision tree algorithm is used in the crop
that underlies all of the system's structures and behaviours. prediction model, and in the Internet of Things, sensors are
Additionally, it shows how one process is dependent on used. The crop prediction model predicts the crop based on
another.It enables us to gain knowledge about our product in the soil moisture.
a style that makes it simple for us to relate to and comprehend

IJISRT23AUG1284 www.ijisrt.com 1692

Volume 8, Issue 8, August 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
IV. PROPOSED METHOD prepared. Techniques for data visualization can be used to
find pertinent links between variables and spot any class
It sounds like the proposed approach involves using imbalances that might skew the model's results. Finally, the
moisture/humidity sensors to automatically irrigate crops data should be split into training and evaluation sets to ensure
based on their water requirements, which are determined by that the model can generalize well to new data. By following
assigning each crop a priority based on its minimum water these steps, we can ensure that the model is trained on high-
requirement. Additionally, the project aims to recommend quality data and can make accurate predictions on new data.
crops to farmers using a Decision Tree Classifier, taking into
consideration factors such as crop yield in previous years,  Model Selection:
market trends, and the season of crop production. To achieve A decision tree is a machine learning technique that
these goals, the project will involve preprocessing the data represents a set of options and their potential outcomes using
provided, preparing a model, and connecting it to a UI a tree structure that resembles a flowchart. Each leaf node
interface using Flask. This will allow farmers to receive represents a result, the branches serve as decision-making
recommendations based on various agricultural parameters, guidelines, and the inside nodes of the tree represent traits or
giving them insight into which crops are likely to yield the attributes. Because of the transparency of their decision-
best results given the current conditions. making process and the ease with which people can
comprehend and interpret their reasoning, decision trees are
Overall, the proposed approach seems like it could be a regarded as white-box algorithms. This is because the tree
useful tool for farmers looking to optimize their crop structure of the algorithm provides a clear and intuitive visual
production and make informed decisions about which crops representation of how the decision-making process works. In
to plant based on a variety of factors. However, the success contrast, black box algorithms like neural networks are more
of the project will depend on the accuracy of the models used opaque, which means that it can be difficult to understand
to make these recommendations, as well as the ability of the how they arrive at their decisions. Decision trees are non-
UI interface to effectively communicate this information to parametric, which means they don't rely on assumptions
farmers in a user-friendly manner. about probability distributions, and they train more quickly
than neural networks. High-dimensional data can be handled
A. Module Identification using decision trees with good accuracy, and their decision
 Data Collection: rules are typically expressed as if-then-else statements. The
The process of gathering data is essential because the complexity of the decision tree model increases with the
quantity and quality of the data can have a big impact on how depth of the tree, as more complex rules are created.
well a machine-learning model performs. Various techniques Information gain and Gini impurity are two popular methods
can be used to gather data, such as web scraping and manual used for selecting the best attribute at each node and
interventions. In the case of crop recommendation in India, a determining the quality of splitting in decision tree models. A
dataset from another source has been used. It is important to metric called entropy is used to quantify the impurity or
ensure that the data used is relevant and appropriate for the unpredictability in a collection of observations. In the context
model being developed. Additionally, it is crucial to give of decision trees, entropy is used to determine the best feature
proper credit and citations to the source of the data to avoid or attribute to split the data at each node. The entropy of a
plagiarism. group of observations is calculated based on the proportion of
observations in each class, where a higher proportion of one
class corresponds to higher entropy. A split that results inlower
entropy is considered to be more informative and is therefore
preferred when constructing a decision tree. Other metrics,
such as Gini impurity or information gain, can also be used to
 Dataset: determine the best split in a decision tree.
The dataset contains 821 individual data points and
consists of 14 columns, which are described below: The The amount of knowledge a feature imparts about a
dataset was narrowed down to 7 features, which include the class is measured using information gain, which also aids in
total number of states in India, the amount of rainfall in mm, deciding the order of characteristics in decision tree nodes.
the total groundwater level, the temperature in degree The sub- nodes are referred to as child nodes, and the main
Celsius, the number of soil types, the season suitable for node is referred to as the parent node. We can assess the
crops, and the types of crops. quality of splitting in a decision tree and raise the model's
precision by employing information gain.

 Data Preparation:  Analyze and Prediction:

Gathering and preparing the data for training is the next The dataset was narrowed down to 7 features, which
step in creating a machine-learning model. This involves include the total number of states in India, the amount of
several tasks such as cleaning the data to remove duplicates, rainfall in mm, the total groundwater level, the temperature
correcting errors, dealing with missing values, normalizing in degrees Celsius, the number of soil types, the season
the data, and converting data types if necessary. Once the data suitable for crops, and the types of crops.
is cleaned, it should be randomized to remove any effects
caused by the order in which the data was collected or

IJISRT23AUG1284 www.ijisrt.com 1693

Volume 8, Issue 8, August 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
The dataset was narrowed down to 7 features, which V. RESULTS AND DISCUSSIONS
include the total number of states in India, the amount of
rainfall in mm, the total groundwater level, the temperature To prepare the dataset, we first divided it into training
in degrees Celsius, the number of soil types, the season and testing sets, which were used to generate and evaluate our
suitable for crops, and the types of crops. crop suggestion prediction model. The dataset was sourced
from anopen-source platform and contained a total of 821 data
points, each with 14 columns of information. Once the model
was trained using the training dataset, it was evaluated for its
accuracy and minimum error. Based on the input data
provided, the final model was then utilized to forecast and
advise the best crops to be sown. Our model was able to
achieve an accuracy of 90.07%, making it an effective tool for
crop suggestion and decision-making.

Fig. 3: Crop prediction

Fig. 4: IOT Connection

VI. CONCLUSION AND FUTURE SCOPE REFERENCES

The importance of crop management in agriculture is [1.] Athani, S., Tejeshwar, C. H., Patil, M. M., Patil, P., &
thoroughly examined in this paper. Modern technology is Kulkarni, R. (2017, February 1). Soil moisture
needed by farmers to efficiently produce their crops. Various monitoring using IoT-enabled Arduino sensors with
agricultural factors have been analyzed using machine neural networks for improving soil management for
learning techniques. A review of the literature was done on farmers and predicting seasonal rainfall for planning
several methods used in agriculture, including soft computing future harvest in North Karnataka — India
methods and blooming neural networks, which are important [2.] Mehta, P., Shah, H., Kori, V., Vikani, V., Shukla, S.,
in making recommendations. Farmers' output can be boosted & Shenoy, M. (2015). Survey of unsupervised
by using personalized and pertinent recommendations that machine learning algorithms on precision agricultural
take factors like production and season into account. To cut data. IEEE Xplore.
back on labor and use less water, agriculture monitoring is [3.] Radhika, Y., & Shashi, M. (2009). Atmospheric
essential. Wireless sensors are used in several systems that Temperature Prediction using Support Vector
have been created to track and forecast soil conditions for Machines. International Journal of Computer Theory
irrigation. Furthermore, crop forecasting has been done using and Engineering.
machine learning approaches.

IJISRT23AUG1284 www.ijisrt.com 1694

Volume 8, Issue 8, August 2023 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
[4.] Koushik Anand, Jayakumar C, Mohana Muthu and
Sridhar A, “Automatic Drip Irrigation using Fuzzy
Logic and Mobile Technology”, Proceedings of
Technological Innovation in ICT for Agriculture and
Rural Development, 2015.
[5.] Karandeep K, “Machine Learning: Applications in
Indian Agriculture”, International Journal of
Advanced Research in Computer and Communication
Engineering, Vol.5, no.4, pp.342- 344, 2016.
[6.] Rumpf, T., A.-K. Mahlein, U. Steiner, E.-C. Oerke, H.-
[7.] W. Dehne, and L. Plümer. "Early Detection and
Classification of Plant Diseases with Support Vector
Machines Based on Hyperspectral Reflectance."
Computers and Electronics in Agriculture, Vol. 74,
no.1, pp.91-99, 2010
[8.] Archana and Priya, ”Design and Implementation of
Automatic Plant Watering System” presented at
International Journal of Advanced Engineering and
Global Technology, vol-04, Issue-01, Jan-2016.
[9.] Joaquin Gutierrez and Juan Francisco, “Automated
Irrigation System using a Wireless Sensor Network
and GPRS Module” presented at IEEE Transactions on
Instrumentation and Measurement, 2013.
[10.] Yunseop Kim and Robert G. Evans, “Remote Sensing
and Control of an Irrigation System using a Distributed
Wireless Sensor Network” presented at IEEE
Transactions on Instrumentation and Measurement,
Vol57, July-2008.
[11.] G. Parameswaran and K.Sivaprasath, ”Arduino Based
Smart Drip Irrigation System Using IOT” presented at
International Journal of Engineering Science and
Computing (IJESC), May 2016.
[12.] R.Subalakshmi and Anu Amal, “GSM Based
Automated Irrigation using Sensors” presented at
Special Issue published in International Journal of
Trend in Research and Development (IJTRD), March-
2016.
[13.] T.R. Lekhaa, “Efficient Crop Yield and Pesticide
Prediction for Improving Agricultural Economy using
Data Mining Techniques”, International Journal of
Modern Trends in Engineering and Science (IJMTES),
2016, Volume 03, Issue 10
[14.] Haedong Lee and Aekyung Moon, “Development of
Yield Prediction System Based on Real-time
Agricultural Meteorological Information” ,16th
International Conference on Advanced
Communication Technology, 2014
[15.] Agro Consultant: Intelligent Crop Recommendation
System Using Machine Learning Algorithms. Zeel
Doshi , Subhash Nadkarni, Rashi Agrawal, Prof.
Neepa Shah.
[16.] Crop Recommendation System for Precision
Agriculture S.Pudumalar*, E.Ramanujam*, R.Harine
Rajashree, C.Kavya, T.Kiruthika, J.Nisha.

IJISRT23AUG1284 www.ijisrt.com 1695