Abstract

This paper presents a simple rain prediction device-based automatic irrigation

management algorithm using a combination of weather parameters and soil moisture
measurements for the water balance required for a crop at each condition during its
growing phase that will reduce farmer intervention for irrigation and avoid unnecessary
irrigation by predicting the rainfall before starting the motor for irrigating the field. This
device is powered by various technologies like deep learning to classify clouds responsible
for rain, machine learning models to predict rainfall based on atmospheric parameters and
the Internet of Things (IoT) using different sensors to collect data from the field. This
algorithm is very appropriate for farmers who are in remote locations and are not able to
use the internet and WIFI due to its unavailability. The device will be attached to the motor,
will take the data from sensors and will do the rain prediction at device level only and will
switch ON/OFF the motor based on the soil moisture value and rain prediction without any
human intervention.

Introduction

Water is the most significant resource for each and every living thing and is used
largely for the agriculture sector. Agriculture uses around 70% of the world's
surface water in irrigating the land to produce crops. Agriculture, which is the sole
source of human food, is the world's main industry and major land use, accounting
for 40% of all available land. Agriculture has a crucial role in a country's economic
development Sah et al. (2018). Furthermore, crop-based food products provide for
78 percent of global average per capita energy demands, while other food sources
such as eggs, milk, and meat account for the remaining 20% Brevik, Eric (2013). As
a result, the
essential necessity is to meet the dietary demands of a growing population, which
can only be met by increasing agricultural productivity. Agriculture and its related
sectors provide 17 percent of India's gross domestic product, according to the
Ministry of Statistics and Program Implementation (GDP, for the years 2009 and
2017). Furthermore, it employs almost two-thirds of India's workforce Deshpande
(2017). In India, agriculture provides the prime source of income for more than 58
percent of rural households, Figure 1 depicts these agricultural dimensions. As a
result, progress in the agriculture division is directly linked to the creation of jobs and
the reduction of poverty in developing nations Srivastava et al. (2016). Monitoring
agricultural activities is the most required necessity of today for minimizing human
intervention in the field because of the scarcity of labour and urbanization. Demand
for food is rising on daily basis and meeting it with traditional agricultural practices is
becoming increasingly difficult. Agriculture monitoring is a top priority because it
helps to reduce labour costs while increasing output which requires to adopt new
technologies and modern practices.

Figure 1: Importance of the agricultural sector for India

Various monitoring activities in the field of agriculture have been carried out using
Artificial Intelligence, such as soil monitoring based on soil conditions to encourage
multi-cropping by forming zones based on similar properties of soil and to help
farmers in taking decisions on the usage of various resources like fertilizers, water,
pesticides based on early disease prediction in the correct amount and at right time.
Today, technology is spreading its wings in every sector, including agriculture so
our farmers must be educated and should be motivated to integrate and make use
of technology as much as possible to increase crop yield by reducing the adverse
impact of traditional practices on the environment. Deep learning offers a vast range
of applications, and its use in industry has advanced dramatically. Deep learning
has a distinct benefit over machine learning in that it provides depth to the latter.
Machine learning paradigm developments are breaking down barriers to required
data analyses in many IoT-based applications. Deep learning models have
demonstrated outstanding results in a variety of fields, including speech
recognition, natural language processing, image recognition, information retrieval,
indoor localization, physiological and psychological state detection, and so on, and
these services serve as the foundation for IoT applications Mohammadi et al.
(2018).
 Wireless Sensor Networks (WSN) and Precision Agriculture (PA) applications bring
together an exciting new area of research that will considerably increase agricultural
production quality, precision irrigation, and cost savings. In addition, the ease of
deployment, system maintenance, and monitoring paves the path for WSN systems
to be accepted in PA. Implementation cost can be reduced and so make WSN a
more appealing solution for all kinds of fields and cultivations using the optimized
methodologies in selecting the appropriate sensor architecture. Internet of Things
(IoT) has proven to be useful in real-time data monitoring and can benefit farmers
through a variety of important approaches by selecting the crops according to the
soil, amount, and time of fertilization at a particular location for a particular crop,
weed detection etc. that can help them grow better crops and manage their fields
more effectively. As food is the most basic need of any human being, this ultimately
aids in the country's general growth.

Need for Precision Agriculture in India

Alarming obstacles to the global food system include global warming, resource
depletion, surging food demand, and growing labour prices. However, the
difficulties are greater and more numerous in India, including small and dispersed
land holdings, declining productivity, depleting natural resources, seasonal
production, climatic fluctuations, lack of an eco-regional approach and stagnant
farm incomes. A time like these calls for the best possible use of scarce resources,
which is where newly developed precision farming technologies can help enhance
agricultural output and profitability.
With the massive challenges of biotic and abiotic pressures experienced by crops
and the enormous need for food grain of 480 million tons (Mt) by the year 2050, the
introduction and acceptance of contemporary technologies in Indian agriculture is
unavoidable.

Drawbacks of Precision Agriculture

● High cost: Farmers may be discouraged from using this farming practice by
high capital expenditures.
● Lack of technical expertise knowledge and technology: Before being used,
pre- cision agriculture techniques still need professional advice as they are in the
initial stages of development.
● Not applicable or difficult/costly for small landholdings: Before the system has
collected enough data to be completely implemented, it may take several years for
small landholdings. It is a really challenging undertaking, especially the data
collect- ing and analysis.
● Heterogeneity of cropping systems and market imperfections
Benefits of Precision Farming for Indian Agriculture
● In India, improving and expanding the use of precision agriculture technologies
can aid in lowering production costs, raising productivity, and more effectively
utilizing natural resources.
● By enhancing productivity, crop quality, sustainability, profitability, environmental
protection, on-farm quality of life, food safety, and rural economic development,
it has the potential to revolutionize modern farm management in India.
● Site-specific uses of irrigation, herbicides, and fertilizers in cotton fields, oil palm
plantations in South India, and coffee and tea gardens in Eastern India can
signifi- cantly lower production costs and chemical pollution in the environment.
● When there are not enough water resources, it can improve irrigation efficiency.
● Farmers can use forecasting to control issues like pests, illnesses, nutritional defi-
ciencies, and plant diseases.
● Furthermore, it increases the sector's opportunities for skilled employment and
intro- duces new techniques for assessing multifunctional factors such as non-
market functions.
● It plays a crucial part in keeping track of greenhouse conditions in farming areas.

Policy Approach to Promote Precision Farming at Farm Level

● Determine the markets where crop-specific precision farming can be most
effec- tively promoted.
● forming multidisciplinary teams of agricultural scientists from various fields, engi-
neers, manufacturers, and economics to research the whole use of precision
agri- culture
● Give the farmers full technological support so they may create prototypes or
models that can be copied on a wide scale.
● Farmers' fields should be the subject of a pilot study to demonstrate the
effective- ness of precision agriculture.
● educating farmers on the effects of using unbalanced levels of agricultural
inputs such as irrigation, fertilizer, insecticides, and pesticides

State of Precision Farming in India Compared to rest of the world

● Although precision farming is thought to have originated in the US, it has since
spread throughout the world due to its advantages in terms of lower costs and
in- creased productivity.
● European countries like the UK, Germany, France, and the Netherlands have
adopted precision agriculture to maximize the use of fertilizer and pesticides and
boost output.
● This trend has been further fuelled by environmental concerns. Countries like Can-
ada, Australia, Argentina, and Brazil are putting precision farming principles into
re- ality due to the necessity for a sustainable farming system.
● Asia, China, and India are among the nations experimenting with precision
farming due to their growing populations, limited landholdings, and rising labour
expenses
● Even though this technology has the potential to be successful, it is still in its
infancy in India. To get this idea out to Indian farmers, organizations like Trimble,
Tata Kisan Kendra (TKK), and Fasal, among many others, are working. TKK, a
project started by Tata Chemicals Limited (TCL), aims to move rural India away
from anti- quated farming methods and toward the satellite and IT era. The TKKs
deliver TCL's extension services to farmers, employing remote sensing to evaluate
soil conditions, look at crop health, look for insect infestations, and predict crop
yield. They help farmers quickly adjust to changing conditions, which increases crop
yield and increases farmers' income.

● Precision farming is being aggressively implemented by the Tamil Nadu Precision

Farming Project to cover a bigger geographic region.
● Despite initial hostility from the farmers, its programs for drip irrigation and crop
pro- duction have received widespread acceptance (Mandal 2009)
● The Central Potato Research Station's (CPRS) investigation of the role of remote
sensing in mapping the variability with regard to place and time (Shanwad 2004)
will aid in the local farming of potatoes.
● The Dindigul area of Tamil Nadu has been chosen by the M S Swaminathan Re-
search Foundation in cooperation with NABARD for the experimentation of
Variable Rate Input Application (Shanwad 2004).
● Precision farming methods are being tested at the institution farm of the Indian
Agri- culture Research Institution.
● Variable Input Application in Different Cropping Systems has been started by Pro-
ject Directorate for Cropping Systems Research (PDCSR), Modipuram and Meerut
(UP), in Collaboration with Central Institute of Agriculture Engineering (CIAE),
Bho- pal.

Challenges in Adopting Precision Agriculture in India

● Due to its distinct pattern of land ownership,

● inadequate infrastructure,
● lack of farmers' willingness to take on the risk,
● social and economic circumstances, and demographic circumstances, the
adoption of precision farming in India is still in its infancy.

Most of the Indian agriculture is carried out on small parcels of land, which restricts
the economic benefits of precision farming equipment that is already in use and
therefore, the farmers of developing nations like India, Brazil, and China rely on
traditional farming rather than tech-driven farming practices.
For the Indian agricultural industry, the successful development and wide-scale
application of precision agricultural technologies is still a long way off. Such smart
farming concepts would enable the Indian agro-industry to undergo a transformation
thanks to the nation's expanding IT sector and extensive research in the agriculture
sector.
The need for automation in the agricultural sector is very essential, and there are
numerous ways to put it into practice. Irrigation is the first area where automation is
required for efficient water use. The soil moisture sensor monitors the moisture
level of the soil and begins watering the farm when the value falls below the
farmer's chosen threshold based on the type of crop and its age. The embedded
system and the Internet of Things have aided in the development of a small system
that monitors the farm's water level without the need for human intervention. The
Internet of Things is mostly used in intelligent watering systems. As efficient use of
existing fresh water is critical, and the water problem can be resolved with
advancements in technology and the application of automation Ahmed et al. (2021).

Novelty: A rainfall prediction device prototype is introduced which is integrated with

common and affordable sensors to sense atmospheric parameters and with a USB
camera to capture sky images on a raspberry Pi where a rain prediction model
based on a machine learning model that take input temperature, pressure and
relative humidity as input and Deep learning model that works on cloud images to
predict rainfall. Edge analytics is used to avoid the use of WIFI or the internet so
the prediction of rainfall based on both models is given at the device level itself
where the input parameters are sensed. The novelty of the device is that it can be
used in many decision-making activities where rainfall plays a major role. It is used
for nowcasting without the help of any internet or WiFi. 5V of power is required to
switch on the device and is portable.

Precision irrigation is a critical step toward achieving food productivity and security
while simultaneously implementing water-preserving techniques to overcome the
unpredictable nature of rainfall and the impact of a water shortage caused by
phenomena like drought in various regions of the world. The aim of Precision
irrigation scheduling is to provide the proper amount of water to each plant at the
right time to overcome the water loss caused by a process such as soil erosion,
deep percolation, or evapotranspiration while avoiding excess- and insufficient
irrigation, Benyezza et al. (2018), Devanand Kumar et al. (2020), Bigah et al.
(2019), Gu et al. (2020). Irrigation Management is a key process in saving water
consumption and other indirect expenditures occurred from other sources of energy
like electricity or fossil fuel for operating pumps with good irrigation management
through proper monitoring for optimal cost-effectiveness Togneri et al. (2019),
Cáceres et al. (2021). The use of artificial neural networks for an estimate and
modelling the non-linear characteristics of reference evapotranspiration has been
proposed and explained by Sharma et al. (2016), Kelley et al. (2019), in addition to
the importance of precise evapotranspiration estimation in aiding precision irrigation
management. This approach uses parameters like temperature, humidity, wind
speed, and solar radiation and was able to accurately estimate water requirements
for the crop that may be used to take appropriate irrigation decisions. Penchalaiah
et al. (2021) created an IoT-based data-driven irrigation system where soil moisture
is predicted using this method, and the results are compared to predictions from
other models. The new method is proven to be more effective, and the soil moisture
content shown is promising. Bhoi et al. (2021) present an irrigation recommendation
that incorporates machine learning methods like regression tree and Support Vector
Machine, as well as agglomerative clustering. The system recommended performed
admirably on both its own gathered data set and the accessible crop data
collection. Risheh et al. (2020) using artificial neural networks and an IoT
architecture, created a dependable system for greenhouse irrigation and
demonstrate the superior
performance of neural networks compared to the current alternative method of
support vector regression using a dataset gathered by conducting tests on various
soils also used transfer learning technique to reduce the processing power and
speed up the training. Mehra et al. (2020) proposed an intelligent IoT-based
hydroponic system using Deep Neural Networks. A prototype for Tomato plant
growth as a case study was developed using Arduino, Raspberry Pi3 and Tensor
Flow based on the numerous input parameters obtained, the system as created is
sophisticated enough to provide the proper control action for the hydroponic
environment. In order to achieve irrigation optimization with weather and soil
conditions as key components, Difallah et al. (2017) created a linear programming
model combined with a knapsack decisional form. The findings showed a reduction
in water use of 28.5%. Better optimization would emerge from taking into account
additional important extrinsic elements such as relative humidity, soil nutrients, wind
speed, and sunshine length. In Alibabaei et al. (2021) the ANN controller is used to
calculate the error by comparing the target soil moisture content to the actual soil
moisture content, and based on that calculation, the valves were opened and
closed. An Internet of Things (IoT)-based wireless sensor network (WSN)
architecture was created by Kamaruddin et al. (2019) and it monitors and manages
the irrigation system either manually or automatically. The communication network
transceiver and CPU used in the suggested manner were NRF24L01 and will be
used to transmit the soil moisture sensor data to the base station. The data from the
sensor node will then be transmitted via the base station to the cloud server. For
the purpose of connecting to an Android application and storing all the data in a
database, this project used Thing speak as a cloud server., Kanmani et al. (2021),
where Deep Learning models are frequently utilised for image and sound
processing, deep learning models have greatly improved the state-of-the-art.

Proposed system
An automated irrigation system is proposed that works with sensors and sense
atmospheric parameters, soil moisture and capture sky image to know the types of
clouds present to predict the rainfall whether it will be no rain, low to medium rain or
medium to high rain which will help in making decision on the time for which motor
should kept on. As shown in Figure 2, the rain prediction model combines machine
learning model using parameters from the atmosphere (temperature, relative
humidity, atmospheric pressure) and deep learning technique (by giving input as
sky/cloud image) to predict the intensity of rain.
Figure 2: Rain prediction device model

The working process is presented in the form of an algorithm as shown in Figure 3,

where soil moisture is recorded continuously and once it reaches a nearby
threshold it triggers a rain prediction model by sensing atmospheric parameters
using sensors that are integrated with Raspberry Pi along with the integrated USB
camera which gets activated and captures live images of the sky.
Next, the atmospheric parameters temperature, relative humidity and atmospheric
pressure are provided to the machine learning model as input, the Random Forest
algorithm is trained on the dataset based on Hyderabad city which is collected from
the NASA Power Data Viewer for Daily Data, that contains a variety of climate
characteristics. The most important ones, however, are selected, and the data is
taken for a 40-year period between May 1981 and 2021. Given that there are
typically 64–65 rainy days per year, a large period is used to gather enough data
and is used to predict the rainfall. The dataset is manually downloaded and created
as per the requirements as well as experimented by Geeta Ambildhuke and Barnali
Gupta Banik (2022). At the same time, the images of the sky/cloud are fetched as
an input to the deep learning model trained on a ground-based cloud dataset to get
the rainfall prediction based on the types of clouds present in the sky at that time is
also demonstrated by Ambildhuke, Geeta Mahadeo, and Barnali Gupta (2021).
Figure 3: Algorithm for automatic water balance irrigation model powered by
rain prediction
Once the inputs are given to the machine learning model and deep learning model
the output is given as the intensity (as probability) of rain, whether it will be No rain
to Low Rain, Low to Medium Rain or Medium to High rain and then depending on
the rainfall prediction the motor will be ON and irrigate the field for particular time as
shown in Figure 4.

Figure 4: Workflow of the automatic Irrigation model

Methodology

The state of the soil water for an irrigated crop must be monitored on a regular
basis to assist the irrigation manager in making irrigation decisions. Irrigation
scheduling is typically accomplished in one of two ways. One method is to use soil
moisture sensors to monitor soil water directly. On the other hand, the soil-water
balance strategy, uses weather data to account for soil-water in the rooting depth.
Weather- based or evapotranspiration (ET) based irrigation scheduling or water
balancing method are common terms for this method.

Evapotranspiration
In a farm environment, ET offers an objective and reliable estimate of the water
requirements of actively growing crops. Irrigators can better schedule irrigations
using evapotranspiration data resulting in higher yields and higher water
productivity.
ET is described as the cumulative process of evaporation from soil and plant
surfaces, as well as transpiration from plant canopies to the atmosphere via the
stomates (tiny holes on the leaf surface). Water is released in the form of water
vapor from the plant surfaces and soil into the atmosphere as part of the ET
process. Advanced approaches can be used to directly measure crop ET. However,
the most typical approach of predicting the ET rate for a specific crop involves first
computing reference ET0 and then using the appropriate crop factors to determine
real crop ET explained in Antonopoulos et al. (2017).
The formula for determining the water need is discussed in Eq. (1) based on
agricultural literature.
ET0 * Kc= Wrequired or (ETc) (1),
Where, Wrequired denotes the amount of water required.
Where ET0 is the reference evapotranspiration - which is the evapotranspiration
from a reference crop and is calculated using standardized Penman-Monteith
equation and Kc is the crop factor. Kc varies by crop development stage as
elaborated in Djaman et al.(2018).

Technology Used

Deep Learning Model is implemented to get the prediction from the cloud images
captured that uses a transfer learning approach to train the pretrained model on
new dataset of cloud images. Machine learning models are trained on three
atmospheric temperatures, humidity and pressure to get rain prediction and
Random Forest Model is selected due to its good accuracy and predictions. The
main feature of this device is the use of Edge analytics where the collection of
data, processing of data and predicting output and automation is all done at the
device level that does not require any cloud infrastructure and internet connectivity
so can be used at any remote location without internet and requires only a power of
5V to start the device as demonstrated by Geeta Ambildhuke and Barnali Gupta
Banik (2022).

Software Use

Python code is used to build the DL Model and also to get the sensor data
collected from the sensors connected to the Raspberry Pi device. Because the
Deep Learning model and machine learning models were trained on Python 3, all
supporting libraries with the required versions Keras 2.4.3, Tensor Flow-2.4.0-rc2,
Scikitlearn-0.20.2, and OpenCV -4.5.3 were installed.

Device Setup
The device consists of Raspberry Pi which is the controller that has a trained
machine learning model (pickle file) and deep learning model (.h5 file). To collect
the real data, sensors like DHT11, BMP 180, USB Camera, are connected for the
prediction of rainfall along with motor and soil moisture sensor. This is used to
automate the irrigation process without any intervention of the farmer and this
device does not require any cloud platform and WIFI as the input collection,
running models and giving prediction is all done at the device level itself. The
specification of DHT11 and BMP180 sensors are shown in Table 1 and Table 2.
As BMP180, OLED Display are I2C devices, they are connected to Raspberry
through I2C interface available on board. I2C devices have their unique address
and can access the modules through these unique address for sending and
receiving the data.

Table 1: DHT11 sensor specifications

DHT11 Sensor Specifications
Temperature range 0°C to 50°C
Humidity range 20% to 90%
Operating power 3.5V to 5.5V
Accuracy ±1°C and ±1%
Price $4 to $ 5

Table 2: BMP180 sensor specifications

BMP180 Sensor Specifications
Pressure range 300 hPa to 1100 hPa
Operating power 3.5V to 5.5V
Accuracy ± 0.2 hPa
Price $3 to $ 4

As Raspberry PI does not have analog data pins, to read the soil moisture data, an
analog-to-digital converter module ADS1115 is used for converting the soil
moisture values received from the soil moisture voltage values to analog values.
Analog to Digital Converter (ADC) used here is an I2C device and the module is
connected to I2C pins on the Raspberry Pi board. Through I2C communication the
data is read and processed as per the requirement of the application. Once the soil
moisture sensor values cross a threshold value, then the prediction is taken and
switch the 5V water pump ON/OFF accordingly which is controlled by a motor
driver shield.

The sensors are chosen based on their simplicity and affordability and are working
very well with the device and captures input data from the atmosphere as the
requirements. Different temperature sensors are available some of which are
compared based on some properties as shown in Table 3.
Table 3: Comparison between similar temperature/humidity sensors

Sensor DHT11 DHT22 SHT71

Parameters Temperature, Temperature, Temperature,
Humidity Humidity Humidity
Measured
Voltage required 3.5 to 5.5 V 3.5 to 6 V 2.4 to 5.5 V
Temperature range 0°C to 50°C -40 °C to 80°C -40 °C to
123.8°C
Humidity range 20% to 90% 0% to 100% 0% to 100%
Accuracy ±2°C and ±2% ±0.5°C and ±1% ±0.4°C and ±3%
response speed Slow Fast Fastest
power High Low Lowest
Consumption
Price $4 to $5 $4.99 to $ 9.99 $30 to $40

Many researchers used different controllers and successfully implemented

automatic irrigation models with different sensors and weather data from nearby
weather stations. For the data collected and analysed as shown in Figure 5 the
Arduino UNO is the most popular node for irrigation automation and monitoring,
Veerachamy et al. (2022). For the requirement of the proposed device deep
learning and machine learning models need to be run on the device level only
Raspberry Pi is chosen for the experiment.

Figure 5: Controllers used in precision agriculture for automatic irrigation

 Results and Discussion

The device prototype is made with the IoT components to sense atmospheric
parameters using sensors like DHT11, BMP 180, USB Camera, and Soil Moisture
Sensor.

Arduino-based Soil moisture Sensor is used to monitor the amount of moisture in

the soil. Soil moisture level can be determined by the output voltage produced
based on resistance. If water is more conductivity will be more and resistance will
be less and vice versa. For easy understanding, the output voltage is scaled
between 0-100 values. When the value of Soil moisture is 100, it means the soil is
completely dry García, L.et al. (2020).

1. Based on the soil nature various threshold values are devised to know the wet-
ness using a soil moisture sensor and th e motor is allowed to work only
on need.

2. Once the soil moisture value goes near the threshold the rain prediction model
gets started and the USB camera gets activated to capture the current sky im-
ages and are passed as input to the deep learning model to get the prediction
of rainfall based on types of clouds present in that location at the current time.

3. Equally, the atmospheric parameters sensed are passed as input to the ma-
chine learning model to predict the rainfall probability based on the sensed
pa- rameters at the current location at the current time.

4. Based on the prediction the motor will be ON for the time to fulfil the need of
crops by keeping track of rain.

The device is used for rain prediction before taking the decision on irrigating the
crop when soil moisture goes below the threshold. The motor pump and controller
are connected by the relay. The relay performs the function of a digital switch,
turning the motor pump ON or OFF in response to commands from the controller.
A 5v Dc motor is used to pump water for the demonstration purpose. The
deployment cost at initial stage, particularly for small-scale farmers, is one of the
key hurdles in adopting digital technologies, machine learning and software
applications in terms of enhancing sustainable precision irrigation.
Figure 6: Automatic irrigation using real time parameters

As depicted in Figure 6. The soil moisture shows little change and remains almost
constant after evening hours as there is no significant loss in moisture from the
plant due to reduced evapotranspiration rate.
This necessitates the digitization of the farm process, as well as the usage of
sensors, actuators, and networking of precision agriculture devices. Water use
efficiency can be enhanced by better predicting irrigation needs, better matching
timing and volume to plant water needs, and adaptively compensating for water
loss owing to evapotranspiration. This automatic irrigation results in a higher yield
while utilizing less irrigation and wasting less irrigation water. The system becomes
smart and can have certain autonomous features for irrigation decision-making as a
result of the model’s training and ultimate implementation. As a result, most of the
irrigation stress experienced by farmers and users can be reduced.

Conclusion and Recommendations

National and international institutes have recognized the notion of climate-smart

agriculture, and various programmes have been launched to create climate-smart
technologies. However, the adoption and distribution of climate-smart technology,
tools, and behaviours is still primarily a continuous and complicated process. The
farming community can only optimise water usage through the use of smart
irrigation systems. It is a smart system because of the auto mode, and it may be
further adjusted for application-specific conditions. The integration of smart
technology like deep learning, machine learning and IoT is majorly driving
agriculture to sustainability and productivity by achieving sustainable precision
irrigation by controlling the usage of water. To fill the technology gap, adopting best
practices is crucial, and digital communication is required. Technology diffusion
heavily relies on agricultural extension, an area where the private sector is
becoming more involved.
For the Indian agricultural industry, the successful development and wide-scale
application of precision agricultural technologies are still a long way off. Such smart
farming concepts would enable the Indian agro-industry to undergo a
transformation thanks to the nation's expanding IT sector and extensive Agri-IT
research. Rapid socioeconomic developments including urbanization, energy
consumption, and economic expansion are opening new possibilities for the use of
precision farming in India. India should employ technology depending on the
demands of the nation's socioeconomic situation rather than blindly embracing the
sophisticated precision agriculture technologies used by Western nations.

References

Ahmed, A. A., Al Omari, S., Awal, R., Fares, A., & Chouikha, M. (2021). A
distributed system for supporting smart irrigation using Internet of Things
technology. Engineering Reports, 3(7). https://doi.org/10.1002/eng2.12352
Alibabaei, K., Gaspar, P. D., & Lima, T. M. (2021). Crop yield estimation using deep learning
based on climate big data and irrigation scheduling. Energies, 14(11), 3004.
https://doi.org/10.3390/en14113004
Ambildhuke, G. M., & Banik, B. G. (2021). Transfer Learning Approach-An Efficient Method
to Predict Rainfall Based on Ground-Based Cloud Images. Ingénierie des Systèmes
d'Information, 26(4).https://doi.org/10.18280/isi.260402
Ambildhuke, G., & Banik, B. G. (2022). IoT based Portable Weather Station for
Irrigation Management using Real-Time Parameters. International Journal of
Advanced Computer Science and Applications, 13(5).
https://doi.org/10.14569/ijacsa.2022.0130533
Benyezza, H., Bouhedda, M., Djellout, K., & Saidi, A. (2018), November). Smart
irrigation system based ThingSpeak and Arduino. In 2018 International conference
on applied smart systems (ICASS) (pp. 1-4). IEEE.
https://doi.org/10.1109/icass.2018.8651993
Bhoi, A., Nayak, R. P., Bhoi, S. K., & Sethi, S. (2021). Automated Precision Irrigation
System Using Machine Learning and IoT. In Intelligent Systems (pp. 275-282).
Springer, Singapore. https://doi.org/10.1007/978-981-33-6081-5_24
Bigah, Y., Rousseau, A. N., & Gumiere, S. J. (2019). Development of a steady-state
model to predict daily water table depth and root zone soil matric potential of a
cranberry field with a subirrigation system. Agricultural Water Management, 213,
1016-1027. https://doi.org/10.1016/j.agwat.2018.12.024
Brevik, E. C. (2013, April). Climate change, soils, and human health. In EGU General
Assembly Conference Abstracts (pp. EGU2013-7).
Cáceres, G., Millán, P., Pereira, M., & Lozano, D. (2021). Smart farm irrigation:
Model predictive control for economic optimal irrigation in agriculture.
Agronomy, 11(9), 1810. https://doi.org/10.3390/agronomy11091810
Deshpande, T. (2017). State of agriculture in India. PRS Legislative Research, 53(8),
6-7.
Devanand Kumar, G., Vidheya Raju, B., & Nandan, D. (2020). A review on the smart
irrigation system. Journal of Computational and Theoretical Nanoscience, 17(9-10),
4239-4243. https://doi.org/10.1166/jctn.2020.9053
 Difallah, W., Benahmed, K., Draoui, B., & Bounaama, F. (2017). Linear optimization model
for efficient use of irrigation water. International Journal of Agronomy, 2017.
Djaman, K., O’Neill, M., Owen, C. K., Smeal, D., Koudahe, K., West, M., ... & Irmak, S.
(2018). Crop evapotranspiration, irrigation water requirement and water productivity
of maize from meteorological data under semiarid climate. Water, 10(4),
405. https://doi.org/10.3390/w10040405
García, L., Parra, L., Jimenez, J. M., Lloret, J., & Lorenz, P. (2020). IoT-based smart
irrigation systems: An overview on the recent trends on sensors and IoT systems for
irrigation in precision agriculture. Sensors, 20(4), 1042.
https://doi.org/10.3390/s20041042
Gu, Z., Qi, Z., Burghate, R., Yuan, S., Jiao, X., & Xu, J. (2020). Irrigation scheduling
approaches and applications: A review. Journal of Irrigation and Drainage
Engineering, 146(6), 04020007. https://doi.org/10.1061/(asce)ir.1943-4774.0001464
Kamaruddin, F., Abd Malik, N. N. N., Murad, N. A., Latiff, N. M. A. A., Yusof, S. K. S., &
Hamzah, S. A. (2019). IoT-based intelligent irrigation management and monitoring
system using arduino. TELKOMNIKA (Telecommunication Computing Electronics
and Control), 17(5), 2378-2388. http://doi.org/10.12928/telkomnika.v17i5.12818
Kanmani, R., Muthulakshmi, S., Subitcha, K. S., Sriranjani, M., Radhapoorani, R., &
Suagnya, N. (2021). March). Modern irrigation system using convolutional neural
network. In 2021 7th International Conference on Advanced Computing and
Communication Systems (ICACCS) (Vol. 1, pp. 592-597). IEEE.
https://doi.org/10.1109/icaccs51430.2021.9441917
Kelley, J., & Pardyjak, E. R. (2019). Using neural networks to estimate site-specific
crop evapotranspiration with low-cost sensors. Agronomy, 9(2), 108.
https://doi.org/10.3390/agronomy9020108
Mehra, M., Saxena, S., Sankaranarayanan, S., Tom, R. J., & Veeramanikandan, M.
(2018). IoT based hydroponics system using Deep Neural Networks. Computers
and electronics in agriculture, 155, 473-486.
https://doi.org/10.1016/j.compag.2018.10.015
Mohammadi, M., Al-Fuqaha, A., Sorour, S., & Guizani, M. (2018). Deep learning for
IoT big data and streaming analytics: A survey. IEEE Communications Surveys &
Tutorials, 20(4), 2923-2960. https://doi.org/10.1109/comst.2018.2844341
Penchalaiah, N., Nelson Emmanuel, J., Suraj Kamal, S., & Lakshmi Narayana, C. V.
(2021). IoT Based Smart Farming Using Thingspeak and MATLAB. In ICCCE
2020 (pp. 1273-1295). Springer, Singapore. https://doi.org/10.1007/978-981-15-
7961-5_117
Risheh, A., Jalili, A., & Nazerfard, E. (2020)). Smart Irrigation IoT solution
using transfer learning for neural networks. In 2020 10th International Conference on
Computer and Knowledge Engineering (ICCKE) (pp. 342-349). IEEE.
https://doi.org/10.1109/iccke50421.2020.9303612
Sah, D., & Devakumar, A. S. (2018). The carbon footprint of agricultural crop
cultivation in India. Carbon Management, 9(3), 213-225.
https://doi.org/10.1080/17583004.2018.1457908
Shanwad, U. K., Patil, V. C., & Gowda, H. H. (2004). Precision farming: dreams and realities
for Indian agriculture. Map India.
Singh, R., Singh, H., & Raghubanshi, A. S. (2019). Challenges and opportunities for
agricultural sustainability in changing climate scenarios: a perspective on
 Indian agriculture. Tropical Ecology, 60(2), 167-185. https://doi.org/10.1007/s42965-
019-00029-w
Togneri, R., Kamienski, C., Dantas, R., Prati, R., Toscano, A., Soininen, J. P., &
Cinotti, T. S. (2019). Advancing IoT-based smart irrigation. IEEE Internet of
Things Magazine, 2(4), 20-25. https://doi.org/10.1109/iotm.0001.1900046
Veerachamy, R., Ramar, R., Balaji, S., & Sharmila, L. (2022). Autonomous Application
Controls on Smart Irrigation. Computers and Electrical Engineering, 100, 107855.
https://doi.org/10.1016/j.compeleceng.2022.107855