Leveraging Artificial Intelligence for Agricultural Advancement: A

Comprehensive Review

Sandipamu Raahalya and Saravanan Raj

National Institute of Agricultural Extension Management (MANAGE),

Rajendranagar, Hyderabad, Telangana, India

Keywords: Artificial Intelligence, Machine Learning, Deep Learning, Precision Agriculture, Innovation.

Abstract: Artificial Intelligence (AI) is bringing about significant changes across various sectors, with agriculture being

one of the fields poised to reap substantial benefits from its applications. This study explores the wide-ranging

and transformative ways in which AI is being utilized in agriculture to streamline processes, boost

productivity, and tackle the challenges confronting the global food system. Beginning with an overview of AI

technologies like machine learning, deep learning, and computer vision, the paper examines their relevance

to agricultural contexts. It then investigates specific instances of AI implementation in agriculture, covering

areas such as crop monitoring and management, precision agriculture, pest and disease detection, yield

prediction, soil health assessment, and autonomous farming systems. Through an extensive review of

literature and case studies, the paper showcases how AI-driven solutions empower farmers by providing real-

time insights, facilitating data-driven decision-making, and optimizing resource utilization. It underscores the

role of AI tools like ChatGPT and Claude in offering agricultural extension services. Additionally, the study

addresses challenges and digital AI startups associated with AI adoption in agriculture, including concerns

over data privacy, technological hurdles, and the necessity for scalable and cost-efficient solutions. It stresses

the significance of interdisciplinary collaboration among agricultural experts, data scientists, and

policymakers to fully leverage the potential of AI technology and foster innovation in the agricultural sector.

1 INTRODUCTION

The anticipated global population growth, expected to

surpass nine billion by 2050, necessitates a

substantial augmentation in agricultural production

by approximately 70% to fulfill the escalating

demand. However, merely about 10% of this

augmented demand can be met by utilizing fallow

land, leaving the remaining 90% reliant on

intensifying current production methods (FAO,

2015). Given this scenario, leveraging state-of-the-art

technological advancements to enhance farming

efficiency becomes paramount. Present strategies

aimed at intensifying agricultural production often

necessitate significant energy inputs, while market

preferences gravitate towards high-quality food

products. Concurrently, the upsurge in population has

resulted in heightened demand for food grains,

leading to inflation in agricultural commodity prices.

In reaction to these challenges, the agricultural sector

has experienced a substantial evolution propelled by

technological progressions, with artificial

intelligence (AI) emerging as a pivotal catalyst for

innovation and efficiency.

The phrase "Artificial Intelligence" was initially

coined during the 1955 Dartmouth Conference by

John McCarthy, who proposed a study based on the

idea that machines could simulate every aspect of

learning and intelligence through precise description.

Today, AI is a crucial field in computer science,

widely applied across various sectors such as

education, healthcare, finance, and manufacturing,

tackling challenges that are difficult for humans to

solve effectively.

Leveraging artificial intelligence enables us to

develop intelligent farming practices that mitigate

farmer losses and yield substantial harvests. AI's

applications span a spectrum of activities within the

agricultural value chain, encompassing precision

agriculture, crop monitoring, supply chain

optimization, and livestock management, thereby

reshaping the entirety of this domain.

In recent years, the successful application of deep

learning models, notably convolutional neural

networks (CNN), has been observed across various

128

Raahalya, S. and Raj, S.

Leveraging Artiﬁcial Intelligence for Agricultural Advancement: A Comprehensive Review.

DOI: 10.5220/0012883600004519

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Emerging Innovations for Sustainable Agriculture (ICEISA 2024), pages 128-136

ISBN: 978-989-758-714-6

domains of computer vision (CV). Examples include

traffic identification (Yang et al., 2019), medical

image recognition (Sundararajan, 2019), scenario text

detection (Melnyk et al., 2019), facial expression

identification (Kolhe, et al., 2011), and face

recognition (Kumar and Singh, 2020). Similarly, in

agriculture, deep learning approaches have been

extensively employed for the identification of plant

diseases and pests, with numerous domestic and

international companies developing WeChat applets

and photo recognition APP software utilizing deep

learning for this purpose. Therefore, methodologies

utilizing deep learning for plant disease and pest

identification hold substantial research value and

possess significant potential for market applications.

Machine learning, an evolving technology, has

proven highly effective in precision irrigation

systems. It replicates human decision-making

processes and addresses complex irrigation

management issues, including multiple variables,

nonlinearity, and time variation. According to

Chlingaryan et al. (2018), machine learning serves as

a robust and adaptable framework for data-driven

decision-making, providing expert intelligence for

the system. The combination of machine learning

with big data technologies has opened up new

avenues for analyzing large volumes of data from

diverse sensors autonomously, without explicit

programming (Liakos, et al. 2018).

This review endeavors to present the present state

of artificial intelligence in agriculture by elucidating

noteworthy considerations and achievements in crop

management, soil management, pest and disease

management, weed management, water use

management, weather forecasting, and supply chain

management. Furthermore, it assesses pressing

challenges encountered in this domain, such as the

uneven distribution of mechanization across regions,

concerns regarding security and privacy, and the

adaptability of algorithms in practical applications,

particularly in scenarios with physical heterogeneity

in plants and extensive datasets requiring processing.

Ultimately, the review emphasizes the significance of

establishing a comprehensive understanding of this

field, providing specific examples, identifying major

challenges, outlining potential applications, and

considering diverse circumstances in different

countries.

2 RESEARCH OBJECTIVES

The majority of agricultural processes and tasks still

heavily rely on manual labor. However, artificial

intelligence (AI) holds the potential to enhance

current technologies and assist in both the most

challenging and routine farming tasks. Given that

agriculture is a labor-intensive sector and labor

shortages are prevalent, automation presents itself as

a viable solution for farmers. It's crucial to note that

farm machinery driven by AI demonstrates greater

efficiency, productivity, and speed compared to

human workers. The primary objectives of this

article's research are as follows:

RO1: To identify and discuss the major applications

of AI in agriculture

RO2: To discuss the challenges associated with the

adoption of AI in agriculture.

3 APPLICATIONS OF AI IN

AGRICULTURE

3.1 Crop Management

The concept of integrating artificial intelligence into

crop management was initially proposed in 1985 by

McKinion and Lemmon, leading to the development

of GOSSYM. GOSSYM was a model designed to

simulate the growth of cotton crops, leveraging expert

systems to optimize cotton production (McKinion

and Lemmon, 1985). Today, various IoT sensors,

including temperature, weather, soil, and

environmental sensors, are utilized to monitor the

growth and root activity of crops in real time on

farms. Data collected by these sensors is transmitted

to cloud storage and displayed on a web application,

allowing smallholders in Africa to observe various

parameters and crop growth performance. Data

collection occurs at one-minute intervals, with IoT

sensors transmitting data to cloud storage

accordingly. Farmers can access this IoT data via

smartphones, enabling them to visualize information

and take necessary actions related to irrigation and

fertilizer applications on their farms (Barakabitze et

al., 2023).

Traditionally, farmers assessed the ripeness of

tomatoes by physically inspecting the field daily and

using their hands to determine their development.

However, modern agricultural practices now require

the assessment of tomato maturity on an industrial

scale. Fortunately, computers have greatly simplified

various aspects of life, including agricultural

processes. For example, machines in factories can

detect wastage and ripeness, while in the field,

different AI technologies enable farmers to evaluate

Leveraging Artiﬁcial Intelligence for Agricultural Advancement: A Comprehensive Review

129

the freshness of tomatoes without physical contact

(Chang et al., 2021; Sharma et al., 2022).

A) Drones and Computer Vision for Crop

Analysis:

Drones primarily gather information by analyzing the

reflected light from crops. In agriculture, growers can

utilize specific types of sensors to collect data that

identifies areas with issues, allowing them to take

appropriate actions. There are two types of sensors

used: thermal sensors and hyperspectral sensors.

Drones rely on the reflection of light from crops to

obtain information. Photosynthesis in plants is driven

by the absorption of visible light. However, Near

Infrared (NIR) photons do not provide energy for

photosynthesis but do emit heat. Plants, as a

consequence, have adapted to reflect near-infrared

(NIR) light. The decay of this reflection mechanism

occurs when the leaf withers. Near-infrared sensors

exploit this characteristic by quantifying the disparity

between NIR reflectance and visible reflectance,

which is referred to as the Normalized Difference

Vegetation Index (NDVI).

Source: agribotix

Figure 1: NDVI and Plant health

A strong NDVI signal indicates a high density of

plants, while a weak NDVI indicates problematic

areas in the field (Ahirwar et al., 2019, Biesel et al.,

2018). These NDVI reports are valuable for various

agricultural purposes as they differentiate between

areas where the crop is growing and areas where it is

not (Enouri et al. (2021). This differentiation enables

targeted fertilizer applications and also reveals the

presence of weeds, pests, and water damage.

Therefore, by mathematically combining these two

signals, it becomes possible to distinguish between

plants and non-plants, as well as healthy plants and

sickly plants (Stamford et al., 2023).

The Normalized Difference Vegetation Index (NDVI)

values typically range between -1 and +1, with higher

values indicating higher levels of chlorophyll content

in vegetation.

3.2 Soil Management

Soil plays a pivotal role in the success of agriculture,

serving as a vital source of nutrients essential for

optimal crop growth and development. It stores key

elements such as water, nitrogen, phosphorus,

potassium, and proteins, which are fundamental for

sustaining healthy plant growth (Eli-Chukwu and

Ogwugwam, 2019; Zha, 2020).

The Soil Risk Characterization Decision Support

System (SRC-DSS) is an AI-based system that

employs fuzzy logic to characterize soil and identify

contaminated soils posing accidental risks. Fuzzy

logic is advantageous for eliminating imprecision or

uncertainty in soil management data, as it requires

only a few parameters to quantify the soil (López et

al., 2008). Parameters that would typically take hours

to estimate can be easily measured by neural

networks. For example, the measurement of soil

hydraulic conductivity presents a direct challenge,

however, it can be accurately predicted by utilizing

measurable soil parameters such as soil texture data

(including sand and clay contents), soil water

retention curve (such as G retention model

parameters), bulk density, and effective porosity

(Ghanbarian-Alavijeh et al., 2010). Moreover,

through the utilization of artificial intelligence, it is

possible to predict biological parameters like soil

enzyme activity. Research studies have demonstrated

that when it comes to prediction, Artificial Neural

Networks (ANN) surpass Multiple Linear Regression

(MLR), and when combined with a Digital Terrain

Model (DTM), ANN offers a superior means of

mapping soil enzyme activity (Tajik et al., 2011).

PEAT, a Berlin-based agricultural technology

startup, has developed a state-of-the-art deep-learning

application known as Plantix. This groundbreaking

application is specifically designed to discern

potential anomalies and deficiencies in soil using a

sophisticated analysis. The analysis is conducted

through the utilization of advanced software

algorithms that establish correlations between distinct

foliage patterns and an array of soil defects, plant

pests, and diseases. By harnessing the capabilities of

the user's smartphone camera to capture images, the

image recognition application can identify plausible

anomalies. Consequently, users are furnished with

invaluable insights about soil rejuvenation

techniques, beneficial advice, and alternative

ICEISA 2024 - International Conference on ‘Emerging Innovations for Sustainable Agriculture: Leveraging the potential of Digital

Innovations by the Farmers, Agri-tech Startups and Agribusiness Enterprises in Agricu

130

remedies. The company asserts that its software

possesses the capability to swiftly identify patterns

with an impressive, estimated accuracy rate of up to

95 percent. Furthermore, the application is

conveniently accessible in most of the local languages

spoken in India (Kumar and Karthikeyan, 2019).

3.3 Pest and Disease Management

Crop diseases pose significant challenges for farmers,

but computer-aided systems offer effective solutions

for disease diagnosis and control. One approach

involves a fuzzy logic model designed to predict

diseases based on the duration of leaf wetness. In this

method, leaf images are pre-processed and segmented

into different areas, such as the background, non-

diseased part, and diseased part. The diseased part is

then sent to remote laboratories for further diagnosis.

Additionally, image processing techniques can be

applied for pest identification and recognition of

nutrient deficiencies. For more accurate and early

disease detection, a deep learning-enabled object

detection model has been developed. This model

focuses on multi-class plant disease detection,

particularly targeting different apple plant diseases in

real orchard environments. The effectiveness of this

model has been demonstrated in complex scenarios,

showcasing its potential for fine-grained and multi-

scale disease detection (Roy and Bhaduri, 2021).

The process of detecting plant diseases and pests

involves three distinct stages. The initial stage

revolves around classification, where the primary

focus is on determining the category of the image.

Subsequently, the second stage entails locating the

diseases and pests within the image, a critical step in

ensuring accurate detection. This stage, akin to the

location task in computer vision, not only identifies

the types of diseases and pests present but also

provides specific spatial information, often

represented by rectangular boxes highlighting the

affected areas, as seen in the case of gray mold.

Finally, the third stage resembles the segmentation

task in computer vision, where lesions caused by

diseases such as gray mold are meticulously

delineated from the background, pixel by pixel. This

detailed segmentation facilitates the extraction of

various attributes like length, area, and precise

location, which in turn contribute to a more

comprehensive evaluation of disease severity and

pest infestation levels. (Wang and Tao 2021).

Rothe and Rothe (2019) adopted a novel strategy

for image segmentation aimed at identifying bacterial

leaf blight, Myrothecium, and Alternaria. They

employed Otsu's segmentation technique to precisely

delineate the image of an infected leaf while retaining

its background, ensuring the accurate isolation of the

affected region from the leaf's natural backdrop. In

the subsequent classification stage, they utilized a

Backpropagation Neural Network (BPNN). Their

approach yielded impressive accuracy rates of

97.14% for Alternaria, 93.3% for bacterial leaf blight,

and 96% for Myrothecium, showcasing the

effectiveness of their methodology in disease

identification and classification.

3.4 Weed Management

Weeds hinder the appropriate growth of crops by

engaging in competition for light, moisture, and

nutrients, as well as by causing interference with

harvesting machinery. Additionally, they contribute

to health issues in both humans and animals and have

a detrimental impact on both the natural ecosystem

and aquatic resources (Ministry of Agriculture, Land

and Fisheries, 2020). Remote sensing techniques,

including the use of Unmanned Aerial Vehicles

(UAVs) and satellites, offer the ability to obtain high-

resolution imagery. This imagery is capable of

detecting weeds at an exceptionally detailed level due

to the advanced camera and spectral band employed.

Consequently, precise identification and monitoring

of weed patches can be achieved. Through the

combination of UAVs and remote sensing techniques,

a highly efficient and timely approach to field

scouting and weed management is made possible

(Ghatrehsamani et al., 2023). Correa et al. (2022)

have devised a neural network model named

RetinaNet, which exhibits remarkable proficiency in

forecasting the presence of the most invasive weeds

in tomato fields. The model accurately identified

98.4% of tomato plants and 0.91% of weed plants.

The implications of this model are significant,

particularly in relation to site-specific weed

management and the implementation of robotic

technology. By doing so, the use of herbicides can be

reduced in comparison to conventional farming

practices. Deep learning techniques were assessed to

identify weeds in bell pepper fields, using RGB

images. The selected models exhibited an overall

accuracy ranging from 94.5% to 97.7%. This

indicates the potential integration of these models

with image-based herbicide applicators, thereby

enabling precise weed management (Subeesh et al.,

2022).

The robot integrated two visual systems. The

initial system utilized gray-level vision to construct a

row structure, guiding the robot through rows. The

second system relied on color-based vision, essential

Leveraging Artiﬁcial Intelligence for Agricultural Advancement: A Comprehensive Review

131

for distinguishing individual weeds from others. The

VIIPA (Variable Injection Intelligent Precision

Applicator), an automated weed-killing robot,

administers the precise dosage of herbicide required

to manage weeds on the farm swiftly (Wu et al.2020).

3.5 Water Use Management

Automation in agriculture is increasingly gaining

importance on a global scale. By incorporating

sensors such as moisture, temperature, and humidity,

alongside IoT devices and machine learning

techniques, irrigation systems can be automated to

cater to the specific needs of various crops, soil types,

and climate conditions.

The incorporation of artificial intelligence (AI)

methods into hydro-meteorological forecasting has

significantly progressed the prediction of extreme

weather events such as floods and droughts, leading

to enhanced preparedness and mitigation strategies

(Zhu et al.2023). The creation of the WaterWise

platform aimed at efficiently managing and analyzing

data from an intelligent Internet of Things (IoT)

system, facilitating tasks like online water leakage

detection, water demand prediction, and water quality

assessment (Allen et al., 2018). This platform

employs a layered model consisting of application,

information communication, and device perception

layers to establish an effective water supply

management system for automating domestic water

management (Marjani, 2017). A smart IoT system

equipped with sensors for water flow, pH, and other

parameters, controlled by a Raspberry core controller

and accessible through a web interface, ensures

efficient control and monitoring of water storage

systems (Shah, 2017).

In agricultural irrigation management, sensors

capture environmental and meteorological data

transmitted to a cloud server database, enabling

farmers to remotely adjust water valves and other

controls based on trends in soil, plant, and weather

data (Chen et al., 2021). Vellidis et al. (2008) devised

a prototype for scheduling irrigation to cotton fields

based on soil moisture and temperature, utilizing

smart nodes for real-time monitoring and irrigation

scheduling (Bisaria et al., 2019). The Korea Water

Resources Corporation, in collaboration with the

International Water Resources Association (IWRA),

implemented a smart water management system

utilizing Information and Communications

Technology (ICT) to provide real-time water data

globally for intelligent resolution of water issues,

including water quality monitoring, efficient

irrigation, leak detection, and flood management

using AI mechanisms (Rocher, 2018). Additionally, a

layer of irrigation architecture supported by machine

learning and digital farming solutions was developed,

incorporating data from UAVs, satellites, soil, and

weather stations to offer predictive recommendations

for irrigation decisions and scheduling (Abioye et al.,

2022).

3.6 Weather Forecasting (Predictive

Analytics)

With the advent of climate change, the significance of

forecasts for crop yields has increased as farmers are

no longer able to rely solely on traditional knowledge.

The utilization of more precise forecasts could

empower farmers to select the most favorable days

for planting or harvesting. Artificial intelligence

techniques, specifically reinforcement learning, are

employed to analyze previous predictions and

compare them with actual outcomes. To enhance

weather forecasting, data is inputted into an algorithm

that utilizes deep learning techniques to acquire

knowledge and generate predictions based on

historical data. Artificial intelligence in the realm of

farming, in conjunction with satellite data, can be

employed to forecast weather conditions, evaluate the

sustainability of crops, and assess the presence of

pests and diseases on farms. The implementation of

Artificial Intelligence (AI) in farming enables the

provision of an extensive volume of data points,

encompassing temperature, precipitation, wind

speed, and solar radiation. Pierre et al. (2023) devised

Fuzzy logic, a machine learning technique, to forecast

weather and predict the timing and quantity of rain

expected within a 24-hour period based on dynamic

changes in air temperature, air humidity, atmospheric

pressure, and wind speed.

To tackle the challenge of determining

appropriate sowing timings for crops in India,

particularly considering the risks associated with

droughts and excessive rainfall, Microsoft

collaborated with the International Crops Research

Institute for the Semi-Arid Tropics (ICRISAT) to

create an AI Sowing App. This application leverages

machine learning and business intelligence from the

Microsoft Cortana Intelligence Suite. A notable

aspect of this app is its accessibility, as farmers only

require a basic feature phone capable of receiving text

messages, without the need for installing sensors or

incurring additional expenses (ICRISAT, 2017). To

determine the optimal sowing period, historical

climate data spanning three decades (from 1986 to

2015) for the specific region of Andhra Pradesh was

analyzed using AI techniques. The Moisture

ICEISA 2024 - International Conference on ‘Emerging Innovations for Sustainable Agriculture: Leveraging the potential of Digital

Innovations by the Farmers, Agri-tech Startups and Agribusiness Enterprises in Agricu

132

Adequacy Index (MAI) was calculated as part of this

analysis to identify the most favorable sowing

window. The MAI is a standardized measure used to

assess how effectively rainfall and soil moisture meet

the water requirements of crops, providing valuable

insights for farmers (Sure and Dikshit 2019).

3.7 ChatGPT: Concerns in

Agricultural Extension

A recent IFPRI blog explores using chatbots powered

by large language models (LLMs) to aid agricultural

extension services, but real-world testing in Nigeria

highlighted significant challenges. LLM-generated

recommendations for cassava farmers lacked

specificity and practicality for smallholders. To

address these issues, a collaborative approach is

needed, involving AI and farming experts, ensuring

accessible data and local knowledge, employing user-

centered design, and enhancing digital literacy among

farmers and extension agents (Jawoo Koo, 2023).

A comparative analysis between two AI chatbots,

ChatGPT and Claude was done to understand their

ability to provide agricultural advice specifically

tailored to the context of wheat farming in Faisalabad,

Pakistan, in both English and local languages. The

evaluation covers their responses regarding the best

time to plant wheat and recommendations for wheat

varieties, as well as their performance in Urdu,

Pashto, Sindhi, and Punjabi languages.

In the English language, both ChatGPT and

Claude provided generally correct information about

the optimal timing for planting wheat in Faisalabad.

However, Claude demonstrated a deeper

understanding of specific wheat varieties suitable for

the region, including newer varieties, whereas

ChatGPT's response was more generic and

noncommittal. When tested in local languages such

as Urdu, Pashto, Sindhi, and Punjabi, both chatbots

struggled to provide accurate and contextually

relevant responses, with varying degrees of success.

While Urdu responses were somewhat satisfactory,

responses in other local languages were less accurate

or outright incorrect. AI chatbots like ChatGPT and

Claude show promise in providing agricultural

advice, but they are not yet ready to serve as

agricultural extension agents, especially in areas with

limited data and content in local languages. However,

it suggests the potential for customized agricultural

advisory chatbots trained on localized content to

better serve farmers in specific regions (AESA, 2021)

Farm Radio International (FRI), in collaboration

with CGIAR, has developed an AI-based solution,

Uliza, to analyze voicemails from over 12 million

listeners of radio shows in rural Africa. This initiative

aims to harness farmers' knowledge and perspectives,

addressing challenges in analyzing the large volume

of responses. The solution utilizes transfer learning in

machine learning models to adapt to local languages

and dialects. Moving forward, the team plans to

further improve transcription accuracy and apply the

solution to new projects, aiming to evaluate its impact

on food security, gender equality, and livelihood

outcomes in 2024 (CIMMYT and IFPRI, 2022).

Acceso is piloting ExtensioBot, an AI tool, to

enhance agricultural extension services for

smallholder farmers in Latin America and the

Caribbean. The decision stemmed from user demand

and the need to address the scarcity of extension

agents. ExtensioBot utilizes Azure AI Speech and

Vision technologies to provide personalized advice

and identify pests and diseases. While still learning,

the tool aims to improve engagement and

effectiveness by addressing common issues such as

generic responses and language barriers.

Collaborative efforts and ongoing training are

prioritized to refine the tool and ensure equitable

access for all farmers, ultimately revolutionizing

extension models to better support smallholders

globally (Mendoza,2024).

Limitations of ChatGPT

The knowledge possessed by ChatGPT is confined to

data from the year 2021. As the data is derived from

the cloud, it cannot exclusively acquire information

from a specialized source. The true beauty of

ChatGPT lies in its capability to derive data from

various sources and generate predominantly unique

responses to the same query. Should we depend on

ChatGPT to gather information from a particular

source, it would not surpass the competence of any

expert system or limited-information mobile

application. While it can compensate for the scarcity

of staff by delivering a comprehensive information

resource to users, it is ultimately up to the user's

discretion to harness the potential of this technology

in the most optimal manner and to exercise caution

before applying the acquired knowledge.

Leveraging Artiﬁcial Intelligence for Agricultural Advancement: A Comprehensive Review

133

Table 1: Digital AI Startups- Implications to Redesigning

of Extension Advisory Services

Digital

start-ups

Description Scope of

integration in

Extension

Advisory Services

Farn2fork 1. Water

monitoring

solutions for better

productivity by

using IoT wireless

soil sensors and

real-time

analytics.

2. Farmers are

contacted via

farmer

associations and

network

1. Predictive

analytics

2. Process

automation

3. Forewarning

advisories

4. Social networks

Aibono 1. Utilizing AI and

data analysis

expertise, it

gathers an

extensive volume

of intelligent

agricultural data

and insights from

farmers and

industry

professionals.

1. Personalization

2. Predictive

analytics

3. Process

automation

Fasal 1.End-to-end

farming app for

horticulture

farmers

2.Assist customers

in making the best

farmin

decisions.

1. Predictive

analytics

2. Process

automation

3. Forewarning

advisories

Agrostar 1. Developed an

m-commerce

platform called

"Direct to Farmer"

to revolutionize

the agribusiness

industry.

2.By simply

giving a missed

call on the toll-free

number 1800,

farmers can

conveniently

obtain agri-inputs

delivered right to

their doorstep.

1. Aggregates

services

2.Access to inputs

3. Supply chain

SatSure 1. IoT and big data

are employed

proficiently to

bestow financial

security upon

farmers, through

the utilization of a

database that spans

15 years and

comprises satellite

images.

2. Furthermore,

recommendations

regarding

clustering

techniques are

made to farmers,

to obtain an

approximation of

the aggregate

agricultural

production. This

data is

subsequently

furnished to agri-

insurance

companies.

1. Financial

inclusion credit

insurance

2. Aadhar-enabled

services

3. Predictive

analytics

Farm

Again

1. IoT devices are

employed for

monitoring and

documenting the

levels of moisture

and soil

conditions,

alongside conduits

facilitating the

provision of water

and fertilizer

inputs.

2.A vast expanse

of 2500 acres of

land has been

transformed into

organic farming

systems.

1. Access to

market

2. Process

automation and

forewarning

advisories

3. Predictive

analytics

Challenges in Developing and Adopting AI in

Agriculture

AI systems require a significant amount of data to

train machines for accurate forecasting or predictions,

with spatial data being relatively easier to collect

compared to temporal data, which poses a challenge

in agricultural settings. One of the main obstacles

ICEISA 2024 - International Conference on ‘Emerging Innovations for Sustainable Agriculture: Leveraging the potential of Digital

Innovations by the Farmers, Agri-tech Startups and Agribusiness Enterprises in Agricu

134

hindering the adoption of AI in agriculture is the lack

of access to technology and infrastructure for many

farmers. Limited access to AI tools like sensors,

drones, and satellite imagery, especially in remote

areas with restricted internet connectivity, hampers

real-time data retrieval. The rise of AI in agriculture

also brings forth concerns regarding data privacy and

security. While extensive datasets are vital for AI

operations, they also expose farmers to cyber threats,

necessitating measures to protect sensitive

information. Ethical considerations arise regarding

AI's potential to widen socioeconomic disparities, as

access to AI technologies could create inequalities

among farmers. Moreover, ethical concerns emerge

regarding AI's impact on job displacement in

agriculture, as automation may replace traditional

farming roles, particularly in rural communities

heavily reliant on agriculture for employment.

REFERENCES

Abioye, E.A., Hensel, O., Esau, T.J., Elijah, O., Abidin,

M.S.Z., Ayobami, A.S., Yerima, O., and Nasirahmadi,

A. 2022. Precision Irrigation Management Using

Machine Learning and Digital Farming Solutions.

AgriEngineering, 4: 70-103.

https://doi.org/10.3390/agriengineering4010006

AESA [Agricultural Extension in South Asia]. 2021.

Available: https://www.aesanetwork.org/blog-140-ai-

enabled-solutions-in-indian-agriculture-the-way-

forward/

Ahirwar, S., Swarnkar, R., Bhukya, S., and Namwade, G.

2019. Application of Drone in Agriculture.

International Journal of Current Microbiology and

Applied Sciences, 8: 2500-2505.

Allen, M., Preis, A., Iqbal, M., and Whittle, A.J. 2013.

Water distribution system monitoring and decision

support using a wireless sensor network. In

Proceedings of the 2013 14th ACIS International

Conference on Software Engineering, Artificial

Intelligence, Networking and Parallel/Distributed

Computing, pages 641–646, Honolulu, HI, USA.

Barakabitze, A.A., Jonathan, J., and Sanga, C. 2023. AI-

driven Internet of Agro-Things adaptive farm

monitoring systems for future agricultural production

and food systems in Africa. Tanzania Journal of

Agricultural Sciences, 22(2): 22-33.

Bisaria, K., Sinwar, D., and Sharma, M.K. 2019. February.

A Study on Real-Time Face Recognition Using

Feature-Based Image Processing Frameworks. In

Proceedings of International Conference on

Sustainable Computing in Science, Technology and

Management (SUSCOM), Amity University Rajasthan,

Jaipur-India.

Chang, C.L., Chung, S.C., Fu, W.L., and Huang, C.C. 2021.

Artificial intelligence approaches to predict growth,

harvest day, and quality of lettuce (Lactuca sativa L.) in

an IoT-enabled greenhouse system. Biosystems

Engineering, 212:.77-105.

Chen, Y.A., Hsieh, W.H., Ko, Y.S., and Huang, N.F. 2021.

An ensemble learning model for agricultural irrigation

prediction. (2021). In Proceedings of the 2021

International Conference on Information Networking

(ICOIN), pages 311–316, Jeju Island, Korea.

Chlingaryan, A., Sukkarieh, S., and Whelan, B. 2018.

Machine learning approaches for crop yield prediction

and nitrogen status estimation in precision agriculture:

A review. Computers and Electronics in Agriculture,

151: 61–69.

CIMMYT [The International Maize and Wheat

Improvement Center] and IFPRI [International Food

Policy Research Institute]. 2022. Artificial intelligence

enables extension services reaching 12 million farmers

in sub-Saharan Africa to respond to farmers’ voices.

https://www.cgiar.org/initiative-result/artificial-

intelligence-enables-extension-services-reaching-12-

million-farmers-in-sub-saharan-africa-to-respond-to-

farmers-voices/s

Eli-Chukwu, N.C. 2019. Applications of artificial

intelligence in agriculture: A review. Engineering,

Technology & Applied Science Research, 9(4): 4377-

4383.

Ennouri, K., Smaoui, S., Gharbi, Y., Cheffi, M., Ben

Braiek, O., Ennouri, M., and Triki, M.A. 2021. Usage

of artificial intelligence and remote sensing as efficient

devices to increase agricultural system yields. Journal

of Food Quality, pp.1-17.

FAO [Food and Agriculture Organization of the United

Nations]. 2015. FAO home page [On-line]. Available:

http://www.fao.org/family-

farming/detail/en/c/897026/ [23 Dec 2019]

Ghanbarian-Alavijeh, B., Liaghat, A., Huang, G.H. and

Van Genuchten, M.T. 2010. Estimation of the van

Genuchten soil water retention properties from soil

textural data. Pedosphere, 20(4): 456-465.

Ghatrehsamani, S., Jha, G., Dutta, W., Molaei, F., Nazrul,

F., Fortin, M., Bansal, S., Debangshi, U. and Neupane,

J. 20223). Artificial intelligence tools and techniques to

combat herbicide-resistant weeds—A review.

Sustainability, 15(3): 1843.

ICRISAT [International Crop Research Institute for Semi

Arid Tropics]. 2017. ihub: A creative platform for

innovations that change the lives of farmer. ICRISAT,

Hyderabad, 4p.

Jawoo Koo. 2023. Can we trust AI to generate agricultural

extension advisories? https://www.ifpri.org/blog/can-

we-trust-ai-generate-agricultural-extension-advisories

Kolhe, S. R., Kamal, H. S., Saini, G. K., and Gupta. 2011.

“An intelligent multimedia interface for fuzzy logic

based inference in crops”. Expert Systems with

Applications, 38(12): 14592-14601.

Kumar, S.A., and Karthikeyan, C. 2019. Status of Mobile

Agricultural Apps in the Global Mobile Ecosystem

using ICT. International Journal of Educational

Development, 15(3).

Leveraging Artiﬁcial Intelligence for Agricultural Advancement: A Comprehensive Review

135

Kumar, S., Singh S.K., Occluded thermal face recognition

on using a bag of CNN(BoCNN). 2020. IEEE Signal

Process Letters, 27:975–979.

Liakos, K.G., Busato, P., Moshou, D. Pearson, S., and

Bochtis, D. 2018. Machine learning in agriculture: A

review. Sensors, 18: 2674.

López, E.M., García, M., Schuhmacher, M., and Domingo,

J.L. 2008. A fuzzy expert system for soil

characterization. Environment International, 34(7):

950-958.

Correa, L.J.M., Moreno, H., Ribeiro, A., and Andújar, D.

2022. Intelligent weed management based on object

detection neural networks in tomato crops. Agronomy,

12(12): 2953.

Marjani, M., Nasaruddin, F., Gani, A., Karim, A., Hashem,

I.A.T., Siddiqa, A., and Yaqoob, I. 2017. Big IoT data

analytics: Architecture, opportunities, and open

research challenges. IEEE Access, 5: 5247–5261.

McCarthy, J., Minsky M.L., Rochester, N., Shannon C.E.

1955. A Proposal for the Dartmouth Summer Research

Project on Artificial Intelligence, 27(4):12-12.

McKinion, J.M., and Lemmon, H.E. 1985. Expert systems

for agriculture. Computers and Electronics in

Agriculture, 1(1): 31-40.

Melnyk, P., You Z., and Li K. 2019. A high-performance

CNN method for offline handwritten Chinese character

recognition and visualization. Soft Computing,

24:7977–87.

Mendoza, G. 2024. Revolutionizing Extension Models with

Artificial Intelligence in Service of Smallholder

Farmers. https://agrilinks.org/post/revolutionizing-

extension-models-artificial-intelligence-service-

smallholder-farmers

Ministry of Agriculture, Land and Fisheries. 2020. What is

a Weed? Information courtesy of the Extension

Training and Information Services (ETIS) Division.

https://agriculture.gov.tt/publications/what-is-a-weed/.

Pierre, N., Viviane, I.V.I., Lambert, U., Shadrack, I.,

Erneste, B., Schadrack, N., Alexis, N., Francois, K., and

Theogene, H. 2023. AI-Based Real-Time Weather

Condition Prediction with Optimized Agricultural

Resources. European Journal of Technology, 7(2): 36-

49.

Rocher, V. 2018. Smart Water Management—Case Study

Report; K-Water: Daejeon, Korea.

Rothe, P., and Rothe, J. 2019. Intelligent pattern recognition

system with application to cotton leaf disease

identification. Innovations in Computer Science and

Engineering, 32: 19–27.

Roy, A.M. and Bhaduri, J., 2021. Deep learning enabled

multi-class plant disease detection model based on

computer vision. Ai, 2(3), pp.413-428.

Shah, J. 2017. An Internet of Things-based model for smart

water distribution with quality monitoring.

International Journal of Innovative Science

Engineering and Technology, 6: 3446–3451.

Sharma, A., Sharma, A., Tselykh, A., Bozhenyuk.,

Choudhury, T., Alomar, M.A., and Sánchez-Chero, M.

2023. Artificial intelligence and internet of things

oriented sustainable precision farming: Towards

modern agriculture. Open Life Sciences, 18(1):

20220713.

Stamford, J.D., Vialet-Chabrand, S., Cameron, I., and

Lawson, T. 2023. Development of an accurate low cost

NDVI imaging system for assessing plant health. Plant

Methods, 19(1): p.9.

Subeesh, A., Bhole, S., Singh, K., Chandel, N.S., Rajwade,

Y.A., Rao, K.V.R., Kumar, S.P. and Jat, D. 2022. Deep

convolutional neural network models for weed

detection in polyhouse grown bell peppers. Artificial

Intelligence in Agriculture, 6: 47-54.

Sundararajan, S. K., Sankaragomathi, B., Priya, D. S. 2019.

Deep belief CNN features representation-based

content-based image retrieval for medical images.

Journal of Medical Systems.43(6):1–9.

Sure, A., and Dikshit, O. 2019. Estimation of root zone soil

moisture using passive microwave remote sensing: A

case study for rice and wheat crops for three states in

the Indo-Gangetic basin. Journal of Environmental

Management, 234: 75-89.

Tajik, N., Peng, Z., Kuyanov, P., and LaPierre, R.R. 2011.

Sulfur passivation and contact methods for GaAs

nanowire solar cells. Nanotechnology, 22(22): 225402.

Vellidis, G., Tucker, M., Perry, C., Kvien, C., and Bednarz,

C. 2008. A real-time wireless smart sensor array for

scheduling irrigation. Computers and electronics in

agriculture, 61(1): 44-50.

Wang, L., Liu, Z., Liu, A., and Tao, F. 2021. Artificial

intelligence in product lifecycle management. The

International Journal of Advanced Manufacturing

Technology, 114: 771-796.

Wu, X., Aravecchia, S., Lottes, P., Stachniss, C., and

Pradalier, C. 2020. Robotic weed control using

automated weed and crop classification. Journal of

Field Robotics, 37(2): 322-340.

Yang. D., Li, S., Peng, Z., Wang P., Wang J., and Yang H.

2019. MF-CNN: traffic flow prediction using

convolutional neural network and multi-features fusion.

IEICE Transactions on Information and Systems,

102(8):1526–36.

Zha, J. 2020. Artificial intelligence in agriculture. In

Journal of Physics: Conference Series, 1693 (1):

012058. IOP Publishing.

Zhu, D., Zhou, Y., Guo, S., Chang, F.J., Lin, K., and Deng,

Z. 2023. Exploring a multi-objective optimization

operation model of water projects for boosting

synergies and water quality improvement in big river

systems. Journal of Environmental Management, 345:

118673.

ICEISA 2024 - International Conference on ‘Emerging Innovations for Sustainable Agriculture: Leveraging the potential of Digital

Innovations by the Farmers, Agri-tech Startups and Agribusiness Enterprises in Agricu

136