Water Optimization in Digital Farming

Pascal Francois Faye

1 a

, Jeanne Ana Awa Faye

and Mariane Senghor

Dept. Mathematics and Computer Sciences,

University Sine Saloum El Hadj Ibrahima NIASS (USSEIN), Kaolack, Senegal

Institut de Management Public et Gouvernance Territoriale,

University Aix-Marseille, Marseille, France

Keywords:

Water Optimization, Internet of Things, Data Analysis, Artiﬁcial Intelligence, Crop Similarities.

Abstract:

In Senegal, agriculture is subsistence and highly dependent on soil, climate, and raining season. Food crops

take up to 46% of the total land and make up 15% of the Gross Domestic Product (GDP), ensuring between

70% and 75% employment. In this work, we propose methods to understand through a sensor network, the

effects of the required irrigation system on six soil types (ferruginous tropical - sandy - loamy - clay - humus-

bearing - clay and loamy) depending to crop production like : - the time interval for inﬁltration or evaporation

of the irrigation water according to the type of soil - the speed of spreading of water in both directions (lateral

and depth) - the set up of four soil’s amendments (peanut shells, livestock manure, poultry manure and plant

mixture) methods for optimized water in crop production. We, also, propose an agricultural calendar for

a good distribution of the farms’ activities over time after ﬁnding the relationship between eighteen crop

production and soil amendments. Our results show the effectiveness of our solution to help water optimization

in agriculture. This means that, taking into account these data, it is possible to understand crop dependencies,

anticipate agro-ecological phenomena and crop water stress that affect the yield of crops.

1 INTRODUCTION

In order to continuously feed humanity, in computer

sciences it has several advances on sensor network,

robot automation, computer vision and artiﬁcial intel-

ligence for prediction, decision making, etc. How-

ever, due to the demographic pressure on cities, cli-

mate change and soil degradation, several research

questions are done in order to face with the gradual

abandonment of land and to ﬁnd solutions or advice

with the reduction of agricultural perimeters. This

is one of the big issues in the peanut basin of Sene-

gal(cf. ﬁgure 1). Computer vision recognition has

been increasingly applied to numerous ﬁelds of agri-

culture with the advancement of computer graphics

and image processing technology. The development

of sensor technologies for smart agriculture, such as

soil and air all-in-one sensors which provide tempera-

ture and humidity parameters, etc., enhance data col-

lection and processing for decision making in agri-

culture. Usually, the data is transferred from the sen-

sor to the sink node for data collection and the server

for processing and decision making. In addition, the

https://orcid.org/0000-0002-2078-5891

use of robots in agriculture that uses a variety of sen-

sors to sense the dynamic of the agricultural environ-

ment and then picks the target using this knowledge

and a decision-making algorithm based on artiﬁcial

intelligence help farms in order to manage their ﬁeld

(ploughing, sowing, harvesting, pest attack, etc.).

In this work we tackle through sensors, the ef-

fects of soil type and the required irrigation system

depending to crop production through - the time in-

terval for inﬁltration or evaporation of irrigation water

depending on the type of soil - the speed of spreading

of water in both directions (lateral and depth) - the

soil amendment methods for optimized water in crop

production. This, to propose an agricultural calendar

for a good distribution of the farms’ activities over

time and ﬁnd the relationship between crop produc-

tion and soils. Thus, to lead to a better planning of

the agricultural calendar for a good distribution of ac-

tivities over time. This in order to take into account

climate changes, soil degradation, etc. for improving

the farms’ crops and to overcome the seasonality of

the agriculture in the peanut basin of Senegal.

In the rest of this paper, section 2 presents our re-

search problem, section 3 highlights our objectives

and section 4 discusses related works. Section 5

472

Faye, P., Faye, J. and Senghor, M.

Water Optimization in Digital Farming.

DOI: 10.5220/0013057400003822

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

In Proceedings of the 21st International Conference on Informatics in Control, Automation and Robotics (ICINCO 2024) - Volume 1, pages 472-479

ISBN: 978-989-758-717-7; ISSN: 2184-2809

Figure 1: Peanut basin of Senegal (ANSD, 2024).

presents some preliminaries and methodologies. Sec-

tion 6 highlights the results before the conclusion in

the section 7.

2 RESEARCH PROBLEM

An agriculture that deals with the environment, the

climate change and the soil degradation has become

an imperative if we aim to make farmer and agri-

culture sustainable. A better understanding of the

farm’s environment is needed to make the right va-

rietal choices and crop options. This is not to men-

tion the rise in temperatures, which leads to increased

evaporation or evapotranspiration, which inﬂuences

yields and the seasonality of the agriculture. In ad-

dition, water is a critical resource that needed to op-

timize for an efﬁcient crop production. That is why,

this work focus on methods to understand, the time

interval for water inﬁltration or evaporation depend-

ing on the soil type and amendment. This, in order

to propose a better agricultural calendar for a good

distribution of the farms’ activities over time and ﬁnd

the relationship between crops and soil activities in

the peanut basin of Senegal (cf. ﬁgure 1). Indeed, we

highlight on the ﬁgure 2, the dependencies between

month, from 2000s to 2023s regarding the agrocli-

matology parameters (Surface Pressure, Temperature,

Humidity, Wind Speed, Surface Soil Wetness, Pro-

ﬁle Soil Moisture, Root Zone Soil Wetness, Precip-

itation, Photosynthetically Active Radiation). Each

square shows the correlation (a measure of dependen-

cies). Values closer to zero means there is no linear

trend between the months. Close to 1 the months are

more positively correlated, and stronger is the rela-

tionship. The legend on the right side help to interpret

correlations.

The results show a strong correlation between the

months in the peanut basin. It is therefore possible to

Figure 2: The result of our data correlation from 2000s to

2023s in the peanut basin.

reﬁne cropping calendars to avoid seasonality in agri-

culture, taking into account other soil-related param-

eters.

3 OUTLINE OF OBJECTIVES

Knowledge of climate change and its effects on the

various sectors of the national economy is a major

challenge for the country’s developers. Various ini-

tiatives are therefore being developed to better iden-

tify the implications of climate variability in the agri-

culture sector. However, while a causal relationship

has clearly been established between the vulnerabil-

ity of the agriculture sectors and a set of parame-

ters like: the climate, the challenge of accurate infor-

mation, the sharing and pooling efforts (Faye et al.,

2022). This calls for a review of existing frameworks,

but it should also help to ﬁnd ways of collaborating,

sharing and other one in order to provide better sup-

port for decision-making, particularly for grassroots

users such as producers. Supporting farmers to better

manage the risks associated with climate variability is

now a major necessity (Faye et al., 2024a)(Faye et al.,

2024b). All economic activities which promote food

security and suitable agriculture must incorporate the

risks of climate change, soil types and water quantity

and water quality into their planning. The aims of this

work are:

1. to understand the time interval for inﬁltration or

evaporation of irrigation water depending on the

type of soil.

2. to calculate the speed of spreading water in both

directions (lateral and depth).

3. to set up several soil amendment methods for op-

timized crop production.

4. to help with water optimization that is adapted

to variations of agrometeology parameters during

Water Optimization in Digital Farming

473

the annual seasons.

5. to ﬁnd similarities between different crops in or-

der to propose an annual soil occupation manage-

ment strategy.

In this way, farmers can make decisions about the

technical itineraries for their crops and the dates set

for the cropping calendars (ploughing dates, sowing

dates, fertiliser application dates, irrigation hours and

other inputs), which can enable precision farming.

This work combines the:

1. laws of probability to model the dynamics and un-

predictable events;

2. Machine learning (ML) models to ﬁnd the better

decision making;

This combination delivers a solution that addresses

well the dynamism and uncertainty challenges tar-

geted in this work.

4 STATE OF THE ART

In (Sellam and Poovammal, 2010), they research the

environmental parameters that affect the crop yield

and related parameters through a multivariate Regres-

sion Analysis. A sample of environmental factors

considers a period of 10 years. The System is applied

to ﬁnd the relationship between explanatory variables

like annual rainfull,area under cultivation, food price

index and hence the crop yield as a response vari-

able and R

value clearly shows that, the yield is es-

pecially hooked into annual rainfall,area under cul-

tivation and food price index are the opposite two

factors that are inﬂuencing the crop yield. This re-

search is often enhanced by considering other fac-

tors like minimum support price, cost price index,

wholesale price index, etc. and their relationship with

crop yield. In (Zhang et al., 2010), the authors show

that NDVI (Normalized Difference Vegetation Index)

and precipitation are the most important predictors for

corn yield in Iowa. In (Zingade et al., 2018), the au-

thors have presented an android based application and

an internet site that uses Machine learning methods

to predict the foremost proﬁtable crop in the current

weather and soil conditions and with current envi-

ronmental conditions. This system helps the farmer

with a sort of option for the crops that will be culti-

vated, which will be helping them over the long run.

In (Sun et al., 2022) they improved a density peak

cluster segmentation algorithm for RGB (Red Green

Blue) images with the help of a gradient ﬁeld of depth

images to locate and recognize target fruit during the

process of green apple harvesting or yield estimation.

Speciﬁcally, the image depth information is adopted

to analyse the gradient ﬁeld of the target image. Usu-

ally, smart agriculture produces enormous quantities

of multidimensional time series data. However, due

to the technological’s limitations, data loss and mis-

representation are frequent problems with the smart

agriculture’s IOT devices. In order to solve the issues

in (Cheng et al., 2022), authors propose a anomaly

detection model that can handle these multidimen-

sional time series data. Meanwhile, a multi-objective

strategy based on supervised machine learning was

utilized in (Uyeh et al., 2022) to identify the ideal

number of sensors and installation locations in a pro-

tected cultivation system. A machine learning tree-

based model in the form of a gradient boosting tech-

nique was speciﬁcally adapted to observed (temper-

ature and humidity) and derived circumstances (dew

point temperature, humidity ratio, enthalpy, and spe-

ciﬁc volume). Time series forecasting was used for

feature variables. In (Maia et al., 2022), sensor data

analysis over two irrigation seasons in three cotton

ﬁelds from two cotton-growing regions of Australia

revealed a connection between soil matric potential

and cumulative crop evapotranspiration (ETcn) de-

rived from satellite measurements between irrigation

events. They explore the distributed averaging is-

sues of agriculture picking multi-robot systems un-

der directed communication topologies by utilizing

the sampled data. A distributed protocol based on

nearest-neighbor information is presented using the

principles of algebraic graph theory and matrix the-

ory.

In contrast to our work, these works do not pro-

pose a cropping calender in order to minimize the risk

depending to climate change, pest migration and other

agrometeorological information and soils parameters

like ours. In addition, we aim to enhance productivity

and sustainability in the peanut basin of Senegal by

taking into account the water optimization and crop-

ping calendar. This, because in Senegal, there are dif-

ferent types of farms and various levels of complex-

ities in terms of organization. We have on the one

hand, family farms with limited levels of organiza-

tions, ﬁnancial capabilities and standard procedures.

On the other hand, there are some farms working on

fruits and vegetables exportation with better organi-

zations and procedures. In (Faye et al., 2023), a set

of results in this peanut basin shown the crop yield

relative to the raining season and to the temperature.

In (Faye et al., 2024a), authors propose an Agricul-

ture Information and Management System based on

some Machine Learning Algorithm (ML) and Inter-

net Of Things device that ensures data collection and

control as well as a data monitoring system via a web

platform for decision-making support in a real-world

ICINCO 2024 - 21st International Conference on Informatics in Control, Automation and Robotics

474

agricultural environments. (Faye et al., 2024b) pro-

vides a set of mechanisms that uses a set of trust

database of agro-climatic parameters and a set of ar-

tiﬁcial intelligence algorithm in order to assess agri-

cultural calendar for a good distribution of the farm’s

activities and to ﬁnd the relationship between crops.

In the continuity of their work, we take into account

soils’ types and water optimization.

The next section provides the methodology of this

work.

5 METHODOLOGY

Our study is motivated by the environmental and cli-

mate challenges that make difﬁcult the prediction on

crop yield, crop diseases and pests in the peanut basin

of Senegal (Fatick, Kaolack and Kaffrine) (cf. Fig-

ure 1). The peanut basin of Senegal has a set of

different climatic characteristics (Faye et al., 2023).

Many challenges are observable due to changes in

land degradation, soil salinization, temperature, ran-

dom rainfall, etc. We conduced a two years study that

has shown the correlation relationship (Figure 3) be-

tween a set of climatic parameters of the peanut basin

by using our dataset from our sensor network (ﬁgure

4).

We follow information on the agriculture require-

ments for planning work, forecasting the development

risks of certain climate-related diseases, monitoring

the water balance of soils, monitoring of temperatures

in connection with the plant development schedule.

The functioning of the crop, the soil and the water

system depends mainly on:

• PS : Surface Pressure (kPa)

• T2M : Temperature at 2 Meters (C)

• QV2M : Speciﬁc Humidity at 2 Meters (g/kg)

• WS2M : Wind Speed at 2 Meters (m/s)

• GWETTOP : Surface Soil Wetness (1)

• GWETPROF : Proﬁle Soil Moisture (1)

• GWETROOT : Root Zone Soil Wetness (1)

• PRECTOTCORR : Precipitation Corrected

(mm/day)

• PAR : Photosynthetically Active Radiation

(W/m

In ﬁgure 3, we have plotted changes in soil parame-

ters between 2002 and 2023 in order to complement

the correlation relationships provided by the ﬁgure 2.

It shows, between 2022 and 2023, the different rela-

tionships and differences between the various agrocli-

matic soil parameters that can inﬂuence plants apart

from soil nutrients.

Figure 3: Our data clustering between 2022 and 2023 using

DecisionTreeClassiﬁer.

Figure 4: Our sensors network topology for local climate

and soil data.

This, in order to study the set of real-time inter-

actions between atmospheric phenomena and all of

agronomic parameters, to deal with the real needs of

farmers in order to avoid the seasonality of their ac-

tivities without to guarantee a good crop yield. Our

sensor network topology is shown in ﬁgure 4.

The data are collected via sensors (e.g. soil mois-

ture, NPK, DHT11, temperature, etc.) with their

controller cards (ESP 8266, ESP CAM, ARDUINO

UNO) which create a mesh network of ﬁeld sensors

that will serve as a medium for transmitting data from

sensors. Different cards transmit their data through

the access points of the network. The data are sent

simultaneously to the ThingSpeak platform and on

our local server via HTTPS (Hypertext Transfer Pro-

tocol Secure) and SQL (Structured Query Language)

requests. In our local server, it has a web applica-

Water Optimization in Digital Farming

475

tion for data visualization to support decision making.

ThingSpeak is a cloud IoT analytics platform service

that allows to aggregate, visualize, and analyze live

data streams. Data are sent to ThingSpeak from our

devices, to create instant visualization. All data from

our local server or from ThingSpeak are aggregated

by our agent platform. An agent is a device or an ap-

plication which can sense the environment compute

some processes and provide results or acts on its envi-

ronment. Our agent concept which extract, organize,

aggregate and interpret sensor data based on our ma-

chine learning models is constrained by the parame-

ters: {Rs, Hs, ϑ

, U, L

Net

}. Rs is its resource(s) and Hs

is its history set which consists of a set of previous de-

cisions. A view ϑ

is the set of sensors in its neighbor-

hood with whom it can directly communicate at time

t. U is its private utility function. L

Net

deﬁnes the de-

pendence level between the received data in a given

sensors network (Net). The utility function U of the

agent is the score used in order to help to improve the

learning rate U =

errors rate

good decision

In the next section, we conduce a set of experi-

ments with a set of soil types in order to optimize

crop production and water inﬁltration in the peanut

basin (see, ﬁgure 5). Choosing the right soil and crop

is important but must take into account water quantity

and water quality. That’s why choosing the right wa-

tering system for the soil and crop, affects soil health

and plant growth. Hence, the importance of main-

taining appropriate humidity, saving water by meet-

ing the speciﬁc needs of crops and preserving soil

structure. A good watering system also helps to max-

imise crop yields and maintain a favourable environ-

ment for optimum development. By precisely adjust-

ing irrigation to the speciﬁc soil conditions, it is possi-

ble to avoid wasting water while promoting optimum

crop growth. The system also speciﬁes the amount

of water required in real time for each crop, based on

agro-ecological parameters and the physicochemical

parameters of the soil. The ﬁgure 5 shows the ex-

perimental set-up that we repeated for each type of

soil, depending to the amendments and sensors’ po-

sition. We apply the same compaction effect (weight

of a person), the same quantity of water and the same

composition: 75% soil and 25% soil amendment. We

placed the sensors at a distance of 5 cm, with a width

of 50 cm and a depth of 30 cm, based on the root sys-

tems of the crops that we consider in this work. Each

sensor can measure humidity, NPK (nitrogen, phos-

phorus, potassium), pH (potential of hydrogen) and

EC (Electro Conductivity), which allows us to deter-

mine the inﬁltration rates.

Figure 5: During our experimental set-up, we repeated for

each type of soil and amendments the same sensors’ posi-

tion.

6 OUTCOMES

We have done our study by taking into account these

sets of parameters.

Dates (month/day/year): From 06/01/2024 to

07/31/2024.

Location: Latitude 14.347943, Longitude -

16.410459.

The agroclimatic parameters collected are:

• PS = Surface Pressure (kPa).

• QV2M = Speciﬁc Humidity at 2 Meters (gkg).

• T2M MAX = Temperature at 2 Meters Maximum

(C).

• T2M MIN = Temperature at 2 Meters Minimum

(C).

• WS2M MAX = Wind Speed at 2 Meters Maxi-

mum (ms).

• WS2M MIN = Wind Speed at 2 Meters Minimum

(ms).

• PREC = Precipitation Corrected (mmday).

• UVA = All Sky Surface UVA Irradiance (Wm

• UVB = All Sky Surface UVB Irradiance (Wm

• EVAP = Evaporation and evapotranspiration after

24 hours.

• Moisture = Means’ soil humidity rate between 0

cm and 10 cm after 24 hours.

• INFD = Speed of depth inﬁltration (cm/s).

• INFL = Speed of lateral inﬁltration (cm/s).

Temperatures in Senegal, range from very warm to

hot, with an annual average temperature of 36 Cel-

sius. At least 4 months of the year are tropical and

frequently sultry with temperatures above 36 Cel-

sius. The distribution of crop types grown in Sene-

gal correlates with the timing of seasonal rainfall (ﬁg-

ure 6 and ﬁgure 7). Moreover, some of the practices

ICINCO 2024 - 21st International Conference on Informatics in Control, Automation and Robotics

476

Figure 6: Our study of the mean variations of the parameters

from 2000 to 2023.

Table 1: Soil parameters without soils’ amendment. These

values are our reference before tests.

Soil Type EVAP Moisture INFD INFL

Ferruginous

tropical soil 30 46 20 15

Sandy soil 31 34 40 45

Loamy soil 40 30 13 13

Clay soil 37 46 16 16

Humus-

bearing soil 10 75 65 64

Clay and

loamy soil 23 50 20 17

of the Green Revolution, especially the use of mod-

ern crop varieties and the addition of synthetic fer-

tilizers and pesticides/herbicides are not sustainable

practices, especially under climate change conditions.

NPK (nitrogen, phosphorus and potassium) inputs are

the most important parameters in maximizing yields

and economic returns to farmers. However, in the

peanut basin it is required to take into account to the

soil type, the water quality and water quantity, the pH

and the salinity (EC) for agriculture calendar.

The ﬁgure 6 illustrates that a considerable number

of climatic parameters exhibit minimal variation from

one month to another. However, it is primarily the

availability of water that is of paramount importance,

and this is the factor that gives rise to the seasonal

variation in certain activities between the months of

June and October. From table I to table V, we select a

set of farms’ soil types of the peanut basin and show

the evolution of evaporation, inﬁltration and moisture

in an average temperature of 36 Celsius. We used

DHT11 and soil moisture sensors to measure evap-

oration rates, inﬁltration rates and moisture levels for

the speculations listed in table 6. This was done to

understand and provide decisions based on soil and

amendment types.

After that, we use a set of endogenous knowledge

to improve soil parameters by using natural soils’

amendment like:

• Peanut shells

• Livestock manure

• Poultry manure

• plant mixture

Table 2: Soil parameters after amendment with peanut

shells.

Soil Type EVAP Moisture INFD INFL

Ferruginous

tropical soil 10 70 70 60

Sandy soil 21 60 60 55

Loamy soil 10 80 83 80

Clay soil 7 90 86 86

Humus-

bearing soil 6 95 85 84

Clay and

loamy soil 7 90 80 80

Table 3: Soil parameters after amendment with livestock

manure.

Soil Type EVAP Moisture INFD INFL

Ferruginous

tropical soil 12 70 70 65

Sandy soil 23 50 50 55

Loamy soil 14 70 73 70

Clay soil 15 80 76 76

Humus-

bearing soil 9 90 80 80

Clay and

loamy soil 14 80 70 70

Table 4: Soil parameters after amendment with poultry ma-

nure.

Soil Type EVAP Moisture INFD INFL

Ferruginous

tropical soil 12 70 70 65

Sandy soil 23 50 50 55

Loamy soil 14 70 73 70

Clay soil 15 80 76 76

Humus-

bearing soil 9 90 80 80

Clay and

loamy soil 14 80 70 70

The results show that the amendments improve

water retention, inﬁltration and decrease evaporation.

In addition, depending on the amendments these im-

provements can double the water retention capacity of

soil types in consideration of our fundamental metrics

(cf. table 1). This makes it possible to rationalize wa-

ter and improve the frequency of watering according

Water Optimization in Digital Farming

477

Table 5: Soil parameters after amendment with plant mix-

ture.

Soil Type EVAP Moisture INFD INFL

Ferruginous

tropical soil 12 64 65 65

Sandy soil 25 40 40 50

Loamy soil 10 75 75 75

Clay soil 15 80 76 76

Humus-

bearing soil 10 83 70 70

Clay and

loamy soil 13 76 67 67

Table 6: Our list of practical case’ speculations.

Carrot, Sweet potato, Eggplant, Lettuce,

Cabbage, Okra, Tomato, Turnip, Melon,

Zucchini, Cucumber, Bell pepper, Chilli,

Onion, Cassava, Potato, Hibiscus sabdariffa,

Parsley.

to the types of soil and crops. Based on the data from

our sensors, it is possible to deduce that farmers can

improve the composition of their soils to optimise wa-

ter use according to the soil types and crops on their

plots. The types of soil studied in this study are natu-

rally available in the peanut basin and improve water

retention and inﬁltration. In order to assess agricul-

tural calendar to help farm’s activities over time and

ﬁnd the relationship between crop production and soil

amendment, we use the following speculations list to

test our soils understanding highlighted in this work.

For each speculation, we consider such needs:

• NPK hec(Kg): nitrogen, phosphorus, potassium

for each hectare.

• Water Day(L): stream-day water requirements

(litre).

• Temperature(C): Temperature (Celsius).

• Light(UV): Means of Ultraviolet (UVA and

UVB).

• Soil type: Soil type (tropical ferruginous soils,

Sandy soils, loamy soils, clay soils, Humus-

bearing soils and Clay and loamy soil).

• Ripening time(day): Number of days before har-

vest.

• Salinity tolerance: Electro conductivity (EC).

For the comparison of needs, we used Elbow method

(Umargono et al., 2020) and the KMeans algorithm

(Onoda et al., 2010) which is an artiﬁcial intelligence

algorithm in order to ﬁnd the crops’ similarities. The

table 7 below, shows the similarity between a set of

speculations regarding their needs, the soil’ parame-

Table 7: Cluster assignment by similarity.

Cluster 1

Cluster

3 Cluster 4

Carrot,

Eggplant,

Zucchini,

Cucumber,

Onion,

Potato

Sweet

potato,

Okra Lettuce

Cabbage, Tomato,

Turnip, Melon,

Bell pepper, Chilli,

Cassava, Hibiscus

sabdariffa, Parsley

Figure 7: Comparative variations of the humidity and the

temperature from 2000 to 2023.

Figure 8: Agricultural calendar for a good distribution of

farm activities to avoid the seasonality of agriculture.

ters and the climate’ parameters that are determining

the crop adaptation.

Depending to the analyse of the ﬁgure 6 and the

ﬁgure 7, we propose the following agricultural calen-

der (ﬁgure 8) by taking into account the possibility of

an out-of-season cultivations. This in order to over-

come the abandonment of agricultural perimeters or

an agriculture depending on the raining season.

This ﬁgure 8 in combination with the tables 1, 2,

3, 4 and 5 permit to ﬁnd a set of other speculations

adaptable in the peanut basin when the salinity (Elec-

tro Conductivity) is between 0.1 and 1.2 and Ph (Po-

tential of Hydrogen) between 5 and 8. The compari-

son with the results in (Faye et al., 2024b) shows how

important is to take into account the types of soil and

types of amendments that can improve the cultural

calendar. This work further proves that understand-

ICINCO 2024 - 21st International Conference on Informatics in Control, Automation and Robotics

478

ing climate parameters alone will not help farmers and

feed the future generations.

7 CONCLUSION

In Senegal, agriculture dependent highly to soil types,

climate, soil salinity and water. This work provides a

set of results in order to know, how to optimize water

in different soil type and it proposes an agricultural

calendar for a good distribution of agriculture activi-

ties over time and ﬁnd the relationship between crops.

Our results show the effectiveness of our solution to

avoid the dependences on raining season. This makes

possible to understand crops’ dependencies and an-

ticipate the agroecological phenomena, the planning

of production facilities and variations in agricultural

yields. In the future we aim to disseminate this tech-

nique in the other agroecological area of the Senegal.

Incorporate an analysis of the socio-economic impact

of our agricultural calendar on local communities by

selecting performances metrics and comparison with

traditional methods.

REFERENCES

ANSD (https://www.ansd.sn/regions, 2024). Agence

nationale de la statistique et de la demongra-

phie/accueil/regions.

Cheng, W., andXiaoting Wang, T. M., and Wang,

G. (2022). Anomaly detection for internet of

things time series data using generative adversar-

ial networks with attention mechanism in smart

agriculture. Frontiers in Plant Science, page

https://doi.org/10.3389/fpls.2022.890563.

Faye, P. F., Faye, J. A., and Senghor, M. (2024a). Decen-

tralized intelligence for smart agriculture. In Proceed-

ings of the 16th International Conference on Agents

and Artiﬁcial Intelligence (ICAART 2024), 1:240–

247, ISBN 978–989–758–680–4, ISSN 2184–433X.

DOI: 10.5220/0012342100003636.

Faye, P. F., Faye, J. A., and Senghor, M. (2024b). Machine

learning for a better agriculture calendar. In Proceed-

ings of the 19th International Conference on Software

Technologies(ICSOFT 2024), pages 307–314, ISBN

978–989–758–706–1, ISSN 2184–2833, pages 307–

314. DOI: 10.5220/0012691800003753.

Faye, P. F., Sall, M. M., Dembele, B., and Faye, J. A. A.

(2023). Plateforme num

erique de guides culturaux

dans un contexte de changement climatique et de

baisse des rendements agricoles. RAMReS Sciences

des Structures et de la Mati

ere, pages vol. 7, no. 1, pp.

85–115, ISSN : 2630–1180.

Faye, P. F., Senghor, M., and Ndione, G. A. (2022). Stable

trade coordinations in smart agriculture. International

Joint Conference on Web Intelligence and Intelligent

Agent Technology (WI-IAT), pages 215–222.

Feng, Q., Wang, S., Wang, H., Qin, Z., and Wang,

H. (2022). Circle ﬁtting based image segmen-

tation and multi-scale block local binary pattern

based distinction of ring rot and anthracnose on

apple fruits. Frontiers in Plant Science, 13:doi:

10.3389/fpls.2022.884891.

Fu, L., Li, S., Rao, Y., Liang, J., Teng, J., and He, Q. (2022).

A novel heuristic target-dependent neural architecture

search method with small samples. Frontiers in Plant

Science, page 10.3389/fpls.2022.897883.

Maia, R. F., Lurbe, C. B., and Hornbuckle, J. (2022).

Machine learning approach to estimate soil matric

potential in the plant root zone based on remote

sensing data. Frontiers in Plant Science, page

https://doi.org/10.3389/fpls.2022.931491.

Onoda, T., Sakai, M., and Yamada, S. (2010). Careful seed-

ing based on independent component analysis for k-

means clustering. 2010 IEEE/WIC/ACM International

Conference on Web Intelligence and Intelligent Agent

Technology, 3:112–115.

Paswan, R. P. and Begum, S. A. (2013). Regression and

neural networks models for prediction of crop produc-

tion. International Journal of Scientiﬁc and Engineer-

ing Research, pages Volume 4, Issue 9,ISSN 2229–

5518.

Sellam, V. and Poovammal, E. (2010). Prediction of crop

yield using regression analysis. IEEE Trans. Knowl.

Data Eng., pages vol.23, no. 10, pp. 1498–1512.

Sun, M., Xu1, L., Luo, R., Lu, Y., and Jia, W. (2022).

Fast location and recognition of green apple based

on rgb-d image. Frontiers in Plant Science, 13:doi:

10.3389/fpls.2022.864458.

Umargono, E., Suseno, J. E., and K., V. G. S. (2020).

K-means clustering optimization using the elbow

method and early centroid determination based-on

mean and median. International Conferences on In-

formation System and Technology (CONRIST 2019,

DOI: 10.5220/0009908402340240:pages 234–240.

Uyeh, D. D., Iyiola, O., Mallipeddi, R., Asem-Hiablie,

S., Amaizu, M., Ha, Y., and Park, T. (2022). Grid

search for lowest root mean squared error in pre-

dicting optimal sensor location in protected culti-

vation systems. Frontiers in Plant Science, page

https://doi.org/10.3389/fpls.2022.920284.

Zhang, L., Lei, L., and Yan, D. (2010). Comparison of

two regression models for predicting crop yield. IEEE

International Symposium on Geoscience and Remote

Sensing (IGARSS).

Zingade, P. D., Buchade, O., Mehta, N., Ghodekar, S., and

Mehta, C. (2018). Machine learning-based crop pre-

diction system using multi-linear regression. Interna-

tional Journal of Emerging Technology and Computer

Science(IJETCS), pages Vol 3, Issue 2.

Water Optimization in Digital Farming

479