International Journal of Computational and Experimental

Science and Engineering

(IJCESEN)
Vol. 11-No.1 (2025) pp. 449-463
http://www.ijcesen.com
Copyright © IJCESEN ISSN:2149-9144
Research Article

An Efficient Smart Flood Detection and Alert System based on Automatic Water
Level Recorder Approach using IoT
Mansi Joshi1, S. Murali2*
1
Indiana University Purdue University Indianapolis, IN, USA;
Email: [email protected] - ORCID: 0000-0002-5247-785X
2
School of Computer Science and Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
* Corresponding Author Email: [email protected] ORCID: 0000-0002-5247-78XX

Article Info: Abstract:

DOI: 10.22399/ijcesen.717 An innovative flood detection system may track an increase in water levels. Deployed in
Received : 02 December 2024 cities or other areas of interest, the system consists of sensors. Both mains energy and
Accepted : 19 December 2024 solar power are viable options for the sensors. To detect impending flooding promptly, a
flood warning system uses reliable and up-to-date sensing equipment such as rain gauges,
water level sensors, and flow rate sensors for the smart alert system. The challenging
Keywords : characteristics of smart flood detection and alert systems are that some people may not
be able to access the warnings, and flash floods may happen too quickly for a warning to
Smart Detection,
Smart alert system, be adequate. Hence, in the proposed method, the Automatic Water Level Recorder
Automatic water level recorder, enabled the Internet of Things (AWLR-IoT), which integrates a low-cost cloud to
Internet of Things, overcome the challenges of the smart flood detection and alert system and increase
Wireless Sensor Network. optimization modelling and efficiency. Among the most destructive natural catastrophes
that may happen on Earth is flooding. After that, a Wireless Sensor Network (WSN) is
used to accomplish the flood prediction utilizing data from sensors enabled by the Internet
of Things. Heavy rainfall and the following water outflow cause flooding in nations with
certain climate conditions. The system monitors humidity, temperature, water level,
water rise rate, and rainfall to identify when a flood is imminent. The function of the
AWLR-IoT sensor is for monitoring and recording in a database with real-time sensing.
This research shows that the low-cost AWLR-IoT sensor has reduced processing time
compared to conventional data processing.

1. Introduction Finding a balance between efficiency, scalability,

and the system's utility and reliability is a continuous
Despite offering a novel approach to a critical effort because sacrificing one could destroy the other
environmental issue, a cloud-based smart flood [6]. Consideration of possible socio-economic
monitoring and alert system that is low-cost consequences is additionally crucial; these include,
encounters numerous obstacles. Concerns regarding are not limited to, how vulnerable groups can obtain
the accuracy and reliability of flood detection access to the system and how it could replace
devices are high [1]. The system aims to utilize traditional monitoring systems and ways of life [7].
inexpensive sensors and analyse data in the cloud; Develop a comprehensive strategy to tackle these
however, a major challenge is a need to guarantee issues head-on to manage flood risks effectively [8].
accurate flood detection in many scenarios with Incorporating reliable cloud infrastructure,
complicated topography [2]. Other worries from the community engagement, strong sensor technologies,
system's dependence on cloud infrastructure include and regulatory considerations into this plan's
data latency and connectivity challenges [3]. This execution is essential [9].
becomes even more problematic in areas that are An affordable smart flood detection and alarm
geographically remote or vulnerable to natural system that operates in the cloud uses various
disasters, where internet connectivity is limited [4]. existing techniques to detect and notify of impending
Moreover, scalability could become a problem as the floods [10]. Installing inexpensive sensors in flood-
system grows to include more people or regions [5]. prone places allows for continuous monitoring of
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

water levels, which is one popular strategy [11]. better warning of flash floods and other flood
Ultrasonic, pressure, and conductivity sensors are a situations that are evolving rapidly.
few technologies that these sensors can use to detect The remainder of the research follows Section II's
changes in water levels reliably [12]. A cloud literature review. A Smart Flood Detection and Alert
platform accepts data collected from various sensors System Hosted in the Cloud at an Affordable Price.
and processes and analyses it in real-time [13]. Using The outcomes of the AWLR-IoT are covered in
machine learning algorithms to scour sensor data for Section III. Section IV contains the results and
trends that might suggest a flood is about to happen discussion, while Section V contains the summary
[14] is common practice. To get a better look at the and recommendations.
flood conditions happening across a larger area, the
system may additionally use remote sensing 2. Literature Survey
techniques like radar data or satellite images [15].
Even if these strategies might work, there are still a This review focuses on several approaches put forth
lot of obstacles to get beyond. An enormous by various scholars, all of which bring something
challenge is the reliability and accuracy of the sensor special to the surface regarding ways to improve
readings, especially in outlying or harsh areas flood control and response capacities.
without adequate infrastructure [16]. In more remote
areas, it may be very labour- and resource-intensive Smart IoT Flood Monitoring System (S-IoT-FM)
to calibrate and maintain the sensors such that they Zahir developed the S-IoT-FM, S. B. et al. [19] and
constantly work. Integrating several data sources and uses a web server to provide real-time water level
ensuring compatibility and interoperability across monitoring. It notifies people promptly, allowing
various sensor kinds and platforms are additional them to take proactive actions for evacuation. Its
technological hurdles [17]. Additional worries about low-cost construction and ease of maintenance
data security, privacy, and connectivity come from improve flood prediction and response, reducing
depending on cloud infrastructure, especially in property and life loss in rural and urban locations.
areas with bad internet service [18]. Another major
obstacle must be solved to increase the system's Internet of Things (IoT)
coverage area without decreasing its operational The suggested approach by Chaduvula, K. et al. [20]
efficiency or cost-effectiveness. Overcoming these uses IoT technology in conjunction with GSM
obstacles and creating a trustworthy flood detection modules and sensors to identify increasing river
and alert system requires a comprehensive plan water levels (IoT-GSM). Using a GSM modem, an
considering technical, logistical, and socio- 8051 microcontroller processes data and sends out
economic aspects. SMS warnings. Float switches, light-emitting
Problem statement diodes, LCDs, and other components make flood
An early warning system and accurate, quick flood monitoring possible. The paper aims to minimize
detection are heavily emphasized in the problem loss of life and property by providing timely
statement for the Low-Cost Cloud-based Smart warnings.
Flood Detection and Alert System. Challenges that
need fixing include the lightning-fast arrival of flash Decision Tree Algorithm (DTA)
floods, the difficulty of getting alerts to everyone, To warn of potential danger, Vinothini, K. et al. [21]
and the need to enhance the system's performance in use the Internet of Things to track flood levels.
light of changing weather patterns. The suggested Sensors track variables, including humidity, water
solution combines AWLR-IoT technology which level, and temperature. Connected devices receive
specifies water level recorders enabled by the data processing from the PIC microcontroller using
Internet of Things with a very inexpensive cloud Wi-Fi. The system uses the cloud and the DTA to
architecture to overcome these problems and categorize data. The results confirm that the flood
improve flood detection and warning capabilities. monitoring and warning capabilities are successful,
with an accuracy rate of 99.6 %.
Objectives
 Developing a reliable system for detecting floods Real-Time Flood Monitoring (R-TFM)
can be accomplished using low-cost sensors and Using NodeMCU-based wireless sensor nodes
cloud computing. connected with the Blynk application, Sabre et al.
 Facilitate the dissemination of flood warnings to [22] provide a R-TFMand early warning system.
all individuals, particularly those who reside in Ultrasonic and rain sensors detect water levels and
underserved or remote areas. the intensity of rainfall, which then trigger alerts
 By improving the alarm system's efficiency and through buzzer and LED. Victims receive email data
responsiveness, it will be possible to provide for forecasts and real-time updates using the Blynk
450
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

app. The efficacy of early flood detection and unexpected inundations. The system is designed to
warning has been validated through testing. quickly identify and notify authorities and residents
about potential flood hazards by using sensors,
ES-LSTM energy from renewable sources of information, and
Hourly precipitation forecasting using an ES-LSTM Internet of Things, or IoT, technologies. It solves the
and RNN-based model with ANN and DT for problems of accessibility and quick reaction by
classification is proposed by Hayder, I. M. et al. combining the AWLR-IoT along with a cheap cloud
[23].With meteorological data from Australia, ES- infrastructure. It highlights the methodology's
LSTM managed a MAPE of 3.17, an RNN of 6.42, potential to improve flood forecast accuracy and
an ANN prediction accuracy of 96.65%, and DT of optimize response efforts in this introduction, which
84.0%. The accuracy of flood forecasts was covers its major components. Table 1 is analysis of
improved using ES-LSTM and ANN, which related works with advantages and limitations.
outperformed other models. Figure 1 shows smart flood detection. A hierarchical
architecture is included into the Smart Flooding
Internet of Things -based solution (IoT-S) Detection system to carefully watch, analyse, and
To better anticipate and control catastrophic events distribute important data linked to floods. In
like floods in India's Bihar state, Khan et al. [24] locations prone to flooding, sensors are placed at
present an IoT-S. A disaster-alert smartphone app strategic intervals to detect changes in the amount of
notifies the proper authorities in the event of an water quickly, rainfall, and other relevant
imminent disaster. The model improves the characteristics; these sensors act as the primary data
precision and velocity of responses, which could collectors. These sensors provide the system with
reduce casualties and property damage caused by crucial insights. The microcontroller/embedded
floods. system co-operates with the flood sensors to
coordinate and oversee their functions. This part
Out of all these options, the AWLR-IoT is the most prepares the data collected by the sensors to be sent
effective one when it comes to flood management. to the Gateway Device, which is the system's nerve
centre. This critical component collects information
3. Proposed Method from various flood sensors and sends it to the cloud
system for processing. The system's scalable
A novel technique is presented in this paper: an computation and storage capabilities provided by the
inexpensive online smart flood monitoring and alert cloud infrastructure are vital for managing massive
system. This is in response to the increasing threat of amounts of sensor data. At this stage, data is
floods, particularly in places that are susceptible to processed and analysed using

Figure 1. Smart Flood Detection.

451
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

Table 1. Analysis of related works with advantages and limitations

Method Advantages Limitations
S-IoT- Real-time water level monitoring.Timely Reliance on web server connectivity May require
FM notifications for proactive actions internet access for notifications
Utilizes IoT technology for flood monitoring Limited coverage area based on GSM network
IoT-
Sends SMS warnings through GSM modem availability May require additional cost for SMS
GSM
alerts
Uses IoT for flood monitoring Complexity in implementation due to cloud
DTA Cloud-based data processing for categorization integration
Potential latency in data processing
Real-time monitoring and early warning system Reliability dependent on stable internet connection
R-TFM
Utilizes wireless sensor nodes and Blynk app Limited scalability in case of extensive deployment
Hourly precipitation forecasting Requires historical meteorological data for training
ES-
Improved accuracy in flood forecasts Computational complexity may hinder real-time
LSTM
operation
Disaster-alert smartphone app for authorities Dependent on smartphone adoption
IoT-S Precision and velocity in responses Requires active participation from authorities for
effective response

algorithms and artificial intelligence to find trends, The system is not complete without the Feedback
patterns, and anomalies that might suggest a flood is Loop, which allows for ongoing optimization and
about to happen in the raw data collected by the modification. The system strengthens resilience in
sensors. Data analysis leads to Alert Generation, flood-prone places by continuously improving
where stakeholders are swiftly notified of possible algorithms, increasing forecasting accuracy, and
flood threats through the system's warnings and adapting to changing flood dynamics. It analyses
alerts. These warnings are delivered across several user input, performance indicators, and
means to guarantee prompt communication and environmental data. A key component in bolstering
response, including public warning systems, mobile resilience and protecting communities from floods,
apps, email, and short message service (SMS). The the Smart Flood Monitoring system integrates these
User Interface is a platform that presents data elements to provide a complete solution for early
visualizations in real-time, flood notifications, and flooding detection, quick reaction, and successful
other relevant information in an easy-to-understand risk reduction.
way. It helps users engage and make decisions.

Tj + YM = argTj max − ∑ q (ilk |Tj ) log q (ilk |Tj ) + IM ([Y0 , Y1 , … , YM−1 ]) (1)
k

In the equation 1, where finding the ideal choice the adverse sum of the conditional odds q (ilk |Tj ) and
variable Tj and supplementary variable YM is the goal the information that share IM . The conditional
of the maximization problem that the equation likelihood of seeing a certain event ilk in relation to
depicts. With a collection of M variables the choice variable Tj is probably represented by the
Y0 , Y1 , … , YM−1 and their respective logarithms
term q (ilk |Tj ).
weighted, the objective function aims to minimize
L
pl 1
minH m(H) = ∑ . Ml (H), where Ml (H) = ∑ mj H (2)
p pl
l=1 j∋Ql

The average loss m(H) can be minimized by within class Q l is denoted by Ml (H), which is
identifying the optimal assumption H using the calculated as the product of the individual losses
minimization problem represented by the equation 2. inside Q l and the inverse of the initial likelihood pl .
The average loss is calculated by adding together all To get the greatest possible classification
the class-specific losses, which are calculated as the performance, the equation seeks to minimize the
product of the prior probability pl and the average loss and identify the hypothesis H that fits
conditional likelihood Ml (H), then multiplying by the data the most. This essentially strikes a balance
the class-specific loss. In this case, the average loss
452
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

Figure 2. Flood Alert System using WSN

between prior experience and conditioned a disaster and alarm system that makes use of Zigbee
probabilities.The Flooding Alert System integrates modules inside a network of sensors (WSN). This
environmental data with current technologies to part allows for the dissemination of severe weather
identify floods early and disseminate warnings. warnings and detailed weather reports to those who
Predicting when and how much rain will fall is need them. The system can detect crucial thresholds
essential in the lead-up to floods, and this method is and send out notifications based on them by using
based on data collected by the hydro meteorological decision tree analysis.
department. Smart placement of node sensors in Multiple tiers make up the Flood Alert System's
flood-prone regions allows for the continuous, real- implementation, and all work together to provide
time monitoring of critical variables including water flood warnings. The hydro logical department
level, rainfall quantity, and relative humidity. provides the initial real-time rainfall data that is used
In a multi-hop sensor network, the sensor nodes act as a basis for flood prediction. Then, the WSN
as the primary data collectors and relay the collected gathers information on the river's flow and present
data to a command-and-control hub. In this case, the water level, giving a picture of the flood's dynamics
present flood state is determined by analysing the as it happens in real time (figure 2). In the third
data that comes in streams using sophisticated implementation stage, threshold values are
processing techniques. The detected water level is determined by processing the acquired data, which
used to classify this state, providing a detailed indicate the likelihood of upcoming floods. Lastly,
insight of the flood severity from low to extreme. the system rapidly notifies riverfront communities
The solution makes use of SMS (short message and ground stations of the imminent threat by
service) notifications to guarantee direct and quick activating SMS warnings when certain criteria are
connection with impacted persons. Residents in surpassed. Communities can better react to flood
flood-prone zones may receive immediate flood dangers and increase overall resilience due to the
status updates through the system, which allows for Flood Alert System's multi-level strategy guarantees
rapid decision-making and proactive risk mitigation. the prompt and targeted distribution of vital
Along with SMS warnings, the design incorporates information.

Eqs (Fql (n1 ). Fql (n2 )) + (X)aggregate_G(1) = n1 + or × n2 + Efdq2 (Xglobal ) (3)

The expected value of the sum of two functions, the function X global under a particular condition q2,
expressed as Eqs (Fql (n1 ). Fql (n2 )), is involved in denoted as Efdq2 (Xglobal ), and the product of n1 and
the connection represented by the equation 3. An n2 , representing the sum of these two variables, are
aggregate operation applied to the variable X, identical to this.
denoted as aggregate_G(1). The anticipated value of

453
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

1
Bwfs (Fpdq1 (Xloca.l )) = × (Fpdq1 (Xloca.l1 )) × (Fpdq1 (Xloca.l2 ) … Fpdq1 (X loca.ln )) (4)
L
The function Fpdq1 (Xloca.l ), represented further, the AWLR-IoT system is an essential go-
as Bwfs (Fpdq1 (X loca.l )), is subject to a weighted between.
After receiving the data, the AWLR-IoT system
average operation in equation 4. Where the weight connects to a WSN to forecast when a flood would
of each term is inversely proportionate to the total occur. The WSN improves readiness by predicting
number of terms, indicated as L. With each instance possible flood occurrences and evaluating hazards
represented by X loca.l , X loca.l1,X loca.l2..., X loca.ln , utilizing real-time data and predictive algorithms.
representing a unique location l1, l2,..., ln, this After processing, the data is sent to a database in the
process determines the mean value of the function. cloud where it may be stored and analysed further.
In Figure 3, the Water Detection System based on a Quick, real-time processing and data evaluation are
combined structure which is AWLR-IoT enabled. made possible by the cloud architecture's scalable
The heart of the system is a collection of sensors computing and storage capabilities. Here, a high-
placed in flood-prone locations to measure things tech warning system is in place to assess incoming
like water level, flow rate, and rain gauges. The main data streams, identify potential flood hazards, and
data gathering devices are these sensors, which promptly issue alerts and warnings. The flood alert
continually monitor important characteristics that system has an alarm mechanism that goes off when
might indicate a flood. The sensors provide their data it detects a flood. Users are then notified through
to the AWLR-IoT framework, which processes and several channels, such mobile applications, email, or
communicates all the acquired data. This part text message. This way, communities and
combines the ability to automatically record water individuals are alerted quickly enough to take the
levels with internet of things technology, making required measures to protect lives and property.
data processing and transfer easier. When it comes Improved flood detection as well as response
to collecting data from sensors and processing it capabilities will lead to safer and more resilient
flood-prone areas, which is the goal of the

Figure 3. Automatic Water Level Recorder enabled the Internet of Things

454
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

Flood Detection System. The solution allows the infrastructure of the cloud. This reduces the
preventive steps to be taken by seamlessly effect of floods and improves overall disaster
combining sensor technologies, IoT capabilities, and preparation.

X loca.l + X bhh_g(j) = Bwfs (Fpdq1 (X loca.l )) + Efdq2 (Xglobal ) (5)

In the equation 5, two variables are shown by the represents the weighted average operation
equation: one, X loca.l , which represents a local performed on a function Fpdq1 (X loca.l ) over many
variable at position l, and another, X bhh_g(j) , which times at location l, and the other that represents the
represents a variable linked to a particular function anticipated outcome of a global variable X global
bhh_g(j). The equation includes two terms: one that under a circumstance q2.

(X)bhhg(j) + X localJ = Efdq2 (Xglobal ) + X localJ − ∃ ∆2L (XlocalL , c) (6)

The context-dependent connection between the certain boundaries. In the case of floods, this allows
variables is shown by the equation 6 for the detection for the prompt implementation of safety measures.
accuracy analysis. The combined value of a The Cloud-based Data Lake is the backbone of the
function-associated variable bhhg(j) and a local Flood Warning System; it allows for the creation of
variable X localJ is shown on the sum of two a web dashboard that displays real-time flood
terms, (X)bhhg(j) and X localJ . By moving equation, predictions. This dashboard is designed to work in
tandem with cloud notification systems, so warnings
two new terms are added: the anticipated outcome of may be sent to relevant authorities through several
a global variable X global under a particular channels, such as smartphone notifications, email,
circumstance q2, and another instance of the local and text message. Stakeholders may get alerts and
variable denoted by X localJ . A term ∆2L (XlocalL , c) is messages across many platforms, and the Cloud-
included in the equation, indicating that there is a based Flood Prediction Dashboard can be linked to
change or variation in the local variable X localL , c other websites using REST API calls. This allows for
with relation to another variable c. For the purpose the interchange of information. The system is more
to provide communities and authorities with up-to- effective because all essential stakeholders are
the-minute predictions of impending floods, the constantly informed about creating flood scenarios
Flooding Warning System is vital. The project through this interoperability. Figure 4 (Web-based
information is kept in a data centre in the cloud information-lake-based flood warning system)
centre, which always makes it accessible. An summarises a full-scale Flood Warning System that
essential part of the system is the real-time automatic uses a Data Lake setup. Authorities and communities
alerts that warn stakeholders by email or text may stay informed and take proactive actions to limit
message as soon as certain parameters go beyond the effect of flooding disasters through real-time
monitoring, automatic notifications, and interaction
with cloud-based services.

Figure 4. Web-based data-lake-based flood warning system

455
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

√6 U
Q U + L = Q buf + LT + [0.5572 + Mp [Mp [ ]]] (7)
∇ U−1

A link between quantities connected to a system's U

dynamics is described by the equation 7 for the Mp [Mp [ ]]]. The variable Q buf is probably a
U−1
response time analysis. The amount of U − 1 buffer quantity, LT is a term involving L and T, and
representing specific variables or parameters is the expression inside the square root implies a
shown by Q U + L. Where it includes many terms,
complicated calculation with √6, U, and Mp, was the
√6 functions of operations.
this is equal to Q buf + LT + ∇
[0.5572 +

p p
1 1 2 Qu
Q buf + T × JU = ∑ QJ × [ ∑(Q j − Q buf ) ] × (8)
p p Ue
j=1 j=1

Several variables and factors are involved in the serialization techniques like pickle to store the
cost-effective analysis depicted by the equation 8. trained model while keeping its precision and
The equationQ buf + T × JU implies that there is a function intact is crucial to this procedure. The user
buffer amount Q buf plus the product of two interface, which is a user-friendly interface for
parameters or variables, which are probably interaction, is a Flask-based responsive online
represented by U and JU . On the right side, which dashboard. Clients may use this dashboard to test or
1 p validate the model by sending new information to
features several terms, this is equal to p ∑j=1 QJtimes
2
the Flask the application programming interface
1 p
[p ∑j=1(Q j − Q buf ) ]. The quantity QJ is probably a (Next, the Flask API gets the model that was stored,
combination of Q and J, the square root of the puts it into memories, and uses it to evaluate the
disparity between QJ and the buffer amount is client's input. It then uses the model's categorization
2 capabilities to provide predictions. Once the
(Q j − Q buf ) , and the ratio of Qu to Ue is categorization is complete, the Flask API returns the
Qu
probablyUe. Showcasing a complex strategy to findings to the user, who may then obtain up-to-the-
guarantee smooth communication between end- minute, accurate forecasts.
users and an established machine learning model, Enclosing the whole system inside a Docker image
Figure 5 demonstrates the full process of handling guarantees that the model can be used across many
warnings through model deployment. The use of platforms and environments. This includes the

Figure 5. Management of alerts enabled by model deployment.

456
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

trained model, the Flask API, and any related simplified, allowing it to be used in more contexts
software packages or runtime libraries. The Docker and by more people. Figure 5 shows a sturdy method
Hub registry then serves as a central repository for for handling notifications brought about by model
this image, making it easy to access and distribute. deployment, which involves enclosing the model
Due to the Docker image's exceptional flexibility, itself and its deployment technology in a Docker
the model and its related components may be easily image. Users will have no trouble interacting with
deployed and executed on any Docker support, the deployed artificial intelligence model owing to
irrespective of the underlying OS or hardware setup. this simplified approach, which improves
Because of its mobility, compatibility problems are productivity, scalability, and accessibility.
eliminated, and the deployment procedure is

4 × J × M ds + 0.0007j
uq + ud = DB 0.22 sf −0.35 × ( )( ) (9)
j2/3 t1/3

A system's reliability parameters and their terms, is equal to DB 0.22 sf −0.35times ×

relationships are shown by the equation 9. On the left 4×J ×M ds+0.0007j
( j2/3 ) ( t1/3 ). In this context, it is probable
side, the total of uq and ud , which probably indicate
velocities or flow rates, is denoted by uq + ud . The that DB 0.22 and sf −0.35 stand for particular
parameters, whereas J, M, j, ds, and t denote
right side of the equation, which includes many
additional variables or coefficients.

Tangle × Tgrade = tan−1 (z⁄y) + (100 %) z⁄y (10)

Two variables, Tangle and Tgrade , are related in the Dataset description: The offered dataset is a great
equation 10 to grade the user satisfaction and tool for detecting flood events since it combines
usability analysis by using the of trigonometric optical and Synthetic Aperture Radar (SAR) image
functions and measures of ratios. The arctangent time series, which are not always easy to find. Since
function, this is equal to tan−1 (z⁄y) + clouds may make it difficult to see floods, SAR data
(100 %) z⁄(100%). In this case, z and y probably is a vital option for reliable detection [25].
Unfortunately, there is often a lack of labelled
stand for user satisfaction. datasets necessary to do machine learning on SAR
data. This dataset provides a wealth of labelled
Finally, a potential option to lessen the destructive optical and SAR image time series, thereby bridging
effects of floods is the Low-Cost Cloud-based the gap. To better identify flood events, researchers
Intelligent Flooding Detection and Alert System. may use this dataset to train machine learning
The system uses cutting-edge sensor technology and models; this will increase disaster response and
Internet of Things connection to deliver precise mitigation capabilities.
alerts in a timely manner, allowing for pre-emptive In the figure 6, an affordable cloud-based smart
steps to be taken to protect lives and property. A flood detection and alarm system is tested for
major step forward in flood detection and reaction detection accuracy within the framework of the
capabilities has been achieved with the incorporation detection accuracy analysis. To study the
of a cost-effective cloud infrastructure, which fluctuations in water level, this investigation will
increases accessibility and processing efficiency. evaluate the accuracy and reliability of the sensors
The findings show that as compared to more placed in flood-prone locations. Factors such as
conventional approaches, the AWLR-IoT sensor's sensor calibration, ambient variables, and the
continuous observation and data recording greatly possibility of false positives and negatives are
decrease processing time. This fresh strategy can considered during the examination. Another facet
significantly increase resistance against one of the that is examined in this research is how well the
most devastating natural catastrophes the planet has system reacts to flood alerts. When floods strike
ever seen. quickly, like in flash floods, this becomes even more
crucial. People test the cloud architecture to see how
4. Results and Discussion well it processes and analyses sensor data in real
time. This aims to test the system's flood detection
The research aims to improve disaster preparedness and notification capabilities. The accuracy study
and response skills by assessing the system examined several obstacles, including data latency,
thoroughly to lower flood risks. network problems, and the incorporation of
numerous data sources produces 97.9%. The
457
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

Figure 6. Detection Accuracy Analysis

Figure 7. Response Time Analysis

Detection Accuracy Analysis seeks to enhance a identifying potential improvement areas and
cost-effective cloud-based smart flood detection and optimizing the system's algorithms and functions.
alert system's capacity to reduce flood risks and In the above figure 7, the primary goal of the
enhance disaster preparedness and response by Response Time Analysis for an Affordable Cloud-
Based Smart Flood Detection and Alert System is to

458
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

assess how quickly the system notifies users when Detection and Alert System. Investment and
floods are about to occur. This research aims to find operational costs related to the flood detection and
out how long it takes to go from seeing a flood alert system's installation and upkeep are both
indicator such as a rise in water level or the severity factored into this study. There are many moving
of rainfall to getting an alarm and sending it to the parts for the total cost, including data transmission
right people. To determine the system's reaction time and processing, cloud infrastructure setup and
and its ability to alert populations to danger, the maintenance, sensor acquisition and installation,
investigation will look at response times in different software integration, and employee training. By
situations and under different environmental comparing these costs to the system's efficacy in
circumstances. To determine bottlenecks and lowering flood damages and losses associated with
improve system performance, people conduct an floods, the analysis aims to ascertain if the system is
exhaustive examination into the factors that impact cost-effective. Several factors are considered during
response times, including sensor placement, data the evaluation process, such as the system's capacity
transfer latency, and the speed of cloud processing. to accurately detect flood scenarios, the timeliness of
The paper considers possible remedies that can its warnings, and the availability of preventive
mitigate the effect of interruptions to system or measures to mitigate floods. To optimize efficiency
network connectivity on reaction times produces without sacrificing performance, the study may
95.9%. The reaction Time Analysis seeks to improve investigate other deployment tactics or technological
a cheap, cloud-based smart flood detection and alert solutions. The Cost-effectiveness Analysis aids
system's efficiency and reliability by analysing decision-makers in allocating resources and
reaction times. This aims to lessen the likelihood of prioritizing investments in flood management
floods and the damage that floods can inflict to infrastructure by measuring the cost-benefit ratio
vulnerable communities. and identifying potential for cost reductions or
In the above figure 8, the capability of the system to efficiency improvements produces 98.9%. While
save costs while increasing the probability of there may be societal benefits to a low-cost cloud-
flooding is examined in the Cost-effectiveness based smart flood detection and alert system, the
Analysis of a Low-Cost Cloud-based Smart Flood

Figure 8. Cost-effectiveness Analysis

459
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

Figure 9. Reliability Analysis

Figure 10. User Satisfaction and Usability Analysis

main objective is to minimize costs without and notify people when floods occur. A system's
sacrificing any of these benefits. dependability is assessed by considering many
In the above figure 9, an affordable smart flood elements, such as the accuracy of data transmission,
detection and alarm system that operates in the cloud the availability of cloud infrastructure, the strategies
is examined in the Reliability Analysis for That for alert dissemination, and the performance of
System to determine its consistency and sensors. The analysis's goal is to find the system's
dependability. The purpose of the system is to detect reliability measures, like its mean time between
460
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

failures and mean time to repair, by looking at past Emphasizing the need of prioritizing user
data and doing simulations. The paper considers how satisfaction and usability for effective adoption,
well the system responds in difficult conditions, like these data demonstrate the system's efficacy in
when the network is down or bad weather is on the decreasing flood risks and improving community
horizon. People are investigating potential sources of resilience.
false positives and false negatives as part of our
endeavours to increase the system's overall 5. Conclusion
reliability and lower the frequency of incorrect
warnings. The paper can include stakeholder and Smart Flood Detection and Alert System in the
end-user feedback to help uncover reliability Cloud at a Low Cost Might Reduce Devastation
concerns or user experience challenges. Doing so from Floods. Rapid flood event identification and
would be done with an eye towards practical alarms are made possible by innovative technology
implementation needs. One way to make a Low-Cost such as Wireless Sensor Networks (WSN) and the
Cloud-based Smart Flood Detection and Alert AWLR-IoT. Widespread deployment is now
System more reliable and trustworthy is to use the feasible, even in resource-constrained rural areas,
insights from the Reliability Analysis. because to the system's scalability and accessibility
This analysis finds areas of weakness and potential enabled by inexpensive sensors and cloud-based
improvement. Because the analysis finds places infrastructure. Predicting when floods may occur,
where things may be better, this is within the realm the system keeps a constant eye on weather
of possibility produces 95.8%. Ensuring a high level conditions like humidity, water level, water rise rate,
of reliability is crucial for improving the system's and rainfall. Being proactive can help us cut
effectiveness in reducing flooding risk and avoiding emergency response times and improve flood
potential effects on affected areas. This change will forecasts. Installing flood detection systems is often
boost the stakeholders' faith in the system's hindered by a lack of funding, especially in less
capabilities. developed nations. Using inexpensive sensors that
In the figure 10, the present research aims to assess are fuelled by solar, or mains electricity can make
the effectiveness and ease of use of an inexpensive, the system more affordable and sustainable.
cloud-based smart flood detection and alert system However, not everyone has equitable access to flood
from the perspective of the end-users. Community warnings, and flash floods can still hit rapidly. To
people, municipal officials, and first responders are enhance the system's accuracy, precision, and ability
all stakeholders whose opinions must be gathered to adjust to evolving flood conditions, further
with the purpose of complete this study. Doing so research and development are necessary.
allows one to gauge the degree of contentment with Communities may become more resilient and better
the system's efficiency and utility. Important factors prepared for floods with the help of a cloud-hosted
that are included in the analysis include how quickly smart flood detection and alert system which does
and clearly flood notifications are sent, how easy it not break the financial institution. Since floods are
is to understand and act upon the information still an existential threat to people's lives and
provided, and how the system is used overall. The livelihoods all across the globe, humanity are in dire
study additionally looks at how easy the system is to need of modern, affordable flood detection
use for different types of users, such as those with technology. Wireless Sensor Network is studied and
physical impairments or low-tech knowledge, so reported in the literature [26-30].
flood alerts can be available to everyone. Listening
to customer feedback can improve the quality of our Author Statements:
user interface, alerting systems, and user support
services. Including end-user feedback into the  Ethical approval: The conducted research is not
review process is an attempt to detect usability issues related to either human or animal use.
and potential solutions. Through this action, people
 Conflict of interest: The authors declare that
can enhance the system's capacity to lessen flood
they have no known competing financial interests
risks and strengthen communities' resilience
or personal relationships that could have
produces 96.5%. To ensure the effective adoption
appeared to influence the work reported in this
and effectiveness of a low-cost cloud-based smart
paper
flood detection and alert system in protecting lives
 Acknowledgement: The authors declare that
and property from floods, it is vital to prioritize user
they have nobody or no-company to
enjoyment and usability. Viewed by stakeholders,
acknowledge.
this will foster a climate of more trust and
acceptance.  Author contributions: The authors declare that
they have equal right on this paper.
461
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

 Funding information: The authors declare that Smart flood resilience: harnessing community-scale
there is no funding to be acknowledged. big data for predictive flood risk monitoring, rapid
 Data availability statement: The data that impact assessment, and situational
awareness. Environmental Research: Infrastructure
support the findings of this study are available on and Sustainability, 2(2), 025006.
request from the corresponding author. The data [12] Abdullahi, S. I., Habaebi, M. H., & Abd Malik, N.
are not publicly available due to privacy or (2019). Intelligent flood disaster warning on the fly:
ethical restrictions. developing IoT-based management platform and using
2-class neural network to predict flood status. Bulletin
References of Electrical Engineering and Informatics, 8(2), 706-
717.
[1] Hashi, A. O., Abdirahman, A. A., Elmi, M. A., Hashi, [13] Basnyat, B., Roy, N., & Gangopadhyay, A. (2021,
S. Z. M., & Rodriguez, O. E. R. (2021). A real-time March). Flood detection using semantic segmentation
flood detection system based on machine learning and multimodal data fusion. In 2021 IEEE
algorithms with emphasis on deep international conference on pervasive computing and
learning. International Journal of Engineering Trends communications workshops and other affiliated events
and Technology, 69(5), 249-256. (PerCom Workshops) (pp. 135-140). IEEE.
[2] Hassan, W. H. W., Jidin, A. Z., Aziz, S. A. C., & [14] Munawar, H. S., Ullah, F., Qayyum, S., & Heravi, A.
Rahim, N. (2019). Flood disaster indicator of water (2021). Application of deep learning on uav-based
level monitoring system. International Journal of aerial images for flood detection. Smart Cities, 4(3),
Electrical and Computer Engineering, 9(3), 1694. 1220-1242.
[3] Suwarno, I., Ma’arif, A., Raharja, N. M., Nurjanah, A., [15] Boulouard, Z., Ouaissa, M., Ouaissa, M., Siddiqui, F.,
Ikhsan, J., &Mutiarin, D. (2021). IoT-based lava flood Almutiq, M., &Krichen, M. (2022). An integrated
early warning system with rainfall intensity artificial intelligence of things environment for river
monitoring and disaster communication flood prevention. Sensors, 22(23), 9485.
technology. Emerging Science Journal, 4(0), 154- [16] Srinithi, A., Nachiappan, D. M., & Vegesna, V. V.
166. (2019). An Embedded Based Integrated Flood
[4] Costa, D. G., Vasques, F., Portugal, P., & Aguiar, A. Forecasting through HAM
(2019). A distributed multi-tier emergency alerting Communication. Muthukumaran and Vegesna, Vinod
system exploiting sensors-based event detection to Varma, An Embedded Based Integrated Flood
support smart city applications. Sensors, 20(1), 170. Forecasting through HAM Communication (January
[5] Rani, D. S., Jayalakshmi, G. N., &Baligar, V. P. (2020, 21, 2019).
March). Low cost IoT based flood monitoring system [17] Perera, D., Seidou, O., Agnihotri, J., Mehmood, H., &
using machine learning and neural networks: flood Rasmy, M. (2020). Challenges and technical advances
alerting and rainfall prediction. In 2020 2nd in flood early warning systems (FEWSs). Flood
international conference on innovative mechanisms impact mitigation and resilience enhancement.
for industry applications (ICIMIA) (pp. 261-267). [18] Munawar, H. S., Ullah, F., Qayyum, S., Khan, S. I.,
IEEE. &Mojtahedi, M. (2021). UAVs in disaster
[6] Al Qundus, J., Dabbour, K., Gupta, S., Meissonier, R., management: Application of integrated aerial imagery
& Paschke, A. (2022). Wireless sensor network for AI- and convolutional neural network for flood
based flood disaster detection. Annals of Operations detection. Sustainability, 13(14), 7547.
Research, 1-23. [19] Zahir, S. B., Ehkan, P., Sabapathy, T., Jusoh, M.,
[7] Babu, V., & Rajan, V. (2019, July). Flood and Osman, M. N., Yasin, M. N., ... & Jamaludin, R.
earthquake detection and rescue using IoT technology. (2019, December). Smart IoT flood monitoring
In 2019 International Conference on Communication system. In journal of physics: conference series (Vol.
and Electronics Systems (ICCES) (pp. 1256-1260). 1339, No. 1, p. 012043). IOP Publishing.
IEEE. [20] Chaduvula, K., Markapudi, B. R., & Jyothi, C. R.
[8] Anbarasan, M., Muthu, B., Sivaparthipan, C. B., (2023). Design and Implementation of IoT based flood
Sundarasekar, R., Kadry, S., Krishnamoorthy, S., & alert monitoring system using microcontroller
Dasel, A. A. (2020). Detection of flood disaster system 8051. Materials Today: Proceedings, 80, 2840-2844.
based on IoT, big data and convolutional deep neural [21] Vinothini, K., & Jayanthy, S. (2019, May). IoT based
network. Computer Communications, 150, 150-157. flood detection and notification system using decision
[9] Van Ackere, S., Verbeurgt, J., De Sloover, L., tree algorithm. In 2019 international conference on
Gautama, S., De Wulf, A., & De Maeyer, P. (2019). A intelligent computing and control systems (ICCS) (pp.
review of the internet of floods: Near real-time 1481-1486). IEEE.
detection of a flood event and its [22] Sabre, M. S. M., Abdullah, S. S., & Faruq, A. (2019).
impact. Water, 11(11), 2275. Flood warning and monitoring system utilizing
[10] Aarthi, M., &Bhuvaneshwaran, A. (2021). IoT based internet of things technology. Kinetik: Game
drainage and waste management monitoring and alert Technology, Information System, Computer Network,
system for smart city. Annals of the Romanian Society Computing, Electronics, and Control, 287-296.
for Cell Biology, 6641-6651. [23] Hayder, I. M., Al-Amiedy, T. A., Ghaban, W., Saeed,
[11] Yuan, F., Fan, C., Farahmand, H., Coleman, N., F., Nasser, M., Al-Ali, G. A., & Younis, H. A. (2023).
Esmalian, A., Lee, C. C., ... & Mostafavi, A. (2022). An intelligent early flood forecasting and prediction
462
 Mansi Joshi, S. Murali / IJCESEN 11-1(2025)449-463

leveraging machine and deep learning algorithms with

advanced alert system. Processes, 11(2), 481.
[24] Khan, R., Shabaz, M., Hussain, S., Ahmad, F., &
Mishra, P. (2022). Early flood detection and rescue
using bioinformatic devices, internet of things (IOT)
and Android application. World Journal of
Engineering, 19(2), 204-215.
[25] https://www.kaggle.com/datasets/virajkadam/sen12
flood
[26] M. Devika, & S. Maflin Shaby. (2024). Optimizing
Wireless Sensor Networks: A Deep Reinforcement
Learning-Assisted Butterfly Optimization
Algorithm in MOD-LEACH Routing for Enhanced
Energy Efficiency. International Journal of
Computational and Experimental Science and
Engineering, 10(4).
https://doi.org/10.22399/ijcesen.708
[27] K. Yasotha, K. Meenakshi Sundaram, & J.
Vandarkuzhali. (2025). Optimizing Energy
Efficiency and Network Performance in Wireless
Sensor Networks: An Evaluation of Routing
Protocols and Swarm Intelligence Algorithm.
International Journal of Computational and
Experimental Science and Engineering, 11(1).
https://doi.org/10.22399/ijcesen.830
[28] M, P., B, J., B, B., G, S., & S, P. (2024). Energy-
efficient and location-aware IoT and WSN-based
precision agricultural frameworks. International
Journal of Computational and Experimental Science
and Engineering, 10(4).
https://doi.org/10.22399/ijcesen.480
[29] Radhi, M., & Tahseen, I. (2024). An Enhancement
for Wireless Body Area Network Using Adaptive
Algorithms. International Journal of Computational
and Experimental Science and Engineering, 10(3).
https://doi.org/10.22399/ijcesen.409
[30] Kosaraju Chaitanya, & Gnanasekaran Dhanabalan.
(2024). Precise Node Authentication using Dynamic
Session Key Set and Node Pattern Analysis for
Malicious Node Detection in Wireless Sensor
Networks. International Journal of Computational
and Experimental Science and Engineering, 10(4).
https://doi.org/10.22399/ijcesen.613

463