Proceedings of the International Conference on Trends in Material Science and Inventive Materials (ICTMIM-2025)

IEEE Xplore Part Number: CFP250J1-ART; ISBN: 979-8-3315-0148-8

IoT and AI for Efficient

Forest Fire Detection, Monitoring, and Risk Prediction
With Real-Time Alerts
2025 5th International Conference on Trends in Material Science and Inventive Materials (ICTMIM) | 979-8-3315-0148-8/25/$31.00 ©2025 IEEE | DOI: 10.1109/ICTMIM65579.2025.10988101

Dr.M.Janardhan Bhupalam Likhitha Oruganti Purandeswari

Associate Professor, Department Student, Department of Internet of
Student, Department of Internet of
of Artificial Intelligence, G. Things, G. Pullaiah College of Things, G. Pullaiah College of
Pullaiah College of Engineering Engineering and Technology, Engineering and Technology, Kurnool,
and Technology, Kurnool, Andhra Kurnool, Andhra Pradesh, India. Andhra Pradesh, India.
Pradesh, India. [email protected] [email protected]
[email protected]

Poteti Rani Golla Harshini

Student, Department of Internet of Student, Department of Internet
Things, G. Pullaiah College of of Things, G. Pullaiah College
Engineering and Technology, of Engineering and Technology,
Kurnool, Andhra Pradesh, India. Kurnool,Andhra Pradesh,India.
[email protected] [email protected]

Abstract - Wildfires pose severe threats to ecosystems, I. INTRODUCTION

wildlife, and human settlements, necessitating efficient early
detection and monitoring systems. This research introduces The IoT- and AI-enabled Forest Fire Detection and Monitoring
a real-time wildfire detection and risk assessment system System is a groundbreaking advancement in environmental
protection, meticulously designed to revolutionize wildfire
powered by IoT and AI. The proposed system continuously
detection, risk assessment, and emergency response. This intelligent
monitors environmental parameters such as temperature,
system stands as a testament to technological innovation, integrating
humidity, and gas concentrations using IoT-based sensors
cutting-edge IoT and AI capabilities to mitigate the devastating
for early fire detection. AI-driven models analyze both
effects of forest fires through early detection and predictive
historical and real-time sensor data to identify high-risk
analytics.Seamlessly merging multiple technologies, the system
areas, improving predictive accuracy. A cloud-integrated
utilizes IoT for real-time environmental data acquisition, GPS for
architecture facilitates data collection, processing, and precise fire location tracking, and GSM for instant emergency
transmission between sensor nodes, cloud servers, and a notifications. Cloud-based machine learning algorithms analyze
mobile alert system. Fire response teams can leverage the historical and real-time data, identifying high-risk zones and
system’s real-time fire alerts, GPS-based tracking, and enhancing fire prevention strategies. A dedicated mobile application
predictive analytics via a smartphone application, enhancing provides real-time alerts, risk assessments, and actionable insights
emergency response coordination. Extensive testing for forest authorities, ensuring rapid and informed decision-
demonstrates high accuracy, efficiency, and scalability, making.A critical gap in existing solutions is the lack of predictive
significantly reducing detection times while improving capabilities, which often results in reactive rather than proactive fire
resource allocation. Designed for cost efficiency and management. Many traditional systems rely on threshold-based fire
adaptability, this framework presents a proactive strategy for detection, leading to delays and inaccuracies in fire prediction.
wildfire prevention, disaster mitigation, and environmental Furthermore, satellite-based monitoring approaches, while effective
protection. in large-scale surveillance, face challenges in real-time
responsiveness and high operational costs. This study addresses
these limitations by incorporating AI-driven models that evaluate
Keywords - IoT, Artificial Intelligence, Wildfire Detection, environmental conditions and forecast potential fire outbreaks. By
Predictive Analytics, Real-time Monitoring, Disaster merging IoT-driven monitoring with AI-based predictions, the
Management, Cloud Computing, Early Warning System. proposed solution minimizes response times, optimizes resource
allocation, and improves overall wildfire management. This research
explores the technical components, operational mechanisms, and
transformative potential of the IoT- and AI-powered forest fire
monitoring system. The findings contribute to the development of an
efficient, cost-effective, and scalable wildfire detection framework,

979-8-3315-0148-8/25/$31.00 ©2025 IEEE 1324

Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
promoting environmental sustainability and reducing wildfire A. Real-Time Fire Detection and Monitoring
impact through-timely-intervention.
The system employs IoT-based sensors, including
temperature, smoke, and gas sensors, to continuously monitor
II. EXISTING SYSTEM environmental conditions and detect early fire indicators.
These sensors collect real- time data, which is processed by a
The first attempts at autonomous forest fire detection were central unit to identify potential fire hazards. Upon detecting
pioneered by Den Breejen et al. (1998) [1], who proposed one of the abnormal changes, the system triggers an immediate alert to
earliest systems for automating the detection process. This provided the prevent fire escalation and ensure timely intervention.
foundation for subsequent breakthroughs in sensor-based technologies.

With the development of wireless sensor networks (WSNs), B. Mobile Application for Fire Alerts and
scientists such as Yu et al. (2005) [2] and Fernández-Berni et al. (2012) Risk Assessment
[4] also developed real-time detection systems using vision-capable
sensor nodes for low-latency, efficient surveillance of fire-risk areas. A mobile application is integrated into the system to
Likewise, Cheon et al. (2009) [9] also put forth a one-chip CMOS- provide real- time fire alerts, GPS-based fire location tracking,
based smoke and temperature sensor with built-in environmental and predictive risk assessment. The application serves as a
monitoring features in small hardware. direct communication channel for forest officials, offering
them instant notifications about fire occurrences. This feature
Various surveys have sought to classify and analyze fire enables quick decision-making and strategic deployment of
detection strategies. Chowdary and Gupta (2018) [3] offered a wide-
resources to mitigate fire-related risks efficiently.
ranging overview of automated detection methods, including satellite,
terrestrial, and UAV systems. Alkhatib (2014) [6] and Singh and
Sharma (2017) [7] provided insight into systems based on sensor To enhance wildfire prevention, machine learning models
networks, emphasizing their scalability and their limitations in the areas analyze historical and real-time environmental data to predict
of deployment cost and energy usage. high-risk fire zones. These models assess factors such as
temperature trends, humidity levels, and previous fire
The use of machine learning (ML) has significantly improved occurrences to forecast potential fire outbreaks. By leveraging
forest fire prediction. Xu et al. (2021) [5][8] proposed an ensemble predictive analytics, authorities can implement proactive
learning-based system that enhanced the detection rate through the use measures, such as controlled burns and resource pre-
of multiple classifiers. Likewise, Hossain et al. (2021) [12] integrated positioning, to minimize fire hazards.
ML algorithms into WSN systems for fire detection and pattern
identification. Recent articles, including Castelli et al. (2023) [13],
provided a recent overview of ML algorithms applied to wildfire C. Emergency Alert System with GPS-Based Tracking
science, with emphasis on algorithm choice and data quality in
predictive modeling.
In case of fire detection, the system immediately sends
emergency alerts through GSM, ensuring that authorities
The Internet of Things (IoT) has been instrumental in the
receive precise fire location data via GPS. This feature enables
detection of wildfires. Rajan et al. (2022) [10] created an IoT-based
emergency responders to take swift action, facilitating faster
system combined with machine learning to provide real-time fire alarms
containment of wildfires and minimizing damage to forests and
and forecasts. Building on this, Bonilla-Ormachea et al. (2025) [14]
nearby communities. The integration of GPS tracking also aids
suggested ForestProtector, which combines machine vision and deep
in efficient coordination among response teams.
reinforcement learning within a smart structure for effective wildfire
surveillance. Adding automation further, Bouguettaya et al. (2024) [15]
wrote about a system involving AI-driven drones and IoT sensors for D. Cloud Based Data Synchronization with Firebase
early warning and quick response, pointing to a trend towards
autonomous surveillance in fire-risk areas. The system utilizes cloud computing through Firebase to
synchronize real-time data from IoT sensors, mobile
Bot and Borges (2022) [11] showed an innovative deep applications, and other connected devices. This cloud-based
learning technique underpinned by learning without forgetting to enable approach ensures seamless communication and data access
long-term adaptation for forest fire detection systems. Such methods across multiple stakeholders, including forest officials, disaster
handle issues such as model drift in changing environments. management teams, and environmental agencies. Storing data in
the cloud allows for historical trend analysis and enhances
Forest fire detection has evolved from basic sensor-based collaboration in wildfire management.
systems to smart, adaptive systems leveraging machine learning, deep
learning, and IoT connectivity. Although basic WSNs offer low-cost E. Iot-Enabled Environmental Monitoring and
coverage, AI-based architectures promise enhanced accuracy,
scalability, and autonomy. The increasing convergence of edge Analysis
computing, vision systems, and autonomous drones holds a bright
future for anticipatory wildfire management.. Beyond fire detection, the system continuously tracks
environmental conditions to provide valuable insights for
long-term analysis. The collected data supports government
III. PROPOSED SYSTEM and environmental organizations in refining wildfire
prevention policies and improving resource allocation. By
The IoT- and AI-enabled Forest Fire Detection and Monitoring leveraging IoT and AI, the system enables data- driven
System integrates IoT, AI, and machine learning to provide real-time decision-making, contributing to effective fire management
fire detection, predictive risk assessment, and intelligent emergency and enhanced environmental protection.
response. Designed for forest environments, the system enhances
wildfire prevention, rapid intervention, and disaster management. The To minimize the devastating impact of forest fires, a smart
primary components of the proposed system include:
1325
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
wildfire detection and monitoring system has been developed, B. Integration of GSM
integrating Internet of Things (IoT) and Artificial Intelligence (AI). The
system utilizes an ESP32 microcontroller to process sensor data, detect
The GSM (Global System for Mobile Communications)
early fire indicators, and trigger timely alerts. It continuously monitors
module shown in fig 1.6 serves as a vital component in this
temperature, gas concentration, and fire presence to assess potential
project, enabling wireless communication for real-time alerts.
wildfire threats. The system ensures real-time tracking and predictive
It ensures that emergency notifications, such as fire or gas
analytics, allowing authorities to take proactive measures. Field tests
leak alerts, are promptly transmitted to relevant authorities or
validate its effectiveness in identifying fire risks, delivering instant
users via SMS. This capability is particularly beneficial in
notifications, and enhancing situational awareness. This cost-efficient
areas with limited internet access, where traditional Wi- Fi-
and scalable solution plays a crucial role in strengthening forest fire
based systems may be unreliable. The GSM module, when
management and environmental protection.
integrated with the ESP32 microcontroller, facilitates
seamless data transmission without requiring a continuous
The Smart Wildfire Prevention System integrates multiple internet connection. Additionally, it operates with low power
functionalities to provide a holistic approach to fire detection and consumption and ensures broad network coverage, making it
emergency response. By addressing key environmental risk factors and a dependable communication tool for emergency responses in
enabling rapid intervention, it aims to reduce wildfire damage, improve this IoT-based fire detection system.
response coordination, and support conservation efforts.

Designed for ease of use and affordability, this system serves C. Sensors Network for Fire Detection
as a practical tool for forest officials, emergency responders, and
environmental agencies. Accurate risk detection is achieved through Sensor technology plays a crucial role in the early and
advanced fire, gas, and climate sensors, alongside efficient data accurate detection of wildfires. The system utilizes three main
processing via the ESP32 microcontroller. The system effectively types of sensors: flame sensors, gas sensors (MQ series), and
identifies fire hazards, monitors air quality, and analyzes environmental temperature & humidity sensors (DHT11). With these
patterns. Machine learning-driven analytics and real- time data integrated sensors, the system delivers high accuracy and
transmission further enhance its efficiency in early fire detection and enables a rapid response. By continuously monitoring
emergency preparedness. By combining IoT, AI, and cloud-based environmental factors and providing real-time data, the system
intelligence, this system empowers authorities with reliable insights to can identify fire risks as soon as they arise, allowing emergency
safeguard forests and communities. responders to act swiftly. The multi-sensor approach enhances
detection precision and accelerates decision-making,
ultimately leading to more effective wildfire management and
minimized damage.

D. Integration of Flame Sensor

The flame sensor shown in fig 1.3 detects infrared radiation

emitted by flames, enabling the system to quickly identify
fires. This early detection reduces response time and helps
contain fires before they spread. The fire sensor operates
based on infrared radiation detection in the 760– 1100 nm
range, allowing it to identify flames even in low-visibility
conditions. It uses a photodiode and a signal amplification
Fig 1. Proposed Block Diagram circuit to detect IR intensity changes. The ESP32 continuously
monitors the sensor's output voltage and triggers an alert if it
exceeds a predefined threshold.

E. Integration of Gas Sensor

IV. HARDWARE COMPONENTS
The MQ135 gas sensor identifies hazardous gases such as
A. ESP32 Microcontroller carbon monoxide (CO) and smoke, which are released during
combustion. These gases serve as early indicators, allowing
The ESP32 microcontroller shown in fig 1.1 is a highly efficient detection of fires before flames become visible, especially in
core component in wildfire detection systems, offering powerful remote or less accessible areas. The smoke sensor utilizes a
performance with its dual-core architecture that enables simultaneous metal oxide semiconductor (MOS) gas-sensing layer that
data processing from multiple sensors, such as flame, gas, temperature, reacts with airborne combustion particles, altering its
and humidity sensors. Its Wi-Fi and Bluetooth connectivity allow for resistance. The ESP32 reads the sensor output via an analog-
both remote monitoring and local communication, transmitting critical to-digital converter (ADC) and applies a thresholding
data to cloud platforms and providing immediate alerts to emergency mechanism to distinguish between normal atmospheric
teams. Optimized for low power consumption, the ESP32 ensures long- variations and hazardous smoke levels. A moving average
term, reliable operation in remote areas, often powered by solar energy filter helps smooth out noise in sensor readings
or batteries. By processing environmental data in real-time, the ESP32
can quickly detect wildfire indicators and trigger alerts, enhancing early
response times. Additionally, it integrates seamlessly with IoT F. Integration of DHT11
technology, continuously collecting and transmitting data for consistent
wildfire monitoring, making it an ideal solution for efficient, scalable, The DHT11 sensors shown in fig 1.4 monitors
and proactive wildfire management in challenging environments. temperature and humidity levels, key factors in predicting fire
risk. By detecting high temperatures and low humidity, the
1326
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
sensor provides early warnings of potential fire hazards and system and machine learning models used in this fire
identifies high-risk areas. The DHT11 sensor measures ambient detection and monitoring project.
temperature (range: 0–50°C, accuracy: ±2°C) and humidity (range:
20– 90%, accuracy: ±5%). It uses a capacitive humidity sensor and a C. Fire Base
thermistor to generate digital signals via a single-wire
communication protocol. The ESP32 periodically samples this data Firebase serves as a cloud-based backend solution for
and transmits it to Firebase for real-time analysis. The ML model real- time data storage, synchronization, and communication
later retrieves historical data to predict fire-prone conditions. between various components of this project. It efficiently
manages sensor data collected from the ESP32, ensuring that
G. Integration of GPS critical fire, smoke, temperature, and GPS location information
is instantly available for analysis and response.The integration
The GPS module shown in fig 1.2 ensures accurate location of Firebase enables seamless connectivity between the
tracking, allowing precise identification of fire-prone zones. Upon hardware and the mobile application, providing real-time alerts
detecting smoke or fire, the ESP32 retrieves GPS coordinates and and status updates to users and emergency responders. Its
transmits them via GSM and Firebase, enabling authorities to pinpoint robust authentication and database management features
the exact location for rapid intervention. The module continuously enhance security and reliability, ensuring that only authorized
monitors position changes, ensuring high accuracy in tracking fire users can access sensitive data. Additionally, Firebase’s
outbreaks and assisting in efficient resource deployment for scalability makes it an ideal choice for handling large amounts
firefighting efforts. Its high sensitivity and fast signal acquisition of sensor data, supporting predictive analytics and intelligent
enhance real-time fire monitoring and response efficiency. decision-making within the system.

D. Iot Technology Integration

IoT technology forms the foundation of this system,

enablingseamless connectivity between hardware components for
real-time monitoring, data processing, and decision-making. By
integrating IoT, the project ensures that sensor data from fire,
smoke, and environmental monitoring modules is transmitted
efficiently to cloud storage and mobile applications.The IoT
framework enhances the system’s ability to provide instant alerts,
remote monitoring, and predictive insights, improving response
times in emergency situations. Its role in automating data
collection and analysis reduces human intervention while
ensuring accuracy and reliability. This smart integration of IoT
makes the system highly efficient for forest fire detection and
environmental safety monitoring.

E.Machine Learning Integration

Machine learning (ML) plays a crucial role in enhancing

V. SOFTWARE COMPONENTS the intelligence of the system by enabling data-driven
decision-making for forest fire detection and risk assessment.
A. MU Editor By analyzing real-time sensor data, ML models can recognize
patterns, detect anomalies, and anticipate potential fire
Mu Editor is a simple yet effective code editor specifically hazards. The system continuously processes environmental
designed for MicroPython development, making it an excellent parameters such as temperature, humidity, and gas
choice for programming the ESP32 microcontroller in this project. It concentration to enable early fire detection and rapid
provides an easy-to-use interface with built-in support for response, thereby improving disaster management efforts.
MicroPython, allowing seamless coding, testing, and debugging of
sensor data processing.The editor's real-time REPL (Read-Eval- For this project, the Random Forest algorithm is selected
Print Loop) functionality enables direct interaction with the ESP32, due to its strong capability in handling nonlinear and complex
facilitating efficient debugging and code execution. Its lightweight environmental data. As an ensemble learning method,
nature and compatibility with IoT-based applications make it a Random Forest constructs multiple decision trees and
suitable tool for developing and deploying embedded software, combines their outputs to improve prediction accuracy while
ensuring smooth integration of sensors and real-time monitoring in reducing the risk of overfitting. This makes it particularly
the fire detection system. effective in analyzing diverse environmental conditions that
influence fire occurrence. By utilizing historical data,
B. Visual Studio Code including temperature variations, humidity levels, gas
concentration trends, and past fire incidents, the model can
classify areas into low-risk, moderate- risk, and high-risk fire
Visual Studio Code (VS Code) is a lightweight yet powerful
zones with high reliability.
code editor used in this project for developing and debugging IoT
and machine learning components. It provides an efficient
One of the key strengths of Random Forest is its
environment for writing MicroPython code for the ESP32
resilience to missing or noisy sensor data, which is a common
microcontroller, ensuring seamless sensor data processing and
challenge in IoT-based monitoring systems. Its ensemble
communication with cloud services.With built-in extensions and
nature ensures that predictions remain stable even if some data
support for multiple programming languages, VS Code enhances
points are inconsistent. Additionally, the algorithm allows for
productivity by offering syntax highlighting, intelligent code
feature importance analysis, helping to identify the most
completion, and version control integration. Its flexibility makes it
influential environmental factors contributing to fire risks. To
an ideal choice for writing, testing, and optimizing the embedded
evaluate the model’s reliability, standard machine learning
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
1327
 performance metrics such as precision, recall, F1-score, and ROC- VIII.Results & Discussion
AUC are used, ensuring robust fire risk classification.Over time, the
system continuously improves by learning from new data,
increasing detection accuracy while reducing false alarms. This When a fire is detected, our IoT and AI-powered
adaptability makes the fire monitoring system more proactive and Forest Fire Detection System initiates an automated response
efficient in disaster prevention. Future enhancements may include to ensure a swift and effective reaction. The GSM module is
hybrid models combining deep learning with ensemble methods, responsible for sending an SMS alert to forest officials,
further refining prediction capabilities. Additionally, integrating providing the exact GPS coordinates of the fire location. This
edge computing can enhance local decision-making, reducing enables rapid intervention by authorities, ensuring immediate
dependency on cloud-based processing. By embedding ML-driven action to mitigate the risk. Figure 2.1 illustrates the Fire Alert
intelligence, this system offers a scalable and adaptive solution for SMS sent to forest officials upon detection and location can
forest fire management, strengthening early warning mechanisms be viewed like that in fig 2.2
and optimizing emergency response strategies.

VII. QUANTITATIVE ANALYSIS

To assess the performance of the proposed IoT and AI-

powered forest fire detection system, a comprehensive quantitative
evaluation was conducted. The analysis focused on key performance
indicators, including detection accuracy, response time, predictive
reliability, and system stability—crucial aspects for early fire
detection and risk assessment.The accuracy of fire and gas detection
was determined by comparing real-time sensor readings with actual
environmental conditions. The system’s ability to precisely detect
Fig 2.1 Fire Alert SMS to Forest Officials
temperature fluctuations, smoke levels, and gas concentrations was
analyzed using standard evaluation metrics, including detection rate
and reduce false alarms.To enhance predictive accuracy, the
system employs the Random Forest machine learning algorithm for
fire risk assessment. Random Forest, an ensemble learning method,
constructs multiple decision trees and aggregates predictions to
improve classification accuracy and mitigate overfitting.The model
was trained using historical environmental data encompassing
temperature, humidity, gas concentration, and past fire occurrences.
The algorithm’s performance was evaluated using precision, recall,
F1-score, and ROC-AUC to ensure robust classification of areas
into low-risk, moderate-risk, and high-risk zones.

Response time is a critical factor in real-time

fire detection. The system's efficiency was measured by evaluating the
ESP32 microcontroller’s processing speed in detecting fire indicators
and triggering alerts via the GSM module. The latency in transmitting
data to Firebase and delivering notifications to the mobile application
was also assessed. The study analyzed the synchronization time
between Firebase updates and mobile app alerts to ensure real-time
communication. The system’s reliability was tested under varying
environmental conditions, including normal weather fluctuations and
simulated fire scenarios. The accuracy of the GSM and GPS modules
in delivering alerts and location coordinates was examined to ensure
minimal delays. Furthermore, stress testing was performed to evaluate
the system’s resilience in prolonged monitoring operations.
Fig 2.2 Gps Location of the forest fire

The results of this quantitative analysis confirm that the

Simultaneously, the system updates Firebase Cloud,
proposed system efficiently detects fire hazards, provides timely alerts,
triggering an instant notification in the forest officials' mobile
and supports predictive risk assessment. The integration of real-time
application. This Android app displays critical sensor data,
sensor data processing, machine learning-based risk prediction, and
including temperature, humidity, gas concentration, and fire sensor
cloud-based communication enhances early detection capabilities.The
values, allowing officials to monitor environmental conditions
Random Forest-based predictive analytics significantly improves the
remotely. Figure 2.3 presents a sample Fire Detection Notification
identification of high-risk zones, enabling a proactive approach to
received on the app.
forest fire management. Future enhancements could explore
optimizing the machine learning model with additional environmental
features and integrating edge computing for faster local
processing.This multi-layered approach ensures improved detection
accuracy, reduced false alarms, and faster response times, making it a
viable solution for mitigating forest fire risks.

1328
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
 By integrating IoT technology, machine learning, and
real- time alerts, the Forest Fire Detection System enhances early
fire detection, improves emergency response, and provides
predictive insights to prevent large-scale wildfires. This proactive
approach significantly strengthens wildfire management, ensuring
faster interventions and minimizing damage.

IX. CONCLUSION

Our Forest Fire Detection and Prevention System

Fig 2.3 Fire Detection Notification leverages IoT and AI to enhance wildfire management through
real-time monitoring, predictive analytics, and automated alerts.
The machine learning model continuously analyzes sensor Fire, gas, and DHT11 sensors continuously collect environmental
data to predict potential fire risks. If a high-risk situation is detected, data, storing it in Firebase. When fire, smoke, or abnormal gas
such as rapid temperature rise or increased gas levels, the system levels are detected, the ESP32 microcontroller sends instant alerts
updates Firebase, which in turn triggers a notification in the forest via the GSM module with precise GPS coordinates, enabling
officials' app. The notification provides messages like "Forest at high swift response. A machine learning model, using the Random
risk: Temperature exceeds safe limits", enabling proactive decision- Forest algorithm, analyzes historical data to identify high-risk
making. Figure 2.4 showcases the Fire Risk Prediction Alert displayed areas, helping authorities take preventive measures. The system
in the app. integrates seamlessly with an Android app, ensuring continuous
updates for better decision-making. Its scalability, cost-
effectiveness, and real-time capabilities make it a practical
solution for forest conservation. Future enhancements could
include drone surveillance and edge computing for faster local
processing. By detecting, predicting, and alerting in real time,
this system plays a vital role in preventing wildfires and
preserving ecosystems.

X. REFERENCES
[1] E. Den Breejen, M. Breuers, F. Cremer, R. Kemp, M.
.Roos, K.Schutte, and J.S. De Vries, “ Autonomous forest
fire detection” Coimbra, Portugal: ADAI-Associacao para o
Desenvolvimento da Aerodinamica Industrial, pp . 2003-
2012, 1998
Fig 2.4: Fire Risk Prediction Alert [2] L. Yu, N. Wang, X. Meng, “Real-time forest fire detection
with wireless sensor networks,” International Conference on
Additionally, the real-time monitoring Wireless Communications, Networking and Mobile
dashboard in the mobile application provides continuous updates on Computing, vol. 2, pp. 1214-1217, 2005.
sensor readings, allowing forest officials to track temperature, [3] V. Chowdary, and M. K. Gupta, “ Automatic forest fire
humidity, gas levels, and fire sensor data. This ensures constant detection and monitoring techniques: a survey,” Intelligent
monitoring and quick response to any unusual environmental Communication, Control and Devices, pp. 1111-1117, 2018.
changes. Figure 2.5 demonstrates the Live Sensor Data Monitoring [4] J. Fernández-Berni, R. Carmona-Galán, J. F.Martínez-
Dashboard. Carmona, and Á. Rodríguez-Vázquez, “ Early forest fire
detection by vision- enabled wireless sensor networks,”
International Journal of Wildland Fire, vol. 21, no. 8, pp.
938-949, 2012.
[5]. R.Xu, H.Lin, K.Lu, L Cao,and and Y.Liu, “A forest fire
detection system based on ensemble learning,” Forests, vol.
12, no .2, pp. 217 225, 2021.
[6] A. A. Alkhatib, “A review on forest fire detection
techniques,” International Journal of Distributed Sensor
Networks, vol. 10, no
.3, pp. 1-11, 2014.
[7] P. K. Singh, and A. Sharma, “An insight to forest fire
detection techniques using wireless sensor networks,”
International conference on signal processing, computing
and control, pp. 647- 653, 2017.
[8] R.Xu, H.Lin, K.Lu, L Cao,and and Y.Liu, “A forest fire
detection system based on ensemble learning,” Forests, vol.
12, no .2, pp. 217 225, 2021.
[9] J. Cheon, J. Lee, I. Lee, Y. Chae, Y. Yoo, and G. Han, “A
single- chip CMOS smoke and temperature sensor for an
intelligent fire detector,” IEEE Sensors Journal, vol. 9, no .8,
pp . 914-921, 2009.
[10] Rajan, B.; Sharma, A.; Kumar, R. "IoT-Based Forest Fire
Detection and Prediction System Using Machine Learning."
Proceedings of the International Conference on Smart Sensors
Fig 2.5 Real-Time Sensor Data Display and Systems (ICSSS), 2022.
1329
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.
[11] K. Bot and J. G. Borges, “Forest fire and smoke detection using
deep learning-based learning without forgetting,” Fire Ecology,
vol. 18, no. 2, pp. 1-12, 2022.
[12] S. M. Hossain, M. A. Hossain, and M. S. Rahman, “Forest fire
detection system using wireless sensor networks and machine
learning,” Scientific Reports, vol. 11, no. 3, pp. 2145-2156, 2021.
[13] M. Castelli, L. Vanneschi, and A. Popović, “A brief review of
machine learning algorithms in forest fires science,” Applied
Sciences, vol. 13, no. 14, pp. 8275-8288, 2023.
[14] K. Bonilla-Ormachea, H. Cuizaga, E. Salcedo, S. Castro, S.
Fernandez-Testa, and M. Mamani, “ForestProtector: An IoT.

[15] A. Bouguettaya, H. Zarzour, A. M. Taberkit, and A. Kechida,

“Towards early forest fire detection and prevention using AI-
powered drones and the IoT,” Internet of Things, vol. 8, no. 1,pp.
1896-1912, 2024

1330
Authorized licensed use limited to: Zhejiang University. Downloaded on July 20,2025 at 14:57:57 UTC from IEEE Xplore. Restrictions apply.