Forest Fire Prediction System Using Node MCU

Saptarshi Pramanik Sambit Chanda (Prof.) Dr. Suparna Biswas Shibam Sanyal
Electronics and Communication Electronics and Communication Electronics and Communication Electronics and Communication
Engg. Engg. Engg. Engg.
Guru Nanak Institute of Guru Nanak Institute of Guru Nanak Institute of Guru Nanak Institute of
Technology Technology Technology Technology
Sodepur, India Sodepur, India Sodepur, India Sodepur, India
[email protected] [email protected] [email protected] [email protected]

on environmental conditions. A wildfire, forest fire, bushfire,

Abstract—This project aims to predict and detect wildland fire or rural fire is an unplanned, uncontrolled and
forest fires using an IoT-based system with Node
MCU. Forest fires, uncontrolled and non- unpredictable fire in an area of combustible vegetation starting
in rural and urban areas. Some forest ecosystems in their natural
prescribed combustion of vegetation, are a state depend on wildfire.
significant concern due to their unpredictability
and the widespread damage they can cause. The In middle latitudes, the most common human causes of
project is motivated by the need for advanced wildfires are equipment generating sparks (chainsaws, grinders,
firefighting systems, especially in environments mowers, etc.), overhead power lines, and arson. In the tropics,
where human intervention is not effective. farmers often practice the slash-and-burn method of clearing
fields during the dry season.

The system uses a temperature sensor and a DHT When thousands of farmers do this simultaneously, much
11 humidity sensor interfaced with Node MCU to of a continent can appear from orbit to be one vast blaze. Coal
detect temperature and humidity changes seam fires burn in the thousands around the world, such as those
indicative of a fire. The sensor data is uploaded in Burning Mountain, New South Wales; Centralia,
and an analysis is conducted based on threshold Pennsylvania; and several coal-sustained fires in China. They
can also flare up unexpectedly and ignite nearby flammable
values to detect a potential fire. material.

These models, which are based on the forest Economic growth in modern industrialized societies has
structure GIS (geographical information system), resulted in factories, complex office buildings, and dense
are used to predict wildfire potential under apartment blocks located in metropolitan areas. Associated gas
various climate futures. Maps are created to stations and oil reservoirs, which are all vulnerable to fire due
to the flammable substances they house, are also found in these
identify areas more susceptible to wildfires, areas. When a fire occurs in such places, firefighting is hindered
enabling targeted management to reduce wildfire by the mazes of crowded buildings, high temperatures, smoke,
chances. and the danger of explosions Current firefighting systems are
based on humans using deluge guns and chemical fire
Keywords—Forest fire prediction, IoT based system, neural repression systems. However, in environments where humans
networks, logistic regression, decision trees, forest structure GIS, cannot work effectively, it is desirable to extinguish a fire
climate futures, SPIN protocol, Tiny OS, remote sensing, emergency quickly using fire-fighting robots. Recently, in order to cope
plan. with catastrophic fire related accidents, research on fire-
fighting robots has advanced in many countries [1].
I. INTRODUCTION
II. LITERATURE SURVEY
The Forest Fire bush or vegetation fire, can be described as
any uncontrolled and non-prescribed combustion or burning of
plants in a natural setting such as a forest, grassland, brush land This comprehensive survey explores machine learning
or tundra, which consumes the natural fuels and spreads based algorithms for forest fire prediction and detection. It covers
various methods and systems, including those based on neural

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

networks, logistic regression, decision trees, and their temporary prototypes and circuit experiments, making them a
ensembles. The study also discusses factors influencing fire popular choice in educational settings.
occurrence and risk assessment [2].
In this research, the authors present an integrated system that
combines IoT (Internet of Things) technology with Node-MCU Earlier versions of breadboards were not reusable. In
for forest fire detection and prediction. The paper provides contrast, stripboards and other prototyping printed circuit
insights into the design and testing of this system. Focusing on boards, which are used for semi-permanent soldered prototypes,
core design criteria and peer-to-peer networking, this study are not easily reusable. Breadboards can accommodate a variety
introduces a low-cost IoT device prototype for permanent on- of electronic systems, from small analog and digital circuits to
site forest fire monitoring. The system leverages Node-MCU complete central processing units (CPUs).
and aims to enhance forest fire detection capabilities [3]. Despite their advantages, modern breadboards have some
limitations compared to permanent circuit connections. They
III. PAPER OVERVIEW have high parasitic capacitance, relatively high resistance, and
Forest fire is getting worse for all these days which can be less reliable connections that are susceptible to physical
detected and predicted using Node-MCU based on IoT. In this degradation and movement. Their signalling capabilities are
project, a temperature sensor, DHT 11 humidity sensor is limited to around 10 MHz, and even at frequencies below that,
interfaced to Node MCU detects the temperature and humidity not all functions operate optimally [6].
produced from the fire. The values are taken from the Sensor
and is uploaded to the cloud i.e., in Thing speak. Using threshold
value, the temperature and humidity value is taken into
consideration and an analysis is made to detect fire. When it is
detected, an email triggered with an alert message with the
sensor values and the fire is predicted [4][5].
The System is designed by using Node MCU, Soil Moisture
Sensor, Temperature Sensor and DHT 11 Humidity Sensor. We
can predict and detect fire in the Forest. We can also travel with
it. With the sensors we can measure the humidity of the air and
the moisture level in the soil.
When humidity will fall down then the chance of forest fire
will rise. Similarly, when light will rise then the resistance of the
LDR will drop and the current will flow and we will get to know
about the fire. Figure 1. Bread Board

In the project of Forest Fire Prediction, we are willing to alert B. Node MCU
the people who are surrounded by the forest that they should The Node MCU is an open-source firmware and prototyping
leave the place as soon as possible. By our project we can also board design. The name "Node MCU" is a combination of
predict the fire in the forest. Thus, we can save the human's life "node" and "MCU" (micro-controller unit), and it specifically
as well as the animal’s life who lives in the forest. By automating refers to the firmware, not the development kits. Both the
the process, becomes more accurate, cost effective and firmware and the board designs are open source.
predictable. The use of this project will help us to make sure that
the people around the forest are safe. The firmware is built on the Eula project using the
Expressive Non-OS SDK for ESP8266, and it uses the Lua
In environments where humans cannot work effectively, it is scripting language. It includes various open-source projects,
desirable to extinguish a fire quickly using fire-fighting robots. such as leucosin and SPIFFS.
Recently, in order to cope with catastrophic fire related
accidents, research on fire-fighting robots has advanced in many For prototyping, a circuit board in a dual in-line package
countries [1]. (DIP) format is typically used. This board combines a USB
controller with a smaller surface-mounted board that houses the
IV. COMPONENTS USED MCU and antenna. The DIP format makes prototyping on
Certain components are used to make this setup successful. breadboards easy. The design, initially based on the ESP-12
module of the ESP8266, features a Wi-Fi System-on-Chip
A. Breadboard (SoC) integrated with a Ten silica Extensa LX106 core, which
The A breadboard, also known as a protoboard, is a platform is commonly used in IoT applications [6].
used for building prototypes of electronic circuits. The term
originally referred to a wooden board used for cutting bread.
However, in the 1970s, the solderless breadboard was
introduced, and it has since become the standard for reusable
prototyping platforms. Unlike traditional methods that require
soldering, solderless breadboards allow for the easy assembly of

Identify applicable funding agency here. If none, delete this text box.
 conductivity. These sensors find applications in agriculture and
remote sensing within hydrology. They help monitor soil
moisture [6].

Figure 2. Node MCU

C. IC (74 HC CD4052)
The CD4052 is a versatile IC used for electronic circuit Figure 4. Soil Moisture Sensor
design. It can function as a 4:1 multiplexer (MUX), taking
inputs from four channels and converting them to a single E. Temperature and Humidity Sensor (DHT 11)
output, or as a 1:4 demultiplexer (DEMUX), taking one input The DHT-11 Digital Temperature and Humidity Sensor is
and providing it to one of four output channels. The selection an affordable and straightforward device. It combines a
of output depends on the settings of pins A and B. This IC can capacitive humidity sensor and a thermistor to measure air
handle both analog and digital voltages, making it suitable for conditions. The sensor provides a digital output on its data pin,
various applications [6]. eliminating the need for analog input pins. However, it’s
essential to time data retrieval carefully since new data is
available only every 2 seconds. Compared to the DHT22, the
DHT-11 is less precise, less accurate, and has a narrower
temperature/humidity range. Despite these drawbacks, it
remains a cost-effective choice for temperature and humidity
measurements [6].

Figure 3. IC (74 HC CD4052)

D. Soil Moisture Sensor

Soil moisture plays a crucial role in both irrigation and
garden settings. Nutrients in the soil nourish plant growth, and
supplying water is essential for regulating plant temperature. Figure 5. DHT 11
Transpiration, where water affects plant temperature, is a key
process. Additionally, plant root systems thrive in moist soil. F. LDR
However, extreme soil moisture levels can lead to anaerobic An LDR (Light Dependent Resistor) is an electronic
conditions, impacting plant growth and soil health. component whose resistance changes based on the intensity of
light falling on its surface. It is commonly used in light-sensing
Now, let's delve into the soil moisture sensor, which gauges circuits. The typical LDR exhibits a resistance that decreases as
the volumetric water content within the soil. Unlike direct light intensity increases, as depicted in the image above.
gravimetric measurements (which involve drying and sample Specifically, in daylight, the LDR’s resistance is around 5000
weighing), these sensors infer water content using properties ohms. In dark conditions, the resistance increases significantly
like dielectric constant, electrical resistance, and interaction to approximately 20,000,000 ohms [6].
with neutrons. Keep in mind that the relationship between
calculated properties and soil moisture may vary based on
factors such as temperature, soil type, and electrical
 and soil moisture, it can anticipate fire risk. Additionally,
changes in light intensity can trigger fire detection through the
LDR sensor. The goal is to alert people near forests to evacuate
promptly in case of fire, ultimately saving human and animal
lives. Automating the process enhances accuracy, cost-
effectiveness, and predictability, ensuring the safety of forest-
edge communities. Forest fires, whether termed wildfires,
bushfires, or rural fires, pose significant risks, and this project
aims to mitigate those dangers.

The following flowchart will describe the working of the

device, more effectively and efficiently.
Figure 6. LDR VI. DEVICE PICTURE

G. Jumper Wires
A jumper wire, also known as a DuPont wire, is an electrical
cable or group of wires with connectors or pins at each end.
These wires are commonly used to interconnect components on
a breadboard or other prototype circuits without soldering. By
bundling the wires into a non-flexing arrangement, space is
optimized, and the risk of short circuits is minimized.

In computers, a jumper consists of a pair of prongs set into

the motherboard or an adapter card. When you set a jumper,
you complete an electrical contact by placing a plug on the
prongs. Essentially, jumpers act as switches, either closing or Figure 8. Device
opening an electrical circuit. They can be added or removed to
modify the function or performance of PC components. A VII. FLOWCHART
collection of jumpers is sometimes referred to as a jumper
block. START

While modern computers often come with pre-set jumpers

due to plug-and-play functionality, some manufacturers Search for Wi-Fi
provide jumper settings in the instruction manual for users who network
wish to customize performance [6].

Connected to Wi-Fi

Connect to google
app-script server

Fetch commands from

server by get request

Figure 7. Jumper Wires Collect the data from

the LDR, soil moisture,
temperature sensor and
store it in google excel
V. DESCRIPTION
The system utilizes Node MCU, Soil Moisture Sensor,
Temperature Sensor, and DHT11 Humidity Sensor to predict DELAY
and detect forest fires. By monitoring humidity levels in the air
 X. CIRCUIT ANALYSIS
VIII. BLOCK DIAGRAM For the circuit we used node MCU board, bread Board,
one 74 HC 4052 IC, soil moisture sensor and a temperature
The below block diagram will help to understand the circuit sensor. First of all, we connected the voltage and ground pin of
diagram and circuit analysis more effectively and efficiently. Node MCU with the following voltage and ground pin of the
The block diagram used here is only used as a reference to also bread board. Then we connect the voltage and ground pin of the
understand the work-flow or data-flow. sensors with the voltage and ground line of bread board. We
have used an IC by which we converted the digital pins of node
MCU to an analogue pin, for that we connected the E, Vee and
GND pin with the ground line and the VCC pin with the voltage
line of bread board. Then do, d1 and AO pin of node MCU is
connected with the So, S1 and 1 Z pin of the IC simultaneously.
Then the analogue pin of sensors will have connected with the
Node MCU 1y1 and 1yo analogue pin of the IC.

Wi-Fi XI. SETUP

connected

Connect to
server

LDR Soil Temperature

sensor Moisture sensor
sensor

Store the data in the excel sheet of

google server Figure 10. Setup Picture

XII. ALGORITHM
IX. CIRCUIT DIAGRAM To understand the algorithm or functioning of the device
we need some mappings or keywords first –
• St – Start.
• E – End.
• Io1 – Input/Output: Define libraries, constants, pins,
variables).
• Io2 - Input/Output: Initialize serial communication at
9600 baud rate.
• Io3 – Input/Output: Connect to Wi-Fi network using
SSID and password.
• Op1 – Operation: Wait until Wi-Fi is connected.
• Io4 – Input/Output: Print Wi-Fi is connected.
• Op2 – Operation: Set client into insecure mode.
• Op3 – Operation: Set pin modes for sensors and
multiplexer.
• Op4 – Operation: Set multiplexer to select soil
Figure 9. Circuit Diagram moisture sensor.
• Op5 – Operation: Initialize DHT 11 sensor.
• Op6 – Operation: Check if last update time is greater
than equals to interval.
 • Op7 – Operation: Read analog value from soil XIII. RESULT ANALYSIS
moisture sensor and convert to percentage.
• Op8 – Operation: Set multiplexer to select LDR.
• Op9 – Operation: Read analog value from LDR.
• Op10 – Operation: Read humidity and temperature
values from DHT 11 sensor in both Celsius and
Fahrenheit values.
• Op11 – Operation: Call send-data function.
• Op12 – Operation: Reset last update time to zero.
• Op13 – Operation: Wait for 1000 ms.
• Op14 – Operation: Increment last update time by 1000
ms.
• Op15 – Operation: Check if client can connect to host
and port.
• Io5 - Input/Output: Print host is connected.
• Op16 – Operation: Construct query string with data
value as parameters. Figure 11. Graph of the values collected from sensors
• Op17 – Operation: Send GET request to web app
using path and query string. The following values collected from the sensors (in real-
• Io6 - Input/Output: Print data is sent along with query time) are stored in google app-script server in google excel
string. sheets and based on those data, the following upper mentioned
• Io7 - Input/Output: Print error in connecting the host. graphs are plotted automatically.
• Cond1 – Condition: Wi-Fi connected?
• Cond2 – Condition: Check if last update time is • Temperature sensor graph – The temperature sensor
greater than equals to interval? graph will show a spike when there is sudden rise in
the surrounding temperature. Here the temperature is
• Cond3 – Condition: Client connected to host?
normal that is why there is a straight base-line. While
forest fire as the surrounding temperature will rise
Following is the path-flow on which the algorithm will
suddenly above its threshold it will indicate a situation
work.
of forest fire.
Op1 (no)
St
• Humidity sensor graph – In the humidity sensor graph
Io1 Io2 Io3 Op1 Op1 (yes) Io4 here we see a sudden dip. This is because when there
is a fire in the surrounding there will be sudden dip the
humidity of the surroundings below the normal
Op6 Op5 Op4 Op3 Op2 humidity level.

• LDR sensor graph – In the LDR sensor graph there we

Op6 (yes) Op7 Op8 Op9
see a sudden hike in the graph that is because when
there is a forest fire there will be huge light emission
Op10
Op6 (no) Op13 which will result in the spike of the LDR graph. Thus,
will indicate a possibility of forest fire.
Op14 Op11
• Soil moisture sensor graph – When there is very low
Op17 Op16 soil moisture level the graph will show a very low
Op12 reading. Thus, it will indicate that the soil is dry, as
well as the surrounding is dry. Thus, it will show a
Io6 Io5 Op15 (yes) Op15 possibility of forest fire.

Op15 (no) XIV. APPLICATIONS

These are the following applications of forest fire
Io7 prediction system –
• It helps to prevent forest fires by estimating the
probability of fire occurrence based on various
factors.
 • It enables the analysis and forecasting of the extent future. This project shows that there is still a need for a good
and severity of fire damage using past data on burnt networking component for Tiny OS. Future works may focus
areas. on improving the protocol that was developed in this project
and creating new networking components for the system [11].
• It monitors and reports the weather conditions in
the region that may affect the fire behavior and Wildfires can reach high into the tree tops and destroy large
spread. parts of the forest. Sometimes people start these fires, and they
can have a big impact on the economy and the environment.
• It alerts the authorities and the public about the The changing climate, with different rain patterns and higher
presence and location of forest fires [7][8]. average temperatures, may make wildfires more frequent. As
more people live in or near forested areas in the Southern
United States, the risk of damage from wildfires may increase
XV. FUTURE SCOPES [12]. Some wildland fires are good for the environment and
some ecosystems need them to stay healthy. These good fires
These are the following future scopes of forest fire
burn low to the ground and usually do not kill big trees. Forest
prediction system –
managers sometimes use fire on purpose to keep some
• The project gives emphasis on using modern
ecosystems healthy. These ecosystems are called fire-
technology of remote sensing and other techniques to
dependent ecosystems, because they need fire to survive [13].
equip the forest authority and organizations in their
work for forest conservation. Due to the lack of ACKNOWLEDGMENT
efficient staff and emergency plan there is a greater
need to use much efficient systems. It is difficult for We would like to acknowledge the many individuals and
authorities to manage huge forest areas and to be organizations who supported and contributed to this paper. We
present at a time of accidents. are very thankful to all those who played a role in the
achievement of this paper.
• Due to human limitation and lack of modern
Above all, we are grateful to our mentor, teacher and
technology, it is not possible to combat forest fire. The
advisor, (Prof.) Dr. Suparna Biswas for her idea, guidance,
installation of this project in a better way can give a
support, and encouragement throughout the whole process. Her
chance to organizations to have efficient control over
mentorship and expertise helped us to refine our research
forest by not interrupting its fragile ecosystem [9].
direction and to realize our ideas.
We also want to express our sincere appreciation to all of the
• The use of this forest surveillance system will be members in our study, who generously shared their time,
helpful in mapping emergency plan and making it experiences, and insights with us. Firstly, Saptarshi Pramanik
more effective during the occurrence of any disaster. for assisting us with his vast knowledge in the field of embedded
systems and the methodologies, techniques and working of
• This system uses advance technology which will help prediction systems and circuit systems used in. Secondly,
in tracing out the forest fire in its initial stage. Sambit Chanda for helping us with the coding and algorithm of
this paper and linking the results to the website and Shibam
• The poaching activities and illegal trade of forest Sanyal for editing the paper.
goods can be stopped by the application of these
sensors. Hence, with the complete information of This paper would not have been possible without the support
incident and its nature, the administration can take the and contributions of so many people. Their readiness to engage
appropriate decisions to combat illicit trading of trees with our work was crucial for the success of this paper, and we
and preventing forest fires [10]. are deeply indebted to them for their participation.
REFERENCES
XVI. CONCLUSION [1] Dhiman, A., Shah, N., Adhikari, P., Kumbhar, S., Dhanjal, I.S. and
Mehendale, N., 2021. Firefighting robot with deep learning and machine
This paper gives a brief overview of a forest fire vision. Neural Computing and Applications, 34, pp.2831-2839.J. Clerk
application that uses a well-known routing algorithm called Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., vol. 2.
Oxford: Clarendon, 1892, pp.68–73.
SPIN. The SPIN protocol in Forest Fire Detection application
[2] https://link.springer.com/article/10.1007/s10694-020-01056-zK.
was only tested in a basic way because of time and resource
[3] https://www.ijeea.in/wp-content/uploads/Volume-9-Issue-1-Paper-4.pdf
limitations. More work could be done to solve the problem with
[4] IOT INTEGRATED FOREST FIRE DETECTION AND PREDICTION
the time synchronization module that would allow the simulator USING NODE MCU by M. Khadar, V. Ranjith, and K. Varalakshmi
to measure the network latency. More detailed simulator runs published in International Journal of Electronics Engineering and
would also help to evaluate the protocol better. It is also Applications in 20212.
important to compare this application with other applications [5] IOT BASED FOREST FIRE PREDICTION SYSTEM by S. S. Patil, A.S.
that use different protocols. This comparison and Patil, and A. S. Patil published in Journal of Emerging Technologies and
Innovative Research in 20205.
implementation of other applications could be done in the
[6] https://www.wikipedia.com
[7] Forest fire and smoke detection using deep learning-based learning IEEE conference templates contain guidance text for
without forgetting | Fire Ecology | Full Text by Veerappampalayam composing and formatting conference papers. Please
Easwaramoorthy Sathishkumar et al. ensure that all template text is removed from your
[8] A smart approach for fire prediction under uncertain conditions using conference paper prior to submission to the conference.
fuzzy logic and machine learning techniques | SpringerLink by R. S. S. Failure to remove template text from your paper may
Prakash et al. result in your paper not being published.
[9] A Design and Development of the Smart Forest Alert Monitoring System
Using IoT by M. Krishnamoorthy et al. published in Journal of Sensors in
2023.
[10] Forest Fire Detection and Prevention System Using IoT by M. Gautam
Singh et al. published in Proceedings of 3rd International Conference on
Machine Learning, Advances in Computing, Renewable Energy and
Communication in 2022.
[11] A multimodal framework for Forest fire detection and monitoring –
Springer.
[12] https://defenders.org/blog/2020/07/fire-adapted-plants-and-animals-rely-wil
dfires-resilient-ecosystems
[13] https://static.colostate.edu/client-files/csfs/pdfs/ColoradoFireEcologyOverv
iew101712.pdf