Intelligent Forest Fire Prediction System

Categories: Catching FireFire

One of the main causes of destruction of archaeological and cultural heritage sites, is wildfires. The existing systems proved to be less intelligent in the case of early warning to a fire break out which is the only way to avoid human losses. The system integrates various sensors including passive infrared (PIR) sensors, a wireless sensor network of temperature and humidity sensors. The signals and measurements collected from these sensors are transmitted to the control centre, which employs intelligent computer vision and data fusion techniques to automatically analyze sensor information with the help of prediction algorithms.

When the amount of data reaches critical volume, modern software techniques are implemented in order to accomplish system goals.


The most common hazard in forests is forests fire. They pose a threat not only to the forest wealth but also to the entire regime to fauna and flora seriously disturbing the biodiversity and the ecology and environment of a region.Wildfires also cause a great loss to human life when a group of people are out there for trekking.

Get quality help now
Writer Lyla
Verified writer

Proficient in: Catching Fire

5 (876)

“ Have been using her for a while and please believe when I tell you, she never fail. Thanks Writer Lyla you are indeed awesome ”

+84 relevant experts are online
Hire writer

Trekkers are sometimes trapped in massive forest fires as they will not be aware of the wildfires before they start.

Wildfires are generally overseen in the following ways:

  • prevention of wildfire
  •  wildfire recognition
  • devising a best recovering approach

Wireless Sensor Networks is a deployment of several devices equipped with sensors that perform a collaborative measurement.Any kind of measurement can be used for a specific purpose.The working process of a sensor network involves three parts .

Get to Know The Price Estimate For Your Paper
Number of pages
Email Invalid email

By clicking “Check Writers’ Offers”, you agree to our terms of service and privacy policy. We’ll occasionally send you promo and account related email

"You must agree to out terms of services and privacy policy"
Write my paper

You won’t be charged yet!

Sending a real time value measured by the sensor followed by the communication protocol that is selected that passes on the gathered information to the external system [1].

Alarm systems could be of less accurate due to the above mentioned complications [2].The fundamental application of Wireless Sensor Networks was military.In future days the lifespan of the batteries were improved which resulted in an increased efficiency of the nodes.WSN is applied in industries like healthcare, mobile phones, cars, household applications.Event detection is a major issue for applications of wireless sensor networks (WSNs). In order to detect an event, a sensor network has to reliably identify which deployment-specific incident has occurred based on the raw data gathered by the sensors on the individual sensor nodes.A detected event is monitored with the help of WSN which helps the government to know the extent of damage to the biomass[3].

Forest fires are caused due to several reasons such as lightning,volcanic eruptions and some human activities. Fuel,oxygen and a source of heat ,these three are the prerequisites for fire and are commonly known as the fire triangle.The availability of these three elements can unleash and intense fire in the forest which is usually termed wildfire or bush fire[4].This method used a fuzzy logic technique to predict the occurrence of wildfire which was found to have a lack of real time response and less in speed [5].

Shortcomings in detection has lead to set up a system that helps to predict the occurrence of wildfires in advance.This prediction involves modern software techniques such as Artificial Intelligence.AI is a machine with the ability to solve problems that are usually done by humans with natural intelligence.This system employs Support Vector Machines which are classifiers that has a hyperplane to separate which fails to perform in case of handling a large amount of data [6].

Regression is one of the Machine Learning algorithms which originated from statistical modelling. Regression outputs a response that is ordered and continuous .This system uses Linear Regression technique which creates a trendline that fits the data points from the dataset.But it’s hard to know how confident you are in a forecast.It seems to be a really challenging to estimate confidence[7].

When the moisture content in the trees falls off the wood starts releasing smoke due to cell breakdown.Moisture content present in the soil and the temperature are taken into account for predicting the wildfires which seems to lack in the amount of data they collect from the forest field[8].Distributed Information system provides a repository of data to the developed models .This system requires a communication network that connects many computers.Since it is found to have a greater potential of bugs it is a risk of employing such models in forest fields[9].

Level of accuracy can be increased using Machine learning technology since it is found to have a good correlation between the input and output values. Its related to statistical analysis and data mining. Basic phases to be considered in machine learning are

  1. Draw data
  2. Process Data
  3. Model training
  4. Test data
  5. Improve

The success of the project depends upon the chosen data. Choosing a set of data that does not fit the statement lads to a failure.The safest way of choosing the data is to make a trial and error method ,so that the data set best fits the statement. Next in order comes the processing of data which involves the manipulation of data for the upcoming analysis and training.

Subsequently a model that suits the requirements is picked,models ability is improved incrementally to predict the outcomes.Real challenge comes at this phase where n number of real time instances are fed to model for examining.The exactness of the system could be improved in distinct paths. Expanding data,considering missing and outlier data, hypothesis generation ,practicing multiple algorithms are a few approaches to refine the model[10].

This paper deals with the fastest prediction algorithm, helps to forewarn trekkers so as to save people and the natural vegetation.This system is designed as a decision support system to provide an alert on forest fires to the trekkers that would act as a great life saver.


The proposed system is developed to predict the occurrence of wildfires. This system is an analogy to the day to day prediction of weather.. In order to perform prediction, it is required to specify the real time parameters monitored during the day that are done with the help of data collected from wireless network of various sensors.


It’s a probabilistic classifier used to build models fast and make quick predictions. This algorithm processes a

huge amount of data from the sensors, calculates the probability of the event to occur with maximum accuracy.


  1. Selection of inputs/parameters
  2. Collecting datasets
  3. Training system
  4. Fetching real time values
  5. Probability calculations
  6. Predicted outcomes

The system takes the following inputs:

  • Fine Fuel Moisture Code(FFMC)
  • Duff Moisture Code(DMC)
  • Drought Code(DC)
  • Initial Spread Index(ISI)
  • Temperature
  • Relative Humidity(RH)
  • Wind
  • Rain

The dataset containing the data such as Fine Fuel Moisture Code(FFMC), Duff Moisture Code(DMC), Drought Code(DC), Initial Spread index(ISI), Temperature, Relative Humidity(RH), Wind and Rain at which the forest fire has occurred is collected from the Tamil Nadu Forest Department.The sample dataset at which the forest fire has occurred is provided below.

The system is then trained using the collected data set for the prediction of wildfires. As these systems learn from this data, they are able to handle increasingly complex tasks.

The process of training the system involves providing an machine learning algorithm with training data to learn from. The training data must contain the correct answer, which is known as a target or target attribute. The learning algorithm finds patterns in the training data that map the input data attributes to the target (the answer that have to be predicted) and it outputs an machine learning model that captures these patterns.

in which naive bayes classifier algorithm is used in this system.The Naive Bayes Classifier technique is based on the so-called Bayesian theorem and is particularly suited when the dimensionality of the inputs is high. Despite its simplicity, Naive Bayes can often outperform more sophisticated classification methods.The Bayesian Classification represents a supervised learning method and a statistical method for classification. Assuming an underlying probabilistic model,it allows us to capture uncertainty about the model in a principled way by determining probabilities of the outcomes. It can solve diagnostic as well as predictive problems.

The representation for Naive Bayes is probabilities.A list of probabilities are stored to file for a learned naive Bayes model. This includes:

  • Class Probabilities: The probabilities of each class in the training dataset.
  • Conditional Probabilities: The conditional probabilities of each input value given each class value.

Naive Bayes classifiers can handle an arbitrary number of independent variables whether continuous or categorical. Given a set of variables, X = {x1,x2,x…,xd}, we want to construct the posterior probability for the event Cj among a set of possible outcomes C = {c1,c2,c…,cd}. In a more familiar language, X is the predictors or hypothesis and C is the set of categorical levels present in the dependent variable. Using Bayes’ rule:

  • P(X / C) – Posterior Probability which represents the degree to which we believe a given model accurately describes a situation given the available data and all of our prior information.
  • P(C/ X) – Likelihood that describes how well the model predicts the data.
  • P(X) – Prior Probability which describes the degree to which we believe the model accurately describes reality based on our prior information.
  • P(C) – Normalizing Constant which makes the posterior density integrate to one.

After calculating the posterior probability for a number of different hypotheses, you can select the hypothesis with the highest probability. This is the maximum probable hypothesis and may formally be called the Maximum a Posteriori (MAP) hypothesis.This can be written as:


The fire prediction is the best way to fight against the wildfires and also to alert the trekkers who decide to go out for a trek.The existing system proved to be less intelligent in the case of early warning to a fire breakout which is the only way to avoid human losses.The proposed system is designed as a decision support system in which the input data from wireless network of various sensors is compared with the trained system.As a result, the system proves to be more intelligent to predict the wildfire breakout in prior and gives an early warning which avoid 100% human loss.


  1. Junguo ZHANG, Wenbin LI, Zhongxing YIN, Shengbo LIU, Xiaolin GUO, “Forest Fire Detection System based on Wireless Sensor Network”,ICIEA 2009,
  2. Li Guang-Hui ;Zhao Jun ;Wang Zhi ,”Research on Forest Fire Detection Based on Wireless Sensor Network”,Published in: 2006 6th World Congress on Intelligent Control and Automation
  3. Xiaoyang Zhang ; Shobha Kondragunta ; Brad Quayle,”Estimation of Biomass Burned Areas Using Multiple-Satellite-Observed Active Fires”,Published in: IEEE Transactions on Geoscience and Remote Sensing ( Volume: 49 , Issue: 11 , Nov. 2011)
  4. Sergio Trilles ; Pablo Juan ; Laura Diaz ; Pau Arago ; Joaquín Huerta,”Integration of Environmental Models in Spatial Data Infrastructures: A Use Case in Wildfire Risk Prediction”,Published in: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Volume: 6 , Issue: 1 , Feb. 2013 )
  5.  Bruna E.Z.Leal;Andre R. Hirakawa ; Thiago D. Pereira,”Onboard fuzzy logic approach to active fire detection in Brazilian amazon forest”,Published in: IEEE Transactions on Aerospace and Electronic Systems ( Volume: 52 , Issue: 2 , April 2016 )
  6. George E. Sakr, Imad H. Elhajj, George Mitri and Uchechukwu C. Wejinya,”Artificial Intelligence for Forest Fire Prediction”, 2010 IEEE/ASME International Conference on Advanced Intelligent Mechatronics Montréal, Canada, July 6-9, 2010
  7. Kansal, Yashwant Singh , Nagesh Kumar , Vandana Mohindru,”Detection of Forest Fires using Machine Learning Technique: A Perspective “,Published in:2015 Third International Conference on Image Information Processing”.
  8. David Chaparro ; Mercè Vall-llossera ; Maria Piles ; Adriano Camps ; Christoph Rüdiger ;Ramon Riera-Tatché,“Predicting the Extent of Wildfires Using Remotely Sensed Soil Moisture and Temperature Trends”,Published in: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Volume: 9 , Issue: 6 , June 2016 ).
  9. Sergio Trilles ; Pablo Juan ; Laura Diaz ; Pau Arago ; Joaquín Huerta,”Integration of Environmental Models in Spatial Data Infrastructures: A Use Case in Wildfire Risk Prediction, Published in: IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (Volume: 6 , Issue: 1 , Feb. 2013 ).
  10.  Divya T.L. ; Vijayalakshmi M.N.,”Analysis of wild fire behaviour in wild conservation area using image data mining”,Published in: 2015 IEEE International Conference on Electrical, Computer and Communication Technologies (ICECCT)

Cite this page

Intelligent Forest Fire Prediction System. (2021, Dec 28). Retrieved from

👋 Hi! I’m your smart assistant Amy!

Don’t know where to start? Type your requirements and I’ll connect you to an academic expert within 3 minutes.

get help with your assignment