Download PDFOpen PDF in browserForest Fire Area Estimate of an Ongoing FireEasyChair Preprint 76054 pages•Date: March 23, 2022AbstractThe most common hazard in forests is wildfire.These fires present a challenge to the invaluable forest resources and the natural ecosystem which includes plants and animals. It drastically affects the biodiversity and ecology of a region.Wildfires are a bigger problem and they occur anywhere ir-respective of the topology. A huge number of fire points were detected in India in the last three years. The increase in the rate of fires is approximately doubling every three years. There are multiple causes of forest fires including natural and man-made. Natural ones include soil erosion, change in climate, global warming, low humidity, high atmospheric pressures, dryness, etc.Taking into account the statistics from last year, India witnessed a 43 percent increase in the number of wildfires in the last decade(2009-2021), witnessing a 125% spike in such fires in a span of just 2 years from 2015 to 2017. Taking into consideration this serious issue we have developed a model which will help us predict the forest fires area when the forest fire is at an early stage of ignition and take appropriate measures. It provides two types of inputs, one via form and the other through SMS. Further, satellite data is integrated and the prediction is done. Machine Learning Algorithms along with GIS data is used to perform predictions. The results of the prediction are displayed on themap at respective coordinates. Once the spread is determined the nearby areas can be evacuated and the lives of many can be saved. Also, appropriate measures can be taken to control the spread of the fire. Keyphrases: Forest Fire Prediction, Gaussian Naive Bayes, K-Nearest Neighbours, Random Forest, decision trees, machine learning
|