Forest Fire Detection Using nRF24L01 Wireless Sensor Network And Prediction by Machine Learning Model

Forest fire is a major problem in recent years which creates natural disasters. An effective method for the detection and prediction of forest fires reduces the devastating effect. An effective method to predict and detect forest fires in real time is implemented in the paper. Real-time forest fire detection is done by considering various parameters like temperature, humidity, and smoke intensity with the help of nRF24L01 based wireless sensor network established using tree topology. Prediction of forest fire is done by integrating neural prophet and logistic regression machine learning models by considering key environmental parameters. The prediction accuracy of 86.67% was obtained by considering the parameters like earth skin temperature, wind speed, relative humidity, and precipitation. The prediction model is able to predict fire occurrence efficiently from 2013 to 2015 which has been validated with the available dataset.