Forest Fire Spread Prediction Method based on BP Neural Network

This paper proposes a method suitable for edge computing to use neural networks to predict the spread of forest fires, aiming to improve the accuracy and efficiency of fire spread prediction, and to achieve edge computing on the drone side with low energy consumption requirements. The BP neural network model is trained by the simulated fire spread raster data obtained by FlamMap, the direction and speed of fire spread are predicted respectively, and according to the Huygens' principle, the vector fire line is obtained by the prediction data fit, and the fire line obtained by FlamMap is compared to verify the accuracy of the method. It can be considered that the same effect can be trained and calculated for the spread of fire through remote sensing data of real fires.