A Method for Forest Fire Monitoring Based on GA-BP

Abstract

Forest fires pose a serious threat to power systems, and cause a large number of casualties and property damage. How to accurately monitor the occurrence of forest fires has become a major problem. Although the BP neural network-based mountain fire monitoring can monitor the occurrence of mountain fires and determine the stage of mountain fires accurately, the BP neural network also tends to fall into the problems of local optimum, slow convergence and even non-convergence. In this paper, we propose a method for monitoring hill fires based on genetic algorithm optimized BP neural network, and also introduce the basic principle of genetic algorithm optimized BP neural network, and also use BP neural network and genetic algorithm optimized BP neural network to determine whether hill fires occur and the stage of hill fires based on the data of CO, smoke concentration and temperature generated by forest fires The advantages and disadvantages of the two algorithms for hill fire monitoring are also compared and analyzed.