The Research Of The Forest Fire Danger Warning Based On The Internet Of Things

Along with the development of Social productivity, as well as computer hardware andsoftware technology with each passing day, Internet of Things as emerging item informationnetwork is advancing rapidly, it is called the world’s third information wave which followingthe computer and the Internet. The development of the Internet of Things will bring hugeeconomic benefits to the community, it will greatly improve the social productivity, and alsofacilitate our daily life in a great extent. The Internet of things bring us such a greatopportunity for development, but at the same time, there are many constraints of industrialdevelopment and unprecedented technical challenge.The paper based on this background, the forest fire danger warning program is designed,the sensing layer and the application layer data are processed separately in different dataprocessing algorithms, identify the critical weather conditions of the forest fire. The studycontent of this paper is as follows:(1)Study the development background of the Internet of Things, researched the relatedconcept of the Internet of Things from the technical level, based on this, analysis theapplication prospects of the Internet of Things.(2)The forest fire danger warning system model is designed based on the Internet ofThings, and for the different geographic or monitoring area that is too large, put forward theidea that divided the different monitoring areas. At the same time, in the program running, atany time to adjust the length of the forecast period according to the actual situation, real-timecorrection to the prediction model of the application layer. In addition to, from the flow levelof the data processing, put forward utilized the focal underdetermined system solver toeliminate the environmental interference error of the sensing layer, cleared the strategy of thevalue of the meteorological elements factor in a forecast period, and use the Neural Networksas the modeling approach of the application layer to identify the critical weather conditions ofthe forest fires.(3)Study the radial basis function neural network which has been popular, in combinedwith the relevant characteristics of the radial basis function neural network, make acorrelation analysis of the radial basis function neural network is applied to the forest firedanger warning. As for the unique advantages of the radial basis function neural network, tolay a good foundation that application the radial basis function neural network to the forestfire danger warning based on the Internet of Things, however, several mature learningalgorithm which there are several disadvantages be further improved so that it applies to theprogram, Such as prediction accuracy and correction capability of the network have a significant inadequate. For this reason, the paper summarizes the results of previous studies,analyzed the several algorithms which have been popular, proposed a two hidden layer radialbasis function neural network model, at the same time use the genetic algorithms as weightstraining methods, and combined with dynamically update the network topological structurewhich predecessors have proposed in the training process, so that the algorithm has a lot ofadvantages compared to the traditional algorithm. Through the experimental simulation of thealgorithm is applied to the non-linear function approximation and the classification of thecommunication enterprise’s customers, moreover, compared with similar RBF neural networkalgorithm, reduces the complexity of time of the two hidden layer radial basis function neuralnetwork, as well as the complexity of computational. this laid a solid foundation for the radialbasis function neural network used in the special scene of the real-time control in the Internetof Things.(4)To make a simulation from the data processing flow level. Firstly, constructing a dataset adapt to the paper based on the traditional forest fire data set. Secondly, experimentalsimulate to the data processing algorithm from the sensing layer and the application layer.Finally, statistical analysis the experimental results, proved the effectiveness and efficiency ofthe program from different perspectives, the program have certain values.