| The occurrence of forest fires is sudden and contingent.Once forest fires occurs,it spreads rapidly with the increase of fire intensity,threatening the living environment and life safety of nearby residents,and also bringing irreversible damage to the earth's ecological environment.Therefore,the fast and effective forest fire identification technology is of great significance to the suppression of forest fires.At present,the commonly used forest fire identification technologies include temperature sensing identification technology and smoke sensing identification technology,but with the continuous updating and development of technology,people have higher requirements for forest fire identification accuracy.In recent years,the rise of Deep Learning technology makes more and more experts and scholars begin to try to apply Deep Learning technology to the field of forest fire prevention,among which forest fire recognition technology based on Convolution Neural Network has made great progress.The main work carried out in this paper includes:(1)The study of Convolution Neural Network algorithms.Introduced and analyzed the stru cture and principle of convolution neural network and its common framework.(2)This paper studied the method of forest fire recognition based on convolution neural network,improves the existing methods of forest fire recognition,and carries out comparative experiments for different factors,and obtains the best method.(3)Design and implement of the forest fire prevention video monitoring system,which completes the functions of video fire monitoring,platform control,video recording,image capture,forest fire identification,fire information alarm and so on. |
PDF Full Text Request