Design And Implementation Of Forest Fire Monitoring System Based On Drone

Forest fire is uncontrollable and can spread freely.If a forest fire occurs,it will threaten and damage the forest itself,the forest ecosystem and humans,it is extremely important to do forest fire prevention work.Because the forest is very vast,the traditional forest inspection methods have the disadvantages of monitoring blind spots,poor real-time performance,and high resource consumption.Therefore,according to the maneuverability of drone,the use of drone is not restricted by geographical and environmental conditions,and the drone is suitable for work in a complex environment.This thesis designed and developed a forest fire monitoring based on drones.And the main work of the thesis is as follows:1.Design of the drone system.According to the actual needs of forest fire monitoring,this thesis completed the selection of the drone and cradle head camera and the design of image acquisition and transmission.2.The design of the intelligent landing station.In order to improve the automation level of drone use and solve the shortage of flying hands,a drone intelligent landing station was designed.When the power of drone was too low,the battery was automatically replaced after landing in the landing station to achieve independent battery life.3.Research of forest fire monitoring algorithm based on drone image.First,the image preprocessing technique was performed on video frames.Then according to the color characteristics of the flame,this thesis proposed a method based on FDI(Fire Detection Index)and R channel to segment the suspected forest fire region and extracted the circularity,are change rate,center-of-gravity height ratio,and LBP texture features of the suspected forest fire area.Finally,the SVM classifier was used to determine whether there was a fire.4.Design of the ground terminal monitoring system.Among them,one was the design of the landing station software based on Labview and the design of the PC interface,which could realize the control of the landing station,parameter display and battery management;the other was the design of forest fire monitoring algorithm based on C++ and the design of the PC interface.The PC interface of forest fire monitoring provided basic functions such as user registration,login,query historical anomaly events,and taking pictures.The user could browse the forest screen through the forest fire monitoring interface,which provided real-time fire dynamic information when the fire occurred and provided relevant decisions of rescue work.The forest fire monitoring system based on the drone used the forest fire monitoring algorithm to process and analyze forest images returned by the drone,and could monitor the forest fire in real time regardless of day and night.It improved the ability of forest fire safety and the automation level of forest fire warning.Compared with the existing algorithms,the proposed algorithm has high accuracy and low false positive rate.After many tests and analysis,the system can run stably for a long time,and has strong expandability.It can be applied to different occasions by carrying different sensors and different algorithms,such as land confirmation,power inspection,vegetation pests,etc.And it has a good application prospect.