| The spread of forest fire has caused huge losses to the property and personal safety all over the world.The forest surface fire is the spread with the strongest damage and the highest risk factor in the process of forest fire spreading.The monitoring,spread and prediction model establishment of forest surface fire is an important method to solve the spread of forest fire.There is a backwardness in monitoring methods for forest surface fire prevention and early warning,and there is no positioning capability on the forest surface fire spread profile in the airborne monitoring system,also,there are some problems,such as inaccurate simulation accuracy of forest surface fire spread model,large amount of calculation and low prediction accuracy of forest surface fire prediction model,which leads to the failure of relatively complete research on forest surface fire monitoring,spread and prediction at the present stage,so it is difficult to popularize and apply.In view of the above development status and problems,in this paper,an airborne monitoring system for forest surface fire based on multi-sensor data fusion formed by UAV carrying lidar,infrared camera and integrated inertial navigation.A multidirectional cellular automata(MCA)spread model of forest surface fire is designed,and a multidirectional cellular automata-ensemble Kalman filter(MCA-ENKF)forest surface fire prediction model based on spread model and data assimilation is designed.From the aspects of technologys and models,this paper breaks through the difficult problems in the field of traditional spread datas monitoring of forest surface fire and the model of spread prediction.The airborne monitoring system of forest surface fire spread based on multi-sensor data fusion is designed.According to the characteristics of forest surface fire spread,the types of multi-sensor are determined,the calibration between multi-sensor is carried out,the rotation transformation matrix between multi-sensors is obtained,the communication mode between airborne monitoring system and ground receiver is constructed.The multi-frame point cloud splicing and accurate location of forest surface fire points are realized by multi-sensor data fusion.The test verifies that the multi-sensor airborne monitoring system has the ability to collect the datas of forest surface fire spread.The model of forest surface fire spread based on multidirectional cellular automata is proposed.To analysis the key factors affecting the spread of forest surface fire,the indoor ignition tests are carried out to analyze the effects of different environmental wind speed,environmental slope,fuel moisture content,fuel load and other main factors on the spread of forest surface fire.The simplified Rothermel surface fire velocity model is obtained.By combining the simplified Rothermel forest surface fire velocity model with cellular automata,a multidirectional cellular automata of forest surface fire spread model is established.The accuracy of the spread model is verified by tests.The forest surface fire prediction model based on ensemble Kalman filter algorithm and multidirectional cellular automata is proposed.Combined with the simulation data of forest surface fire spread model,forest surface fire spread observation data and data assimilation algorithm,the forest surface fire prediction model based on ensemble Kalman filter cellular automata(MCAENKF)is established.The best parameters such as the set number and data assimilation frequency of the forest surface fire prediction model are determined.The tests are carried out to verify the accuracy of the model in different environmental slope,environmental wind speed,fuel moisture content and fuel load.The ignition test of field forest surface fire is designed.It is verified that the airborne monitoring system with multi-sensor data fusion is high monitoring capability at different flight altitude,flight direction and flight speed.Under the influence of factors such as slope of different combustion environment,wind speed of combustion environment,combustible moisture content,fuel load and other factors,the reliability of forest surface fire spread model and prediction model are verified.The establishment and experimental application of forest surface fire spread and prediction model of airborne monitoring system based on multi-sensor data fusion are completed. |
