| Forest fire is one of the key factors threatening forest environment and public safety,which is catastrophic to forestry development,industrial production and people’s life.As a new type of aerial monitoring platform,unmanned aerial vehicle(UAV)has the advantages of rapid deployment,flexibility and practicability.It is a fast,accurate and efficient monitoring method on forest fire by using UAVs.The key technology of monitoring and detecting of the fire monitoring of UAV is the stabilization and recognition,the electronic image stabilization and the fire detection and identification system of surveillance video are of great significance.In this paper,unmanned helicopter is studied as the platform to practice the monitoring of forest fires,equipped with a binocular camera combining a visible and an infrared camera.The stabilization and recognition system of UAV fire monitoring had been designed and realized.The main content of this paper includes the following parts:1)By analyzing the cause of jitter in aerial image,an electronic image stabilization algorithm based on optical flow method is proposed.The algorithm uses L-K optical flow method to estimate the local motion of the corners of the image.Then the least squares is applied on global estimation.Moving filtering is conducted by mean filtering and Kalman filtering respectively,and then a compensation of the image is given based on the global motion vector.The results show that the algorithm can reduce the influence of jitter in the image and improve the quality of the video image.2)The characteristics of forest fire infrared image are analyzed.Circularity,area change rate and texture feature are selected and extracted respectively as the flame features.A fire image recognition algorithm based on support vector machine(SVM)is designed and the classifier model is obtained by training samples.In the actual fire detection test,the correct recognition rate of fire is 84.77%,proving that the algorithm is effective.3)A UAV Forest Fire Monitoring System is designed and implemented.In terms of hardware,the binocular camera is made,and the CM-T3730 is used as the airborne embedded core and air-ground communication interaction system.In terms of software,the airborne image algorithm and the ground control station are designed and implemented.4)On-line debugging and open air flight test are carried out based on this system.The online debugging is mainly to test the integrity of the whole system on the ground.The debugging results show that the system meets the functional requirements of this paper.The results of flight show that the monitoring video image is clear and coherent by using the electronic image stabilization algorithm,and the image quality has improved.Fire identification algorithm can accurately detect the fire and provide ground staffs the fire position. |
