Research On Forest Fire Identification And Location Algorithm Based On UAV

At present,the rapid development of drone technology has provided new research and development directions for various fields,and drones have gradually emerged in the field of forest fire prevention.In the application of forest fire prevention monitoring,video network monitoring has the defects of increased cost,difficult infrastructure maintenance and fixed location.Therefore,this paper introduces the UAV technology,based on the identification and positioning of forest fires by drones,to achieve early detection and quasi-positioning,and to improve the effectiveness of forest fire prevention decisions.Mainly done three aspects of research.The method of forest fireworks identification based on drone is studied and an improved algorithm is proposed.The improved algorithm based on color index has good processing effect,which can meet the video image data returned by UAV during real-time processing in real time,and combined with the method of forest fire calibration to remove suspected fire,which can be accurately determined.The occurrence of forest fires.The forest fire location algorithm based on drone is studied..According to the geometric position relationship between the UAV and the target area and the geometric relationship between the UAV shooting point and the ignition point,the coordinates of the image and the ignition point are calculated by the coordinates of the UAV shooting point.The position coordinates,and the calculated image vertex coordinates and track coordinates can be displayed in the 3D image scene.The experimental verification of forest fire identification and forest fire location algorithm was carried out.By comparing the processing results,processing speed and recognition accuracy of the general algorithm and the improved algorithm,it is found that the improved algorithm is better than other algorithms,and then based on the recorded shooting position information,the four vertex coordinates and the ignition point of the corresponding time image are calculated.Coordinates,and displayed in the 3D image scene,mark the location of the forest fire occurrence area.The results show that the combination of improved color index-based recognition algorithm and localization algorithm can achieve the location of forest fire.By comparing the results of inter-frame difference method,background subtraction method and improved algorithm processing image,combined with the results of forest fire identification and forest fire calibration method,it is proved that the research method can meet the forest fire identification requirements under the general scene and canaccurately judge The forest fire occurs in the area,and then the positioning operation is performed on the processed image,and the function of the ARCGIS component in the platform is called,thereby finally realizing the forest fire identification and positioning algorithm function of the drone video.