Image Segmentation Methods for Kamikaze FPV Drones Targeting to Aid Critical Energy National Infrastructure Assets Protection
摘要
The real state of affairs regarding risks has been considered damage to assets of the national energy infrastructure and options for their protection. Methods of segmentation of images obtained from kamikaze FPV drone were considered. Their advantages and disadvantages are listed. The paper proposes the development of an improved image segmentation method for kamikaze FPV drones targeting to aid critical energy national infrastructure assets protection. Experimental studies were conducted on the segmentation of a real image from kamikaze FPV drone by the Canny edge detector and Sobel operator. The characteristics of images obtained from the drone and the main disadvantages of the considered segmentation methods were determined. Such data were taken into account when developing the improved image segmentation method for kamikaze FPV drones. Visual analysis result of work of the proposed method showed that it allows to detect an aerial object on a segmented image.