Insulator identification method based on infrared image

UAV aerial has been used more and more widely in transmission line inspection. However, the automatic recognition of the visible image still can't meet the actual requirements, fault diagnosis is still largely done by hand, and infrared images contain fewer environmental factors in comparison and can basically reflect a variety of insulator faults, so it is more practical than visible image in the identification and fault diagnosis of insulator. In this paper, a method of insulator identification based on infrared image is proposed; in turn extracting surf feature points of image, K nearest neighbor matching, adaptive spatial clustering and segmentation correction are performed, then the axis and axis temperature of the insulator is extracted. The method gives full play to the advantages of infrared image and can accurately identify the insulator; it lays the foundation for insulator fault diagnosis based on infrared image.