Research and Application of Insulator Detection Method Based on Information Fusion of Inspection Characteristic Quantity

Abstract

The insulator image recognition is an important prerequisite to achieve its operational state monitoring and fault diagnosis. Based on the complex features of the aerial image background, this paper proposes a kind of insulator detection method based on the information fusion of image characteristic quantity. The detection method improves the traditional Canny edge detection method and Hough transformation image recognition algorithm in order to give a digitized description of its characteristics and achieve automatic recognition of the insulator image. The experimental results show that, the proposed segmentation method can distinguish the target and background, with complete and continuous edge, less voids in the target, high positioning accuracy and rapid speed. This method can lay a foundation for the insulator state monitoring and fault diagnosis without human intervention.