A Transmission Line Icing Detection Method based on Improved Canny and Hough Transform

Abstract

To address the issues of noise interference in edge detection of transmission line icing images and inaccurate thickness calculation due to scattered straight line detection, a detection method based on an improved Canny operator and optimized Hough transform is proposed. The Canny adaptive double threshold is achieved through the percentile sorting of gradient magnitude to enhance the robustness of edge extraction; the optimized Hough transform with line fitting is adopted to fuse scattered edges into complete icing edge lines ; and the thickness conversion is completed in combination with the camera imaging principle. Experiments show that this method can effectively improve the accuracy of icing thickness detection and has engineering application value.