A Method of Transmission Conductor-Loosened Detect Based on Image Sensors

Abstract

The conductor-loosened fault of the transmission line caused by its own mechanical load, breeze vibration, and galloping will gradually worsen by the passage of time and even lead to a series of irreversible accidents such as conductor breakage and tower fall. In order to detect the loosened conductors as soon as possible and minimize the possible damages, a detection method for the loosened conductor based on the vision sensor is proposed. First, the conductor images are captured by the image acquisition system, using the inspection robot, and then, the energy gradient algorithm with guided filtering (GEG) is performed to highlight the foreground and strengthen the hierarchy, so that the conductor area can be obtained by the Otsu threshold segmentation. Second, the surface contour of the conductor is extracted, filtered, and numbered. Finally, according to the arrangement characteristics of the outermost Al-strands of the conductor, we established a conductor-loosened model that includes the analysis of the four feature quantities of contour length D, direction angle θ, curvature W, and spacing T and the thinning process based on the step-vertical limit (SVL) to judge the conductor fault and locate the loosened strands. We analyze the performance of the technology by tests, and the results demonstrate that the proposed method can quickly detect the conductor-loosened fault and achieve good performance.