Detection and Condition Assessment of Icicle Bridging for Suspension Glass Insulator by Image Analysis

Ye Zhang,Xinbo Huang,Jianyuan Jia,Yongcan Zhu,Long Zhao,Xiaoling Zhang

DOI: https://doi.org/10.1109/tim.2020.2984965

IF: 5.6

2020-01-01

IEEE Transactions on Instrumentation and Measurement

Abstract:Icicles generated on the insulator sheds will constantly shorten the insulating gap distance between adjacent sheds as growth, thereby inducing significant decrease in electrical insulation of the insulators, icicle bridging even threatening the operating security of power systems. In order to evaluate whether the ice-covered insulator needs to deicing or the icicles bring in potential dangers, a study on detection and condition assessment of icicle bridging for ice-covered suspension glass insulator by image analysis is proposed in this article. First, the initial icicles regions are obtained by the local variation threshold segmentation algorithm based on the optimized texture energy image. Second, the insulator sheds are segmented and extracted based on the normalized color difference operator, and then the complete icicles regions are obtained through the image difference and postprocessing. Finally, the icicle length is calculated by finding its smallest axis-parallel enclosing rectangle, and two adjacent insulator sheds spacing which is calculated by ellipse fitting is used as bridging threshold, so the condition of icicle bridging can be defined as the safe, observation, and danger. The results demonstrate that the icicles can be effectively enhanced and accurately segmented by the presented method in this article, and the icicle bridging condition of ice-covered insulator can be assessed with average detection accuracy of 92.2% and average time consumption of 3.910 s, by which the safe operation of ice-covered insulator can be monitored and guaranteed.