Faulty Insulator Diagnosis for UAV Videos Based on Repetitiveness Feature

Abstract

In this paper, we proposed a faulty insulator diagnosis method for insulator faulty detection based on repetitiveness feature from UAV video sequence. The repetitiveness feature which describes the relationship of each structural element of insulator is employed for the insulator faulty diagnosis, and is robust to noise, camera motivation and complex backgrounds. Base on the repetition of structural element, our method can not only deal with single insulator with multi-faults, but also work on double insulator strings. The recognition results on insulator dataset and UAV video show the effectiveness of our method.