Detection of pollution components on the surface of transmission line insulators based on hyperspectral technology

Insulators of transmission lines are easy to accumulate pollution on their surfaces due to long-term exposure to the atmospheric environment, and the pollution flashover accidents it causes seriously endanger the safety of power system. And the flashover voltage is closely related to the pollution components on insulators' surface. This paper proposes to use hyperspectral imaging technology to establish the identification model of pollution components, which makes up for the shortage of traditional insulator detection methods. Firstly, contaminated insulator samples with different pollution components were prepared in the laboratory, their hyperspectral spectral lines were obtained for comparison, then the differences between the spectral lines of the samples with different pollution components were analyzed. At last, a random forest model was built to realize the identification and classification of the samples with different pollution components, and the accuracy of the model was more than 95%. The results show that the method for detection of insulator surface pollution component based on hyperspectral imaging technology is feasible and provides a new idea for the online detection of insulators of transmission lines.