Insulator contamination degree non-contact detection method and device

Abstract

The invention discloses an insulator contamination degree non-contact detection method and device. Area-of-interest hyperspectral spectral line full-wave band data of insulators with different contamination degrees can be subjected to unsupervised feature learning to obtain adjacent-level connecting weights and feature representation; an insulator contamination degree prediction model based on a level extreme learning machine is built according to the connecting weights, feature representation and the spectral line data of the insulators with different contamination degrees; the insulator contamination degree prediction model is optimized; the contamination degrees of all local areas of the insulators are judged by the optimized insulator contamination degree prediction model, and finally,large-scale, rapid and accurate non-contact detection of the insulators with unknown contamination degrees is realized. A hyperspectral technique can be effectively assisted by introducing a level extreme learning machine algorithm to judge the contamination degrees of the insulators, the detection process is simple and efficient, cleaning operation of the insulators is more targeted, pertinent and objective, and a scientific basis is provided for non-contact insulator contamination degree judgment.

Description

technical field [0001] The present application relates to the technical field of analysis and measurement control, in particular to a non-contact detection method and device for pollution degree of insulators. Background technique [0002] The pollution flashover of the insulator seriously affects the safe operation of the power supply system. Under fog, dew, drizzle or acidic wet deposition, the flashover voltage of the insulator will drop greatly after the surface of the insulator is dirty and absorbs water, resulting in the occurrence of pollution flashover accidents. Therefore, it is very necessary to know the degree of pollution on the surface of insulators in a timely, simple and large-scale manner, to formulate targeted and accurate cleaning cycle strategies, and to realize fast, non-destructive, non-contact pollution and live detection of insulators. [0003] At present, the detection of pollution level of insulators is mainly through regular spot checks of salt den...

AI Technical Summary

Problems solved by technology

[0003] At present, the pollution degree of insulators is mainly detected by regular spot checks of the salt density and dust density of the insulator surface pollution. This method takes a long time and cannot reflect the local pollution degree of the insulator surface.

The existing hyperspectral detection methods may lack corresponding efficient and accurate discrimination methods and the means to achieve large-scale detection. The application of the method still requires the subjective judgment of workers, and it is impossible to achieve an objective and accurate judg

Images