Insulator is one of the main electrical equipment in power grid.Pollution flashover of insulators seriously threatens the safety and stability of power system.It is of great significance to mater the insulation status of polluted insulators timely.In this paper

the frame is extracted from the ultraviolet video of insulator discharge

the extracted image is segmented by ultraviolet spot

the of spot region area and its time series are calculated

and the relationship between the spot area and leakage current in different discharge stages is analyzed

based on which the insulating state of insulator is given and flashover warning is carried out.An artificial fog chamber with continuously adjustable humidity was constructed.Video of discharging contamination of disc insulators was recorded by UV imager.The ultraviolet characteristics of discharging were extracted.Combining with naked eye observation

leakage current signal and other related parameters

the insulating state of insulators was divided into four stages:"safe"

"warning"

"alarm"and"dangerous".The arithmetic mean

maximum

standard deviation and discharge frequency of the spot area sequence were extracted

and the sample data were analyzed and clustered.The practical application of the project shows that the spot area sequence is effective for the insulation condition assessment of insulators.