This work aims to classify the contamination state in insulators using statistical signal processing approaches. When subjected to high voltage, insulators can radiate radio frequency signals (corona effect). The histograms, normality, and correlation coefficient statistical methods are used to estimate the pollution state in glass insulators when subjected to 8 kV. It is shown that in particular situations the histograms can be used to distinguish clean and dirty insulators. The histogram limitation analysis can be improved using the correlation coefficient and the normality or Gaussianity test. Indeed, it is shown that using these parameters into an analysis per sub-bands, it is possible to estimate the pollution state of the insulators. That is, the analysis using these tools checks if the insulators spectra under test are noticeably different from the clean one, used as reference. It is achieved eliminating the fast variation of the correlation coefficient based on the amplitude and width of the peaks. The tests were done up to the frequency of 1 GHz using measured data.