A self-adaptive filtering algorithm based on morphology theory is proposed to deal with the noise in online monitoring data of leakage current of arrester.The effects of the open-closing filter and the close-opening filter are compared

and the statistic bias of these traditional morphological filters is analyzed.The open-closing filter and the close-opening filter are combined

in which the weighting coefficient can be adjusted with the self-adaptation approach.Simulation shows that the proposed algorithm can suppress the impulse noise and white noise in on-line monitoring data of leakage current of zinc oxide arrester to restore the general regularity of the data.