Condition Evaluation of the Contaminated Insulators by Visible Light Images Assisted With Infrared Information
Lijun Jin, Zhiren Tian, Jianyong Ai, Yingyao Zhang, Kai Gao · IEEE Transactions on Instrumentation and Measurement · 2018
Pollution flashover is harmful and has caused huge economic losses. Thus, it is important to evaluate the contamination condition of insulators to prevent the pollution flashover. The objective of this paper is to discriminate the insulator contamination grades by the noncontact method. In this paper, a method of visible light images assisted with infrared information is proposed. The proposed method evaluates the contamination condition using the visible light images, and is corrected by the infrared information in the wet environment. A mathematical model is established to study the relationship between the equivalent salt deposit density (ESDD) (mg/cm2) and the color features. The influences of illumination L and relative humidity (RH) on the color features are analyzed and their eliminated methods are proposed. The infrared image information is proposed to assist the visible light image method when RI ≳ 60%. In the experiment, the visible light images of the artificial polluted insulators are acquired by a digital camera under different illuminations L and RI!s, and the average value of V component Vave in the YUV color space is selected for the discrimination of contamination grades. In the test, the proposed method gives an 84% accuracy rate when RI!2. Therefore, it might be concluded the condition of naturally polluted insulators could be evaluated by the proposed method of visible light images assisted with infrared information.