基于紫外图像的接触网棒瓷绝缘子污秽状态检测

为了检测接触网绝缘子的污秽状态,采用紫外成像检测技术对接触网常用的棒瓷绝缘子进行研究。对绝缘子人工涂污并对污秽绝缘子在人工气候室内进行了放电试验,拍摄紫外图像和视频,对图像进行处理,提出用阈值滤波对二值图像去噪的方法,提取并定义了"光斑面积百分比"这一图像特征。通过对不同相对湿度下不同污秽等级绝缘子的污秽放电试验,得出湿度和污秽等级对绝缘子放电的影响规律,并将此作为污秽状态评估的依据。建立以光斑面积百分比、相对湿度为输入,绝缘子污秽程度为输出的模糊推理模型,并将其与视频图像处理部分融合,构建基于紫外视频图像的外绝缘污秽状态检测系统。

Rod Porcelain Insulator Filth State Detection of Catenary Based on Ultraviolet Image

In order to detect the filth state of catenary insulator, this paper use the ultraviolet image detection technique to study the rod porcelain insulators of catenary. The insulators were smeared, and then discharged in phytotron, which were recorded by ultraviolet images and videos. Then the images were processed. Based on the extracted characteristics, the threshold filtering was used for binary image denoising, and the "flare area percentage" was defined to describe the characteristics of images. By the discharge experiments of different pollution level insulators under different relative humidity, the influence of humidity and pollution level on insulator discharge was analyzed. A fuzzy inference model was established, where the average flare area percentage and relative humidity were set as the inputs and the insulator filth degree as the output. A UV video image based filth state detection system for external insulation was formed by combining the inference model with the video image processing system.