摘要
提出了利用数字图像处理技术和曲线拟合理论自动检测复合绝缘子表面接触角的方法。针对绝缘子表面水珠图像的特点,首先利用中值滤波、灰度拉伸等算法预处理图像,去除高频噪声,并从背景中突出水珠轮廓;然后利用Robert算子边缘检测、迭代最优阈值区域分割等算法将水珠与背景分离;再采用骨架提取迭代和八邻域轮廓跟踪等算法提取水珠的边界曲线;最后以正交多项式为基函数拟合水珠边界曲线,继而求导得到接触角。该方法能够现场检测运行中的绝缘子,测量过程不需人工干预,客观精确。多次测量同一复合绝缘子的接触角,均方误差在2%之内。
A method to auto-detect contact angles of water drop on the surface of a composite insulator with image processing and curve fitting is proposed in this paper. Firstly, according to the image characteristics of water drop on insulator＇ s surface, the median filter algorithm and the gray stretch algorithm were deployed to suppress noise and highlight the outline of water drop from background. Secondly, in order to separate water drop from background, the process of image segmentation was implemented using the edge detecting algorithm of Robert and the iterative region partition algorithm on optimized threshold. Then, the proposed method combined sketch extraction iteration algorithm and eight neighborhood outline tracking algorithm to extract the corresponding boundary curve of water drop. Finally, the polynomial fitting was deployed to fit the boundary curve of water drop, and contact angles were calculated by differentiate the polynomial deduced from the polynomial fitting. The proposed method could detect composite insulator in operation. Moreover, the detection could be processed without manual intervention which makes the detect result accurate and objective. Experiments show that, implementing the proposed detection of contact angles to the same composite insulator many times, the root mean square error is below 2%.
出处
《电机与控制学报》
EI
CSCD
北大核心
2009年第2期232-237,244,共7页
Electric Machines and Control
基金
国家自然科学基金(50677047)
关键词
绝缘子
憎水性
接触角
图像处理
electric insulators
hydrophobicity
contact angle
image processing
