Composite insulator is prone to internal defects such as air gaps which seriously endanger the safety of power systems. To detect the internal defects in composite insulators, a terahertz (THz) imaging method based on edge detection algorithm was proposed in this article. First, air gap defect sample consisting of planar multilayers was imaged based on the envelope area of the THz reflection waveform. Then, Canny operator was applied to extract the edges of the defects. After that, the time interval of defect characteristic pulses was used to calculate the defect depth and then 3-D imaging of defects was, therefore, obtained. The experimental results showed that the imaging method based on envelope area was very effective in determining the edges of air gap defect and it had better performance compared with those using pulse amplitude or time interval as parameter for defect imaging. According to the results, the accuracy was quite satisfactory when the defect depth was above 0.339 mm. To verify the effectiveness of the proposed method, the developed THz imaging system was also applied to the rod of composite insulator. It was found that the relative error was smaller along the axial direction of the rod than the lateral direction.

中文翻译：

---

基于边缘检测算法的复合绝缘子缺陷太赫兹成像方法


复合绝缘子容易出现气隙等内部缺陷，严重危害电力系统的安全。为了检测复合绝缘子内部缺陷，提出了一种基于边缘检测算法的太赫兹（THz）成像方法。首先，基于太赫兹反射波形的包络面积对由平面多层组成的气隙缺陷样品进行成像。然后，应用Canny算子提取缺陷的边缘。之后，利用缺陷特征脉冲的时间间隔来计算缺陷深度，从而获得缺陷的3D成像。实验结果表明，基于包络面积的成像方法对于气隙缺陷边缘的确定非常有效，并且与使用脉冲幅度或时间间隔作为缺陷成像参数的方法相比具有更好的性能。结果表明，当缺陷深度在0.339 mm以上时，精度相当满意。为了验证所提方法的有效性，所开发的太赫兹成像系统也应用于复合绝缘子棒材上。结果发现，沿杆轴向的相对误差小于横向的相对误差。