基于数字射线检测技术的电缆熔接头主绝缘缺陷研究

近年来高压电缆熔接头技术逐步在现场开展应用,但制作过程中关键工艺缺乏有效的管控手段。为了解决此类问题,利用图像深度智能处理数字射线(Image Depth Intelligent Processing Digital Ray,IDIP-DR)检测技术,对高压电缆熔接头关键制作工艺进行检测,检测后发现主绝缘存在气泡施工工艺问题,并对缺陷产生的原因进行分析,提出相应的工艺管控建议。

基于Canny算法的10 kV电缆主绝缘表面缺陷识别方法

主绝缘表面存在划痕与污渍类施工缺陷容易导致绝缘故障,采用图像识别技术可以智能管控施工缺陷。针对10 kV电缆主绝缘表面划痕与污渍缺陷边缘不易检测的问题,本文提出一种基于Canny算法的10 kV电缆主绝缘表面缺陷识别方法。本文首先利用尺寸不变特征变换(scale-invariant feature transform,SIFT)方法完成对电缆主绝缘图像的拼接,再借助Canny算法计算主绝缘表面缺陷边缘的梯度与幅值、抑制极大值信息点、去除伪边缘,得到由划痕与污渍产生的缺陷边缘信息。实验结果表明,由本文算法检测后电缆主绝缘表面缺陷图像的边缘轮廓具有较高的纯净度和完整性,缺陷边缘更加清晰,算法平均准确率达到95.42%,验证了算法的有效性。

Investigation of Residual Donor Defects in Undopedan d Fe-Doped LEC InP

The free electron concentration of as-grown liquid encapsulated Czochralski (LE C) InP measured by Hall effect is much higher than the concentration of net dono r impurity determined by glow discharge mass spectroscopy.Evidence of the existe nce of a native donor hydrogen-indium vacancy complex in LEC undoped and Fe-do ped InP materials can be obseved with infrared absorption spectra.The concentra tion increase of the donor complex correlates with the increase of ionized deep acceptor iron impurity Fe concentration in Fe-doped semi-insulating (S I) InP.These results indicate that the hydrogen-indium vacancy complex is an im portant donor defect in as-grown LEC InP,and that it has significant influence on the compensation in Fe-doped SI InP