中文摘要：

真空开关分闸速度对真空电弧形态和弧后介质强度的恢复速度有重要影响,实现其分闸速度高精度检测并研究分闸速度对电弧特性影响是真空开关领域重要课题之一。为实现触头分闸速度高精度检测,本文提出用Zernike矩算法从真空开关分闸序列图像中精确跟踪动触头位置,进而计算出触头分闸速度的检测方法。首先利用Canny算子对所有可能的电弧边缘进行像素级定位,缩小精确定位的检测范围,然后用Zernike矩算子对所有可能的电弧边缘重新检测,最后实现触头分闸速度的亚像素级精确测量。实验结果表明：本算法能够高效的实现真空开关分闸速度的高精度检测,动触头边缘定位能精确到小数点后两位,算法运行速度较像素级精度提高了35%。

英文摘要：

A novel algorithm was developed to precisely evaluate the contact opening speed of vacuum switch by processing the arc images taken with charged coupled device（CCD） camera,based on Zernike moment algorithm. First,the precise positioning range was narrowed by locating all the possible edges of the arc at pixel level with Cannyoperator. Next,all the possible edges of the arc were re-evaluated with Zernike momentsoperator. Finally,the opening speed of the vacuum switch was accurately determined at sub-pixel level. The newly-developed algorithm was tested. The test results show that the novel algorithm is capable of accurately determining the opening speed of the contacts. The position of the moving contact edge was located with a precision up to two decimal places; and the calculation speed of the new algorithm was 35% faster than that of the algorithm at pixel level.