基于图像亚像素真空开关分闸速度的检测

Novel Image Processing Algorithm for Opening Velocity Evaluation of Vacuum Circuit Breaker

操动机构运动参数是影响真空开关开断性能的重要因素之一,如何高精度检测操动机构运动参数是该领域重要研究课题。本文以操动机构分闸速度为研究对象,据电弧实验触头运动序列图像中电弧边缘与触头边缘重合的特点,分别采用二次多项式插值法、灰度矩法亚像素边缘检测法对动触头精确定位,实现真空电弧亚像素边缘点提取,采用霍夫变换剔除干扰点,而后对有效的电弧边缘点拟合,得到电弧真正的边缘位置,最后用帧间平均速度代替触头运动瞬时速度。结果表明,基于灰度矩的亚像素边缘检测法对动触头定位精度可达0.01~0.35像素,且分闸速度拟合曲线较平滑,拟合误差较小,且其检测精度具有一定的稳定性,可行性好。

A novel algorithm was developed to precisely determine the opening velocity of a vacuum circuit breaker by processing the arc images taken with a charged coupled device （CCD） Camera. Two schemes： gray level moments algorithmand image gradient gradation,were used to accurately locate the edge positions of the fixed and moving electrodes. Moreover, the interference edge points were removed by Hough transform, and the true arc edge line was determined by least square data fitting of the remaining edge points. The instantaneous velocityof the mov- ing electrode can be calculated in terms of the average inter-frame speed. The results show that the newly-developed sub-pixel detection algorithm, gray level moments algorithm, is capable of more precisely locating the moving elec- trode with an error of 0.01pixel to 0.35pixels. The fitted opening velocity with smaller fitting error was found to be smooth, accurate and fairly stable.