高压设备放电紫外图像特征量提取方法研究

Extraction Method of Ultraviolet Image Parameters of Discharge for High-voltage Device

紫外成像仪提供光子数来表征放电强度,但是因为存在环境噪声干扰,光子数并不准确,有时甚至误差很大。为了更准确地描述高压设备的放电状态,文中提出了利用数字图像处理技术提取图像参量来表征放电特性的新方法。首先将试验得到的紫外图像灰度化,再通过阈值分割法二值化,利用数学形态法滤波以及通过小区域消除算法提取了放电区域;之后,采取不规则图形边界跟踪算法得到放电区域边界点的坐标,并对光斑面积、边界周长、长轴、短轴等参数进行定义并计算;最后采用radon变换算法,获得沿高压设备表面的放电强度空间分布曲线。变电站现场绝缘子表面放电实例应用表明了上述方法能正确有效地评估放电特性。

Ultraviolet（UV） imager provides UV photon number to characterize the discharge intensity.However, the photon number is inaccurate due to environmental noise interference. For the purpose of more accurate description of high-voltage device discharge characteristics, a new method for extracting UV image parameters based on image processing technique is proposed in this paper. The discharge regions are extracted with the method of threshold segmentation, morphological filtering, and small region elimination algorithm. Then the coordinates of the boundary points of each discharge region are obtained through the multi-zone boundary tracking algorithm, and the parameters of facular area, perimeter and other parameters are defined and calculated to quantify the discharge intensity. Further, the radon parallel line projection algorithm is applied to get the discharge intensity distribution on surface of high-voltage equipment. A practical discharge detection example of a substation insulator demonstrates that the proposed method can evaluate the discharge characteristics effectively.