一种基于CV模型的电晕放电紫外成像分割方法

为了判断高压设备发生电晕放电的位置和大致的放电强度,利用紫外成像仪对设备进行记录并判断高压设备的绝缘缺陷。通过C-V(Chan-Vese)模型方法对紫外图像光斑进行图像分割提取,且引入了光斑面积和光斑边界周长两个特征量来判断设备放电状态。结果表明,C-V模型提取光斑自动化程度高,耗时短,抗噪能力更强,具有很好的全局优化,能够在一定程度上减少漏检的放电区域,且引入的特征量能够较好地表征放电状态。

A method of uv image segmentation for corona discharge based on CV model

In order to judge the position of the corona discharge in the high-voltage equipment and the approximate discharge intensity, the device is recorded by the ultraviolet imager and the insulation defect of the high-voltage device is judged. The image of the ultraviolet image spot was extracted by the Chan-Vese (CV) model, and the feature area of the spot area and the perimeter of the spot boundary was introduced to judge the discharge state of the device. The results show that the CV model extracts the spot with high degree of automation. The utility model has the advantages of short time consumption, stronger anti-noise ability, good global optimization, and can reduce the discharge area of the missed detection to a certain extent, and the introduced feature quantity can better characterize the discharge state.