基于Grabcut分割和填充物面积判别的复合绝缘子断串诊断

针对传统 GrabCut 算法需要人工初始化而引起图像分割效率低的问题，结合Otsu方法，提出了一种新的Grabcut自动化算法对复合绝缘子进行分割。首先，对原始图像进行HSV空间转换和加权的灰度化处理。其次，对V通道图像和灰度化后的图像进行Otsu分割并进行或逻辑融合，以此来确定目标绝缘子区域，并结合最大连通域定位绝缘子位置坐标完成Grabcut框取初始化，实现Grabcut的自动化。最后，针对绝缘子断串判别准确率低的问题，通过对绝缘子分割图像的最小外接矩形加框、填充、去原图的方式，提出一种新的面积判别方式来诊断并定位故障位置。实验结果表明：Grabcut自动化算法可以很好地分割出目标绝缘子，分割准确率可以达到96.6%以上。所提出的面积判别方法对于具有断串故障的绝缘子检测率可以达到96.6%以上，对于无故障的误检率为6.7%以下。

Broken string diagnosis of composite insulator based on Grabcut segmentation and filler area discrimination

In view of the low efficiency of image segmentation caused by manual initialization of traditional Grabcut algorithm, a new Grabcut automatic algorithm for composite insulator segmentation is proposed based on Otsu method. First, the original image is transformed into HSV space and weighted grayscale. Secondly, the V-channel image and the grayed image are segmented by Otsu and fused by logic to determine the target insulator area, and the initialization of Grabcut frame is completed by combining the position coordinates of the largest pass domain positioning insulator, so as to realize the automation of Grabcut. Finally, in view of the low accuracy of insulator broken string identification, a new area discrimination method is proposed to diagnose and locate the fault location by adding frame, filling and removing the original picture of the smallest external rectangle of the insulator segmentation image. The experimental results show that the Grabcut automatic algorithm can segment the target insulator well, and the segmentation accuracy can reach more than 96.6%. The detection rate of the proposed area discrimination method can reach more than 96.6% for the insulator with broken string fault and less than 6.7% for no fault. This work is supported by the Natural Science Foundation of Heilongjiang Province (No. LH2019E016).