基于统计形状模型的绝缘子掉串检测算法

The Detection Algorithm of Cracked Insulator Based on Statistical Shape Models

摘要:

在电力系统中,利用计算机视觉和图像处理技术对绝缘子的状态进行检测,在保障电力系统的安全运行方面具有非常重要的作用。针对绝缘子掉串问题,本文提出了一种基于统计形状模型的绝缘子掉串检测算法,该算法首先对输入图像进行预处理；其次提取感兴趣的目标区域,该区域内应包含绝缘子信息；然后利用GLOH描述子来精确定位绝缘子,再利用统计形状模型在图像分割方面的优势来精确分割绝缘子；最后对分割出的绝缘子的位置进行分析,以达到检测绝缘子是否掉串的目的。实验结果说明了本文方法可以准确地判断出绝缘子是否存在掉串现象。

Abstract:

In the power system, there is a very important role in the protection of the safe operation of the power system by using computer vision and image processing technology to detect the cracked Insulator. Aiming at the problem of the cracked insulator because of self leads to the shedding of glass insulator, this paper presents a detection algorithm of cracked insulator based on Statistical Shape Models(SSMs). The algorithm firstly preprocesses the input image. Secondly, we extract the interest target region which should contain insulator information. Then GLOH descriptor is used to accurately locate the insulator, and then the SSMs are used to accurately segment the insulator with their advantages in the image segmentation; Finally, that the insulator is cracked or not is determined by analyzing the positions of the separated insulators. The experimental results show that the algorithm can accurately locate the cracked insulator.