基于聚类分析和Adaboost算法的绝缘子串识别

Insulator recognition based on clustering analysis and Adaboost algorithm

绝缘子串作为输电线路中最重要的基础设施之一,对其准确识别是实现运行状态监测与故障诊断的重要前提。为了提高绝缘子串识别的准确率,提出一种基于平聚类分析和Adaboost的绝缘子串识别方法。首先,利用LSD线段检测提取图像中的线段;然后,统计所有线段的长度、方向及中心点位置,通过聚类分析检测平行线段,初步确定绝缘子串区域;最后,提取初定位区域的不变矩特征,运用训练的Adaboost分类器进行进一步识别。实验中,绝缘子串在平均耗时0.23s的基础下识别准确率达到91.5%。结果表明,所提方法具有较好的识别准确率和快速性。

As one of the most important infrastructures of transmission line, precision location of insulators is the primary premise of condition monitoring and fault diagnosis. In order to improve the positioning accuracy, an insulator recognition method is put forward based on clustering analysis and Adaboost algorithm in this paper First, the LSD line segment detection is used to extract the line segrnents and the angular points from the pictures Then the length, direction and center point position of a segments are gathered statistics. Cluster analysis is used to detect parallel lines to identify the insulator string region nitially. And finally, the moment invariant of preliminary targeted area is extracted, trained Adaboost classifier is used to get further identification. The experiments are completed, in which the location accuracy can reach 91.5% with the short operation time of 0.23s per image The results show that this method has high accuracy and good rapidity