基于角点特征的电力绝缘子视觉识别方法

Visual recognition method of power insulator based on corner feature

为实现电力绝缘子有效清洗作业,针对电力绝缘子自主检测和识别定位问题,分析绝缘子串的几何特征,提出了一种基于角点特征的电力绝缘子视觉识别方法。首先,详细阐述了角点检测的基本原理,给出了角点检测的理论模型;然后,以电力绝缘子为检测对象,提出了绝缘子串角点检测的基本流程;最后,在Open CV环境下对绝缘子角点检测进行了仿真。结果表明,该方法能够较好地检测到绝缘子的角点,将整串绝缘子上所有检测到的角点连接得到几何区域,即可包含绝缘子串,从而实现整串绝缘子的视觉检测。

In order to realize the effective cleaning operation of electric power insulators,this paper proposes a visual recognition method for electric power insulators based on corner feature by analyzing the geometric characteristics of insulator strings.Firstly,it elaborates on the basic principle of corner point detection and gives a theoretical model of corner point detection,then,taking power insulators as the detection object,it proposes the basic process of insulator corner point detection.Finally,a simulation analysis of insulator corner point detection is carried out in Open CV environment.The results show that the algorithm can detect the corner points of insulators better,and the geometric area obtained by connecting all detected corner points on the whole string of insulators can contain the insulator string,thus realizing the visual detection of the whole string of insulators.