基于融合分割与形态特征分析的绝缘子自爆识别OA

Aiming at the problem that the traditional method has low recognition rate of insulator self-ex-plosion defects and it is difficult to segment double-string insulators,the author proposes an insulator self-explosion recognition method based on fusion segmentation and connected domain morphological feature analysis.Firstly,the infrared image of insulator is segmented by double threshold,and the visible light image is segmented by HSI color feature,and the target area is segmented by logical operation.Then,morphological filtering and adhesion connected domain separation are carried out on the insulator target area;After that,the minimum circumscribed rectangle is located in the connected area of insula-tors,and the mathematical model of the midpoint of the circumscribed rectangle is constructed.The opti-mal separation line is found by least square method to realize the separation of double-string insulators,and finally a threshold is set for the distance between adjacent insulators to realize the identification of self-explosion defects.The experimental results show that this method has better recognition perform-ance than other common traditional methods.