改进型PCNN在绝缘子图像分割中的应用  

The Application of Improved PCNN in Infrared Insulator Image Segmentation

作　　者：徐雪涛[1] 李宝树[1] 崔克彬[2] 魏文力[1] 

机构地区：[1]华北电力大学电气与电子工程学院.河北保定071003 [2]华北电力大学控制与计算机学院,河北保定071003

出　　处：《电瓷避雷器》2013年第3期24-29,共6页Insulators and Surge Arresters

摘　　要：针对高压输电线路绝缘子状态检测中光照不均匀、对比度不强的红外绝缘子图像分割,提出了一种基于类内绝对差准则的改进型PCNN图像分割算法。对最小类内绝对差法进行改进,引入背景与目标的面积差因子,确定最佳分割阈值后通过PCNN算法迭代优化进行图像分割。与经典OTSU算法、基于最小类内绝对差准则的PCNN算法进行比较。实验结果表明,本文方法能够取得更好的分割效果,并且具有较强的实用性。In high voltage transmission line insulator state detection, for light heterogeneous, contrast is not strong infrared insulator image segmentation, a new method of improved PCNN image segmentation based on the interclass absolute difference is proposed. The minimum interclass absolute difference method was improved, the area difference factor between background and target is introduced, then image segmentation is carried out through PCNN algorithm iterative optimization method after the best threshold is determined. And compared with classical OTSU algorithms, PCNN method based on the minimum interclass absolute difference. The experimental results show that this method is more efficient and has strong applicability.

关 键 词：绝缘子 类内绝对差 PCNN 图像分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]