基于区域定位的绝缘子图像分割  

Segmentation of the Blue Insulator Images Based on Region Location

作　　者：马帅营[1,2] 安居白[1] 陈舫明[3] 

机构地区：[1]大连海事大学信息科学技术学院,辽宁省大连市116602 [2]大连民族学院现代教育技术中心,辽宁省大连市116605 [3]绍兴电力局,浙江省绍兴市312000

出　　处：《电力建设》2010年第7期14-17,共4页Electric Power Construction

摘　　要：按绝缘子颜色对图像进行分类,对其中颜色特性突出的蓝色绝缘子图像,结合倾斜校正和投影法区域定位,获得较小范围下的绝缘子图像,进而利用饱和度分量图像和最大类间方差法进行绝缘子的分割,得到的分割结果精确,绝缘子损坏部位突出,为后续的绝缘子故障判断提供了依据。Insulator images are divided into different types.The blue insulator image,which is very prominent in color,is used to get a concentrated insulator image exclusive of complicated background by means of tilt correct and projection.Through segmentation of the treated insulator images by using the saturation component image and the maximum variance between clusters,an accurate segmented image is obtained and the damaged location stands out ,which can provide a basis for the follow-up fault diagnosis.

关 键 词：绝缘子图像 图像分割 区域定位 饱和度分量 最大类间方差法 阈值 

分 类 号：TP753[自动化与计算机技术—检测技术与自动化装置]