航拍图像中绝缘子串的轮廓提取和故障检测  

Contour Extraction and Fault Detection of Insulator Strings in Aerial Images

作　　者：方挺[1] 董冲[1] 胡兴柳[1] 王彦[1] 

机构地区：[1]安徽工业大学电气与信息工程学院,安徽马鞍山243000

出　　处：《上海交通大学学报》2013年第12期1818-1822,共5页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金资助项目(51007002;61201109)

摘　　要：针对绝缘子爆裂和缺失将导致整个输电线路绝缘能力下降的常见故障,提出一种通过计算绝缘子数目实现故障检测的算法.输电线路绝缘子串中绝缘子数目相对固定,通过计算绝缘子的数目,进而达到故障检测的目的,可应用于航拍直升机智能巡检系统中.利用Otsu算法与形态学滤波对图像进行预处理;提出免疫遗传Snake算法优化绝缘子串凹陷区域边界,得到较为完整的绝缘子串轮廓;用椭圆拟合和连通区域方法计算出绝缘子的数目.该算法能够有效地提取绝缘子轮廓并计算出绝缘子数目,为进一步实现精准、稳定检测绝缘子缺失和爆裂等故障奠定了良好的预研基础.Based on the fact that insulator bursting and missing are threats to the safety of electric power, this paper proposed an algorithm to detect fault by calculating the number of insulators. Because of the fixed number of insulators, an image processing method was adopted to detect the fault of insulators by calculating the number of insulators, which can be applied to intelligence patrol system. Otsu and morpho- logical filtering were used to preprocess the image, the immune genetic snake algorithm was adopted to op- timize insulator concavity regions to get a complete contour of the insulator strings, and the ellipse fitting and connected region algorithm were used to calculate the number of string of insulators. The proposed al gorithm can extract the initial contour of insulator strings effectively and calculate the number of insula- tors, establishing a good foundation for further study of insulator fault detection.

关 键 词：绝缘子轮廓提取 免疫遗传算法 SNAKE模型 故障检测 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]