基于植被指数的绝缘子表面藻类检测及表征方法的研究

Detection and Characterization Method of Algae on Insulator Surface Based on Vegetation Index

在一些温暖湿润地区,运行的绝缘子表面出现了藻类生长现象,且在涂有RTV的绝缘子和复合绝缘子上现象更严重。藻类生长在绝缘子表面会导致其耐污闪性能下降。本文通过数字图像处理技术检测绝缘子表面的藻类,研究了超绿指数EXG、动态阈值超绿指数DEXG、Hague植被指数HVI、植被提取颜色指数CIVE等四种常用植被指数在附有藻类的绝缘子图像中对藻类区域和背景区域的分割效果,采用相对错分率评价定量评价图像分割效果,并用藻类覆盖率表征绝缘子表面藻类的生长程度。研究发现EXG和HVI在绝缘子图像中的分割效果较好,DEXG有藻类区域识别不足问题,而CIVE有藻类区域识别过度问题。文末提出采用EXG和HVI相结合的优化算法进行图像分割,进一步提高藻类识别准确度。

Algae has been found growing on the surface of insulators installed in harm and humid areas,especially those coated with RTV and composite insulators.The growth of algae will lead to the degradation of anti pollution flashover performance of insulators.Based on digital image processing technology,the segmentation of insulator images with algae is studied in this paper,using four common vegetation indexes,that is EXG,DEXG,HVI and CIVE,and the effect of image segmentation is compared.The relative misclassification rate is used to evaluate the image segmentation effect quantitatively and the algae coverage rate is used to characterize the growth degree of algae on insulator surface.It is found that EXG and HVI have better segmentation effects in insulator image,while DEXG identifies algae as the background mistakenly and CIVE identifies the background as algae mistakenly.In the end,an optimized algorithm based on EXG and HVI is proposed to segment the image,which further improves the accuracy of algae identification.