摘要
为解决绝缘子缺陷检测效果不佳等问题,提出了基于Faster-RCNN算法的无人机巡检图片绝缘子缺陷检测方法。通过对绝缘子运行过程中的缺陷数据特征进行采集分析和绘制,判断巡检图片中的绝缘子运行状态,并结合Faster-RCNN算法对巡检图像进行局部分割和增强处理,从而提高图像检测的精准性,实现对绝缘子缺陷的有效检测。最后通过实验证实,基于Faster-RCNN算法的无人机巡检图片绝缘子缺陷检测方法具有较高达95%的准确性。
In order to solve the problem of poor insulator defect detection effect,a defect detection method based on Faster-RCNN algorithm for unmanned aerial vehicle inspection images is proposed.Through collecting,the defect data features are analyzed and drawned in the insulator operation process,the insulator operation state in the inspection image is judged,and the inspection image is locally segmented and enhanced by combining Faster-RCNN algorithm,so that the accuracy of image monitoring is improved,and the effective detection of insulator defects is realized.Finally,the experiment proves that the defect detection method of UAV inspection picture insulator based on Faster-RCNN algorithm has a 95% higher accuracy.
作者
宁柏锋
董召杰
NING Baifeng;DNG Zhaojie(Shenzhen Power Supply Co.Ltd,Shenzhen Guangdong 518000,China;Digital Grid Research Institute,China Southern Power Grid.,Guangzhou 518000,China)
出处
《自动化与仪器仪表》
2020年第7期194-197,共4页
Automation & Instrumentation
基金
中国南方电网有限责任公司科技项目(No.090000KK52170124)。
