摘要
传统方法在架空输电线路异物定位识别中应用效果不佳,不仅异物正确定位识别次数比较少,而且识别时间比较长,无法达到预期的异物定位识别效果,因此,文章提出基于深度学习的架空输电线路异物定位识别方法。采用无人机航拍技术采集输电线路图像,采用图像分割算法对原始图像进行分割处理,提取异物图像目标区域并对目标区域中值滤波进行处理。利用深度学习网络模型提取输电线路轮廓特征,通过特征比对判断输电线路是否存在异物,并利用欧式距离定位到异物位置。经实验证明,该设计方法正确识别次数较多,且识别时间较短,在架空输电线路异物定位识别方面具有良好的应用前景。
The traditional method has poor application effect in the positioning and recognition of foreign objects in overhead transmission lines.Not only the number of correct positioning and recognition of foreign objects is relatively small,but also the recognition time is relatively long,which cannot achieve the expected foreign object positioning and recognition effect.Therefore,this paper proposes a deep learning-based overhead transmission line foreign object positioning and recognition method.The drone aerial photography technology is used to collect transmission line images,and the image segmentation algorithm is used to segment the original image.The target area of the foreign object image is extracted and the median filter of the target area is processed.The deep learning network model is used to extract the contour features of the transmission line,determine whether there is a foreign body in the transmission line through feature comparison,and use the European distance to locate the foreign body position.Experiments have proved that the design method has many correct recognition times and short recognition time,which has good application prospects in the positioning and recognition of foreign objects in overhead transmission lines.
作者
易敏
YI Min(State Grid Sichuan Electric Power Company Meishan Power Supply Company,Meishan,Sichuan 620010,China)
出处
《计算机应用文摘》
2023年第19期113-116,共4页
Chinese Journal of Computer Application
关键词
深度学习
架空输电线路
异物定位识别
无人机航拍技术
图像分割算法
deep learning
overhead transmission lines
foreign body positioning identification
aerial photography technology
image segmentation algorithm