基于探地雷达的输电线路走廊隐性病害智能识别

Intelligent identification of hidden disease in transmission line corridor based on ground penetrating radar

为消除输电线路走廊地下运维盲区,及时发现杆塔地基内部以及周围地下环境隐性结构病害,避免出现走廊范围内地面塌陷等安全事故,基于探地雷达的高效无损地下探测手段,提出了一种基于改进FasterR-CNN的输电线路走廊隐性病害识别模型,并针对现有探地雷达数据集样本量不足的问题,提出了一种利用生成式对抗网络扩充地下隐性病害探测数据的方法,降低构建数据集难度。实验结果表明:所提出模型的检测精准度达到92.34%,能有效识别出复杂背景下输电线路走廊隐性病害。研究结论可为未来输电线路走廊地下隐患排查提供工程应用理论依据。

In order to eliminate underground blind areas in transmission line corridor operation and maintenance,timely discover hidden structural diseases inside the tower foundations and the surrounding underground environment,thus avoiding safety accidents such as ground collapse within the corridor,based on the efficient and nondestructive ground penetrating radar means,a transmission line corridor hidden disease identification model based on improved Faster R-CNN is proposed,and a method of expanding underground hidden disease detection data using generative adversarial network is proposed to reduce the difficulty of constructing datasets for the problem of insufficient sample size of existing ground-penetrating radar datasets.The experimental results show that the detection accuracy of the proposed model reaches 92.34%,which can effectively identify the hidden diseases in transmission line corridors in complex backgrounds.The research results can provide engineering application theoretical basis for the underground hidden danger detection of transmission line corridors in the future.