机载视觉融合下输电塔金具缺失叠加残差检测  

Detection of transmission tower hardware missing overlay residual under airborne visual fusion

作　　者：孔志战 张静 孙喆 高肖松 顾燕丰 KONG Zhizhan;ZHANG Jing;SUN Zhe;GAO Xiaosong;GU Yanfeng(State Grid Shaanxi Electric Power Co.,Ltd.,Xi’an 710000,China;Xi’an Electric Power College,Xi’an 710000,China;Jiangsu Huacheng Xiehong Technology Co.,Ltd.,Wuxi 214400,China)

机构地区：[1]国网陕西省电力有限公司,陕西西安710000 [2]西安电力高等专科学校,陕西西安710000 [3]江苏华成协弘科技有限公司,江苏无锡214400

出　　处：《电子设计工程》2025年第9期53-56,61,共5页Electronic Design Engineering

基　　金：陕西省教育厅2022年科研计划一般专项(22JK0388);陕西省电力公司2023年科技项目(5226FX230004)。

摘　　要：现实输电塔环境中存在各种干扰和噪声,导致金具出现部分缺失、重叠缺失及结构混乱等问题,进而造成检测结果存在误差。为此,提出机载视觉融合下输电塔金具缺失叠加残差检测方法。利用机载视觉获取输电塔金具图像,融合梯度特征与灰度特征,根据损失函数和数据集中每个坐标点的平均距离,构建叠加残差检测结构,修正每个步骤的残差,实现输电塔金具缺失叠加残差检测。由实验结果可知,该方法 A点坐标为(1.1,1.62)、B点坐标为(1.6,1.68)、C点坐标为(1.93,1.4)、D点坐标为(1.4,1.3),与实验框选择结果一致,具有精准检测效果。Due to all kinds of interference and noise in the real transmission tower environment,there are problems such as partial missing overlapping missing,and structural disorder in the fittings,which in turn leads to errors in the detection results.Therefore,a method for detecting the missing overlay residuals of transmission tower hardware under airborne vision fusion is proposed.The image of transmission tower hardware is obtained by airborne vision,the gradient feature and gray feature are fused,and the superimposed residual detection structure is constructed according to the loss function and the average distance of each coordinate point in the data set,and the residual error at each step is corrected to realize the missing superimposed residual detection of transmission tower hardware.It can be seen from the experimental results that the coordinates of point A of the method are(1.1,1.62),point B is(1.6,1.68),point C is(1.93,1.4)and point D is(1.4,1.3),which are consistent with the experimental frame selection results and have accurate detection effect.

关 键 词：机载视觉融合 输电塔 金具缺失 残差检测 

分 类 号：TN108.1[电子电信—物理电子学]