无人机航拍的输电线路弧垂测量研究

Research on transmission line sag measurement based on UAV aerial photography

针对目前的弧垂测量方法存在效率低、成本高的问题,该文提出了一种输电线路弧垂测量的新方法。该方法基于融合卷积块注意力模块(CBAM)注意力机制的间隔棒自动分割算法CBAM-Mask-RCNN,结合光束法平差、空间前方交会和空间曲线拟合等经典算法,以无人机巡检视频数据为基础,低成本、高效率地实现弧垂测量。经验证,该文提出的CBAM-Mask-RCNN算法在自建数据集上AP指标达到了72.44%,优于Yolact++、U-Net和Mask-RCNN等算法。通过消融实验也证明了该文所采用的CBAM-block模块能有效提高算法的分割性能。此外,在5个不同输电线档距中进行弧垂测量实验,该文方法的误差均在±2.5%以内,平均误差为0.31%,满足工程要求,验证了该方法的有效性。

Aiming at the problems of low efficiency and high cost of current arc sag measurement methods,this paper proposed a new method for arc sag measurement of transmission lines.The method was based on the CBAM-Mask-RCNN algorithm,which incorporated the convolutional block attention module(CBAM)attention mechanism,and combined classical algorithms such as beam adjustment,spatial front rendezvous,and spatial curve fitting,to achieve arc sag measurement in a cost-effective manner based on UAV inspection video data.It was experimentally verified that the CBAM-Mask-RCNN algorithm proposed in this paper achieved an AP index of 72.44%on the self-built dataset,outperforming algorithms such as Yolact++,U-Net,and Mask-RCNN.The CBAM-block module used in this paper was also proved to be effective in improving the segmentation performance of the algorithm through ablation experiments.In addition,arc sag measurement experiments were carried out in five different transmission line stall distances,and the errors of the method in this paper were all within±2.5%,with an average error of 0.31%,meeting the engineering requirements and verifying the effectiveness of the method.