<p>Real-time monitoring of the distance between highly mobile equipment and transmission lines is of great significance to ensure the safe maintenance and construction of these infrastructures. We propose a&#xa0;novel identification and measurement technique that utilizes Light Detection and Ranging (LiDAR) scanning to achieve 3D perception. Firstly, a&#xa0;local dense 3D point cloud map is created through a&#xa0;short-term Simultaneous Localization and Mapping (SLAM) method. Then, a&#xa0;deep 3D neural network is implemented by integrating voxel downsampling and spatial distribution features of the transmission lines to improve recognition capabilities. The method employs fast Euclidean distance for instance segmentation, while misclassified points outside the base of transmission towers are eliminated through contour extraction. Experimental results on real datasets validate that our proposed method not only fulfills the real-time measurement requirements but also surpasses the performance of existing algorithms</p>