An Algorithm Based on Image and Point Cloud Data Fusion for Detection of Transmission Lines External Force Damage Risk

To address potential external force damage to transmission lines caused by mechanical operations of engineering vehicles, this paper proposes a LiDAR-based distance calculation algorithm between intruders and transmission lines, integrating camera images for intruders' proximity labeling and displaying. Initially, LiDAR acquires point cloud data, which undergoes statistical filtering to remove noise points. Subsequently, transmission line point clouds are extracted using elevation filtering and Euclidean clustering. Equation fitting and parameter solving based on the least squares method extend the power line point cloud length for distance calculation. YOLOv8 is utilized to detect engineering vehicles in camera images, providing pixel coordinates and category information. Data fusion of point cloud and image is achieved using Zhang's calibration method, obtaining coordinate transformation between them. The engineering vehicle's point cloud channel is derived from its pixel coordinates. Finally, the traversal method calculates the nearest distance between the engineering vehicles and transmission lines, with detection results displayed on both image and point cloud data. Experimental results across various scenarios validate the algorithm's effectiveness and robustness.