Investigation and Experimental Verification of Substation Fast Modeling Method Based on Three-Dimension Laser Point Cloud Data

Abstract Due to the characteristics of large data volume, uneven distribution, and excessive background noise of substation point cloud data, substation modeling processing with traditional methods have problems such as noise residue, difficulty in equipment segmentation, and low model matching efficiency during the process of point cloud modeling in substations. Therefore, based on the modeling principle of point cloud data, an adaptive voxel filtering algorithm is proposed. This algorithm calculates the neighborhood density of each point as the judgment criterion, and processes point clouds in different density areas with different compression parameters to ensure high compression ratio while preserving low-density details inside the substation equipment. In addition, a fast-matching method based on image recognition was proposed to solve the problem of low model matching efficiency, with a recognition accuracy of 99%. Finally, the correctness and effectiveness of the proposed algorithm were verified through examples, confirming that it improves modeling speed while solving the processing of substation point cloud data.