Grid Engineering Point Cloud Device Clustering Segmentation and Three-Dimensional Shape Retrieval Technology for Grid Digital Handover Method

The power grid infrastructure design model requires the transfer of digital design data to the power grid operation and maintenance, but the current digital transfer lacks intelligent components such as serial transfer data and automatic feedback correction for specific application scenarios. There are problems in the clustering and segmentation of single devices in the power grid engineering point cloud, as well as the three-dimensional shape description and retrieval of power grid devices. This paper studies the technology of point cloud target extraction and verification. Based on the characteristics of the 3D handover model, point cloud data is preprocessed using a direct pass filter. Point cloud filtering methods based on terrain filtering and 3D model filtering are used to achieve point cloud ground processing. An improved region growth method based on the boundary eigenvalue of the 3D handover model is used to segment the point cloud model, and a deep neural network is introduced, research on 3D key point extraction methods and feature descriptors for point clouds suitable for complex power equipment to improve the accuracy of apparent 3D shape matching and retrieval of power grid equipment. Through the research methods in this article, it is easier to segment and extract individual points from point clouds and achieve automatic association mapping of main equipment extraction and equipment point clouds with corresponding models in the model library.