Power pipeline point cloud semantic recognition and three-dimensional modeling method based on PointNet + +

By combining the improved PointNet++ network with prior geometric knowledge of power engineering, the entire process of 3D modeling of power pipelines is automated, solving the problems of low automation, poor accuracy and insufficient scene adaptability in existing technologies, and improving modeling efficiency and accuracy.

CN122473356APending Publication Date: 2026-07-28SHANGHAI MINGCHUAN SURVEYING & MAPPING TECH CO LTD
4 Cites 0 Cited by

Patent Information

Application Number
CN202610679975.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-05-18
Publication Date
2026-07-28

AI Technical Summary

Technical Problem

Existing technologies cannot simultaneously meet the requirements of high automation, high precision, strong robustness and wide scene adaptability for 3D modeling of power pipelines, especially in complex scenarios where it is difficult to achieve accurate semantic segmentation and parametric modeling of multiple components.

Method used

By employing an improved PointNet++ deep learning network combined with prior knowledge of power engineering geometry, and through a fully automated approach encompassing point cloud preprocessing, semantic recognition, parametric fitting, and topology remodeling, we achieve semantic segmentation and parametric modeling of multiple components in power pipelines.

Benefits of technology

It significantly improves the automation and accuracy of 3D modeling of power pipelines, enhances the adaptability to complex scenarios, meets the high-precision requirements of power engineering, and supports the processing of multiple data sources.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an electric power pipeline point cloud semantic recognition and three-dimensional modeling method based on PointNet + +, and belongs to the technical field of crossing of three-dimensional laser point cloud data processing, deep learning semantic segmentation and electric power engineering three-dimensional modeling. According to the method, a full-process automatic processing system from original point cloud input to complete three-dimensional model output is built, and modeling is completed through four links of point cloud preprocessing, semantic recognition, parametric fitting and topology reconstruction. In the point cloud preprocessing link, statistical filtering and radius filtering are combined for denoising, FPFH feature matching and ICP precise registration are adopted, and a voxel down-sampling technology is adopted; and generating a standardized point cloud data set. According to the method, the problems of low automation degree, poor precision and weak scene adaptability of three-dimensional modeling of the power pipeline in the prior art are solved, manual intervention can be reduced by 80% or above, the average semantic segmentation precision of the component reaches 95.2%, various power pipeline scenes such as a transformer substation, a cable pipe gallery and a power transmission line are comprehensively adapted, and the construction efficiency is improved. And technical support is provided for full-life-cycle digital management of power pipelines.
Need to check novelty before this filing date? Find Prior Art