Oblique photography point cloud building reconstruction method based on deep neural network

CN118967940AActive Publication Date: 2024-11-15UNIV OF ELECTRONICS SCI & TECH OF CHINA

Patent Information

Application Number
CN202411159678.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-22
Publication Date
2024-11-15
Estimated Expiration
2044-08-22

Smart Images

  • Figure CN118967940A_ABST
    Figure CN118967940A_ABST
Patent Text Reader

Abstract

The invention discloses an oblique photography point cloud building reconstruction method based on a deep neural network, and relates to the technical field of image data processing, and the method comprises the steps: constructing a single building point cloud data set, a building surface extraction network M1 and a loss function L; the M1 is trained to obtain a building surface extraction model M2; obtaining a point cloud set by using an RANSAC algorithm according to a to-be-reconstructed single building point cloud, and forming a set X1 by using all the point cloud sets; predicting the predicted SDF value of each point in the X1 by using the M2 to obtain a set X2; and performing surface reconstruction on the to-be-reconstructed single building based on the X2. The invention provides the integrated and full-automatic building surface extraction and reconstruction method, manual participation is not needed, and based on less point reconstruction, the reconstruction efficiency can be improved, and the number of planes can be reduced as much as possible, so that the building representation is more concise, the occupied storage space is smaller, the data volume is smaller, and the calculation complexity is reduced.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Point cloud reconstruction deep neural network training method, reconstruction method and controller

    CN114580621A

  • City building three-dimensional model monomer reconstruction method based on point cloud

    CN116310192A

  • Method for automatically detecting building structure and generating 3D model based on laser radar

    WO2019242174A1

Cited By

  • 3D reconstruction method based on real estate building cluster point cloud-to-MESH neural network

    CN120707769A

  • A 3D reconstruction method based on a real estate building cluster point cloud to MESH neural network

    CN120707769B