EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes

Collaborative geometric reconstruction of building structures can significantly reduce communication consumption for data sharing, protect privacy, and provide support for large-scale robot application management. In recent years, geometric reconstruction of building structures has been partially st...

Full description

Saved in:
Bibliographic Details
Main Authors: Changyu Qian, Hanqiang Deng, Xiangrong Ni, Dong Wang, Bangqi Wei, Hao Chen, Jian Huang
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Applied Sciences
Subjects:
Online Access:https://www.mdpi.com/2076-3417/15/13/7133
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849320894827069440
author Changyu Qian
Hanqiang Deng
Xiangrong Ni
Dong Wang
Bangqi Wei
Hao Chen
Jian Huang
author_facet Changyu Qian
Hanqiang Deng
Xiangrong Ni
Dong Wang
Bangqi Wei
Hao Chen
Jian Huang
author_sort Changyu Qian
collection DOAJ
description Collaborative geometric reconstruction of building structures can significantly reduce communication consumption for data sharing, protect privacy, and provide support for large-scale robot application management. In recent years, geometric reconstruction of building structures has been partially studied, but there is a lack of alignment fusion studies for multi-UAV (Unmanned Aerial Vehicle)-reconstructed geometric structure models. The vertices and edges of geometric structure models are sparse, and existing methods face challenges such as low feature extraction efficiency and substantial data requirements when processing sparse graph structures after geometrization. To address these challenges, this paper proposes an efficient deep graph matching registration framework that effectively integrates interpretable feature extraction with network training. Specifically, we first extract multidimensional local properties of nodes by combining geometric features with complex network features. Next, we construct a lightweight graph neural network, named EKNet, to enhance feature representation capabilities, enabling improved performance in low-overlap registration scenarios. Finally, through feature matching and discrimination modules, we effectively eliminate incorrect pairings and enhance accuracy. Experiments demonstrate that the proposed method achieves a 27.28% improvement in registration speed compared to traditional GCN (Graph Convolutional Neural Networks) and an 80.66% increase in registration accuracy over the suboptimal method. The method exhibits strong robustness in registration for scenes with high noise and low overlap rates. Additionally, we construct a standardized geometric point cloud registration dataset.
format Article
id doaj-art-d6c16a34312e414fa6a87aea037f3a61
institution Kabale University
issn 2076-3417
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Applied Sciences
spelling doaj-art-d6c16a34312e414fa6a87aea037f3a612025-08-20T03:49:55ZengMDPI AGApplied Sciences2076-34172025-06-011513713310.3390/app15137133EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural ScenesChangyu Qian0Hanqiang Deng1Xiangrong Ni2Dong Wang3Bangqi Wei4Hao Chen5Jian Huang6College of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollege of Intelligence Science and Technology, National University of Defense Technology, Changsha 410073, ChinaCollaborative geometric reconstruction of building structures can significantly reduce communication consumption for data sharing, protect privacy, and provide support for large-scale robot application management. In recent years, geometric reconstruction of building structures has been partially studied, but there is a lack of alignment fusion studies for multi-UAV (Unmanned Aerial Vehicle)-reconstructed geometric structure models. The vertices and edges of geometric structure models are sparse, and existing methods face challenges such as low feature extraction efficiency and substantial data requirements when processing sparse graph structures after geometrization. To address these challenges, this paper proposes an efficient deep graph matching registration framework that effectively integrates interpretable feature extraction with network training. Specifically, we first extract multidimensional local properties of nodes by combining geometric features with complex network features. Next, we construct a lightweight graph neural network, named EKNet, to enhance feature representation capabilities, enabling improved performance in low-overlap registration scenarios. Finally, through feature matching and discrimination modules, we effectively eliminate incorrect pairings and enhance accuracy. Experiments demonstrate that the proposed method achieves a 27.28% improvement in registration speed compared to traditional GCN (Graph Convolutional Neural Networks) and an 80.66% increase in registration accuracy over the suboptimal method. The method exhibits strong robustness in registration for scenes with high noise and low overlap rates. Additionally, we construct a standardized geometric point cloud registration dataset.https://www.mdpi.com/2076-3417/15/13/7133three-dimensional reconstructiongraph neural networkgraph structure registration
spellingShingle Changyu Qian
Hanqiang Deng
Xiangrong Ni
Dong Wang
Bangqi Wei
Hao Chen
Jian Huang
EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
Applied Sciences
three-dimensional reconstruction
graph neural network
graph structure registration
title EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
title_full EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
title_fullStr EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
title_full_unstemmed EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
title_short EKNet: Graph Structure Feature Extraction and Registration for Collaborative 3D Reconstruction in Architectural Scenes
title_sort eknet graph structure feature extraction and registration for collaborative 3d reconstruction in architectural scenes
topic three-dimensional reconstruction
graph neural network
graph structure registration
url https://www.mdpi.com/2076-3417/15/13/7133
work_keys_str_mv AT changyuqian eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT hanqiangdeng eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT xiangrongni eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT dongwang eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT bangqiwei eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT haochen eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes
AT jianhuang eknetgraphstructurefeatureextractionandregistrationforcollaborative3dreconstructioninarchitecturalscenes