PolyGNN: Polyhedron-based Graph Neural Network for 3D Building Reconstruction from Point Clouds

Abstract

We present PolyGNN, a polyhedron-based graph neural network for 3D building reconstruction from point clouds. PolyGNN learns to assemble primitives obtained by polyhedral decomposition via graph node classification, achieving a watertight and compact reconstruction. To effectively represent arbitrary-shaped polyhedra in the neural network, we propose a skeleton-based sampling strategy to generate polyhedron-wise queries. These queries are then incorporated with inter-polyhedron adjacency to enhance the classification. PolyGNN is end-to-end optimizable and is designed to accommodate variable-size input points, polyhedra, and queries with an index-driven batching technique. To address the abstraction gap between existing city-building models and the underlying instances, and provide a fair evaluation of the proposed method, we develop our method on a large-scale synthetic dataset with well-defined ground truths of polyhedral labels. We further conduct a transferability analysis across cities and on real-world point clouds. Both qualitative and quantitative results demonstrate the effectiveness of our method, particularly its efficiency for large-scale reconstructions. The source code and data are available at https://github.com/chenzhaiyu/polygnn.

Accepted for publication in ISPRS Journal of Photogrammetry and Remote Sensing

References in corpus (7)

• PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
• Topology Adaptive Graph Convolutional Networks
• Supervised Fitting of Geometric Primitives to 3D Point Clouds
• Virtual laser scanning with HELIOS++: A novel take on ray tracing-based simulation of topographic 3D laser scanning
• City3D: Large-Scale Building Reconstruction from Airborne LiDAR Point Clouds
• Reconstructing Compact Building Models from Point Clouds Using Deep Implicit Fields
• Neural Vector Fields for Implicit Surface Representation and Inference