APC2Mesh: Bridging the gap from occluded building façades to full 3D models

This paper addresses the challenge of creating digital twins of urban buildings from airborne LiDAR data when building surfaces are partially hidden. The proposed APC2Mesh method combines point completion techniques with a neural network-based 3D mesh reconstruction approach. Complete points generated from occluded ones are used as input to a linearized skip attention-based deformation network for mesh generation. Testing across three different scenes demonstrates the approach effectively reconstructs buildings with varying architectural styles while managing noise and incomplete data coverage, offering improved fidelity for urban digital twin applications.

Recommended citation: Akwensi, P. H., Bharadwaj, A., & Wang, R. (2024). "APC2Mesh: Bridging the gap from occluded building façades to full 3D models." ISPRS Journal of Photogrammetry and Remote Sensing.
Download Paper