RoofDiffusion: Constructing Roofs from Severely Corrupted Point Data via Diffusion
This work addresses a domain-specific problem for remote sensing and 3D building modeling, enabling better use of low-cost sensors and reducing UAV flight overlap, but it is incremental as it builds on existing diffusion and completion techniques.
The paper tackles the problem of accurately completing and denoising roof height maps from severely corrupted point data, such as sparse or occluded LiDAR scans, to improve 3D building reconstruction. It introduces RoofDiffusion, a self-supervised diffusion method that handles up to 99% point sparsity and 80% roof area occlusion, outperforming state-of-the-art methods on benchmarks and enhancing reconstruction with City3D.
Accurate completion and denoising of roof height maps are crucial to reconstructing high-quality 3D buildings. Repairing sparse points can enhance low-cost sensor use and reduce UAV flight overlap. RoofDiffusion is a new end-to-end self-supervised diffusion technique for robustly completing, in particular difficult, roof height maps. RoofDiffusion leverages widely-available curated footprints and can so handle up to 99\% point sparsity and 80\% roof area occlusion (regional incompleteness). A variant, No-FP RoofDiffusion, simultaneously predicts building footprints and heights. Both quantitatively outperform state-of-the-art unguided depth completion and representative inpainting methods for Digital Elevation Models (DEM), on both a roof-specific benchmark and the BuildingNet dataset. Qualitative assessments show the effectiveness of RoofDiffusion for datasets with real-world scans including AHN3, Dales3D, and USGS 3DEP LiDAR. Tested with the leading City3D algorithm, preprocessing height maps with RoofDiffusion noticeably improves 3D building reconstruction. RoofDiffusion is complemented by a new dataset of 13k complex roof geometries, focusing on long-tail issues in remote sensing; a novel simulation of tree occlusion; and a wide variety of large-area roof cut-outs for data augmentation and benchmarking.