Research map / 3D vision / Gaussian splatting
Reconstructing Satellites in 3D From Amateur Telescope Images
Zhiming Chang, Boyang Liu, Yifei Xia, Youming Guo, Boxin Shi, He Sun
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add controlled Gaussian growth regulation (ablation) | full pipeline · Simulation 1 | SSIM: 0.87 -> 0.90 | Improved |
| add BnB pose refinement (ablation) | full pipeline · Simulation 1 | PSNR: 22.51 -> 25.33 | Improved |
| add BnB pose refinement after initialization (ablation) | pose pipeline · synthetic pose estimation | ATE: 2.61 -> 2.43 | Improved |
| add point cloud filtering (ablation) | full pipeline · Simulation 1 | SSIM: 0.62 -> 0.90 | Improved |
| add image pre-processing (ablation) | full pipeline · Simulation 1 | PSNR: 20.87 -> 25.33 | Improved |
| add modified SfM init, controlled growth, BnB and filtering to original GS (ablation) | Gaussian Splatting · Simulation 1 | PSNR: 20.96 -> 25.33 | Improved |
| add VRT denoising after stacking and sharpening (ablation) | pre-processing pipeline · Simulation 1 | PSNR: 27.60 -> 27.74 | Improved |
| add wavelet sharpening to stacking (ablation) | pre-processing pipeline · Simulation 1 | PSNR: 27.00 -> 27.60 | Improved |
| add SelfDeblur to stacking (ablation) | pre-processing pipeline · Simulation 1 | PSNR: 27.00 -> 23.89 | Hurt |
| constrain NeRF sampling region (BARF Con.) (ablation) | BARF · Simulation 1 | PSNR: 17.64 -> 24.20 | Improved |
| use our modified SfM initialization instead of Dust3r | pose initialization · synthetic pose estimation | ATE: 138.81 -> 2.61 | Improved |
| GS with BnB instead of NeRF-- or BARF for reconstruction | reconstruction · Simulation 1 | PSNR: 16.09 -> 25.32 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- BARF — BARF: Bundle-Adjusting Neural Radiance Fields (ICCV 2021) · better on 5 of 5 shared results
- Dust3r — DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Gaussian Splatting for Efficient Satellite Image Photogrammetry (CVPR 2025 · similarity 0.86)
- RPC-GS: Gaussian Splatting with native RPC Rendering for Satellite Imagery (NeurIPS 2026 · similarity 0.86)
- COSMIC: Compress Satellite Image Efficiently via Diffusion Compensation (NeurIPS 2024 · similarity 0.84)
- Event-3DGS: Event-based 3D Reconstruction Using 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.83)
- Sparse2DGS: Geometry-Prioritized Gaussian Splatting for Surface Reconstruction from Sparse Views (CVPR 2025 · similarity 0.83)
- Satellite to GroundScape - Large-scale Consistent Ground View Generation from Satellite Views (CVPR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
