SpotlessGS: Relightable 3D Gaussian Splatting under Dynamic Illumination for Robotic Perception
Authors: Liang Hong, Jiaxin Wei, Simon Schaefer, Stefan Leutenegger, Jaehyung Jung
Organizations: Mobile Robotics Lab, School of Computation, Information and Technology, Technical University of Munich. · Mobile Robotics Lab, Department of Mechanical and Process Engineering, ETH Zurich.
Robots operating in dark or poorly lit environments rely on onboard lights, which often produce uneven illumination that degrades downstream perception tasks. Prior approaches based on 2D image enhancement lack reliable supervision and fail to preserve multi-view geometric consistency. To address these limitations, we extend Dark Gaussian Splatting (DarkGS) toward a more accurate and flexible relightable 3D reconstruction framework. First, we eliminate the need for explicit light parameter calibration by jointly optimizing lighting parameters within the Gaussian Splatting framework. Second, we introduce a low-frequency illumination model based on spherical harmonics (SH) to capture spatially varying residual and ambient lighting effects. Third, we incorporate an MLP-based Bidirectional Reflectance Distribution Function (BRDF) to model non-Lambertian reflectance. Experiments on synthetic and real-world datasets demonstrate that our method effectively mitigates illumination artifacts while improving rendering quality and quantitative performance over prior approaches. We further validate its benefits for robotic perception through a downstream task.
Figures & tables
Fig. 1 : Robotic platforms and representative onboard images under localized spotlight illumination. Onboard lighting causes over- and under-exposure, producing shadows and saturated highlights that degrade perception. Our method reconstructs a geometrically consistent 3D scene and relights it with uniform illumination, mitigating these artifacts while preserving structure. Robots and onboard images are from [ 1 , 2 , 3 ] .
Fig. 2 : Overview of the SpotlessGS framework.
Fig. 3 : Photo-realistic simulator in Blender for generating synthetic dataset.
Fig. 4 : Qualitative comparison on the synthetic (Uneven, Saturated) and self-collected (Self-col.) dataset.
Dataset
Method
PSNR ↑
SSIM ↑
LPIPS ↓
Synthetic Dataset
Vanilla 3DGS [ 19 ]
6.04
0.391
0.424
DarkGS [ 10 ]
10.66
0.685
0.460
Ours
12.67
0.727
0.381
Self-collected Dataset
Vanilla 3DGS [ 19 ]
7.26
0.333
0.630
DarkGS [ 10 ]
11.09
0.749
0.604
Ours
17.46
0.804
0.513
TABLE I : Quantitative Evaluation on Synthetic and Real-World Datasets.
Fig. 5 : Qualitative comparison on the DarkGS dataset [ 10 ] and real-world dataset in the wild [ 2 , 3 ] .
Fig. 6 : Downstream COLMAP reconstruction comparison. Left: Reconstruction from unevenly illuminated images. Right: Reconstruction from our relit images. Top: sparse point clouds with camera poses. Bottom: enlarged camera trajectories. Colors indicate temporal progression (blue → red). The relit images yield a smoother and physically plausible trajectory with more geometrically consistent point cloud.
Method
Reprojection Error ↓
Observations per Image ↑
Original (Uneven)
1.164 px
480.75
Vanilla 3DGS [ 19 ]
0.733 px
290.76
DarkGS [ 10 ]
0.611 px
881.78
Ours
0.463 px
960.28
TABLE II : Quantitative Comparison on COLMAP reconstruction.
Fig. 7 : Qualitative results of the ablation study on SH-based ambient illumination and the MLP-based BRDF.
Dataset
Method
PSNR ↑
SSIM ↑
LPIPS ↓
Syn. Dataset
DarkGS (Baseline) [ 10 ]
10.66
0.685
0.460
+ SH Ambient Light
12.62
0.724
0.381
+ MLP-based BRDF
10.85
0.721
0.403
+ SH + MLP (Ours)
12.67
0.727
0.381
Self-col. Dataset
DarkGS (Baseline) [ 10 ]
11.09
0.749
0.604
+ SH Ambient Light
15.96
0.768
0.524
TABLE III : Ablation study on the effect of SH-based ambient light and MLP-based BRDF on the synthetic (Syn.) and real-world self-collected (Self-col.) datasets.
Fig. 8 : Effect of Multi-view Rendering
Method
PSNR ↑
SSIM ↑
Tmp. PSNR ↑
Tmp. SSIM ↑
Cycle-GAN [ 9 ]
28.97
0.580
14.92
0.615
Ours
32.05
0.887
15.65
0.664
TABLE IV : Effect of multi-view rendering on synthetic dataset.
Novel-view synthesis and 3D reconstruction from sparse posed images are central to robotics and AR/VR. Yet, feed-forward 3D Gaussian reconstruction fails under lowlight due to noise, color shifts, and unreliable correspondence. We propose DelowlightSplat, a lowlight-aware feed-forward Gaussian splatting framework for clean novel-view rendering. We build a controllable multi-view lowlight benchmark by degrading only context views while keeping target views clean. We introduce a lightweight Lowlight Adapter for residual enhancement to improve matchability, and couple it with cost-volume-based multi-view inference to directly predict clean 3D Gaussians. Experiments show that DelowlightSplat significantly outperforms previous feed-forward method and two-stage pipeline under lowlight conditions.
Fuzhen Jiang, Zengtian Xie, Zhuoran Li
Hangzhou Dianzi University · Hangzhou, China · Zhuhai College of Science and Technology +1
Reconstructing 3D scenes under real-world low-light conditions remains challenging due to severe sensor noise, low signal-to-noise ratios, and degraded photometric consistency, which destabilize geometry estimation and novel view synthesis. Existing approaches often rely on well-lit reference data for reliable Structure-from-Motion (SfM) initialization under degraded inputs or apply per-view enhancement methods that introduce cross-view inconsistencies. To address these limitations, we propose \textbf{NOVA-GS}, a unified noise-aware framework for low-light 3D Gaussian Splatting that subsumes enhancement, denoising, and geometry optimization within a single process. Our method leverages VGGT-based feed-forward estimation to obtain robust camera poses and geometry directly from degraded inputs, eliminating the need for SfM. Building on this initialization, NOVA-GS integrates three coupled components: a structure-aware enhancement module for exposure correction, a self-supervised denoising module with blind-spot masking for pseudo-supervision, and a consistency-driven Gaussian Splatting optimization enforcing cross-view geometric coherence. We further introduce a noise-guided spherical harmonic regularization to suppress view-dependent artifacts in noisy regions. Extensive experiments on diverse real-world low-light datasets demonstrate improved geometric fidelity, color consistency, and robustness without requiring paired supervision or well-lit references. https://shaurya2524.github.io/nova-gs/
Shaurya Pavan A, Vemunuri Divya Madhuri, Yash Pradeep Gawande +1
Gaussian splatting (GS) has garnered significant attention in VR/AR and digital content creation due to its explicit parameterization and efficient rendering capabilities. However, existing GS-based methods for deformable objects face two key limitations: (i) illumination is erroneously baked into textures, causing physically inconsistent responses under dynamic deformations and lighting changes; (ii) snapshot-based reconstruction restricts post-reconstruction material editing. To address these challenges, we propose Deformable and Relightable GS (DR-GS), a unified Gaussian framework that integrates physically-based inverse rendering, relighting, and deformation-aware manipulation. Through explicitly disentangling geometry, illumination, and material representations, DR-GS overcomes the limitations of static snapshots, resolving unrealistic appearance under varying conditions while enabling post-reconstruction parameter editing. Extensive experiments show that DR-GS achieves leading visual quality across static reconstruction, dynamic deformation, and relighting, reliably preserving reflections and specular highlights on glossy surfaces. It further establishes a fully decoupled geometry-illumination-material pipeline, enabling high-quality 3D asset creation and comprehensive post-editing.
Jiaxin Li, Tong Wu, Yi Wei +2
Shanghai Innovation Institute, China · Zhejiang University, China · Department of Artificial Intelligence, Westlake University, China +2