GS-3I: Gaussian Splatting for Surface Reconstruction from Illumination-Inconsistent Images

2025-03-16

AI Summary

This work addresses geometric distortions in 3D Gaussian Splatting (3DGS) surface reconstruction caused by illumination inconsistencies—such as single-view underexposure and inter-view exposure variations. To tackle this, we propose a robust reconstruction framework comprising two key components: (1) a CNN-driven tone-mapping correction module that mitigates optimization bias induced by underexposed regions during Gaussian parameter learning; and (2) a normal compensation mechanism that jointly leverages single-view depth estimation and multi-view photometric consistency constraints to explicitly model and correct illumination-induced geometric inconsistencies in surface normals. Evaluated on complex real-world scenes, our method significantly improves reconstruction accuracy and robustness compared to existing illumination-sensitive approaches. It demonstrates strong practical utility for geometry-aware tasks essential in robotic autonomous exploration, where reliable 3D perception under varying lighting conditions is critical.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Machine Learning: Learning with Manifolds

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Accurate geometric surface reconstruction, providing essential environmental information for navigation and manipulation tasks, is critical for enabling robotic self-exploration and interaction. Recently, 3D Gaussian Splatting (3DGS) has gained significant attention in the field of surface reconstruction due to its impressive geometric quality and computational efficiency. While recent relevant advancements in novel view synthesis under inconsistent illumination using 3DGS have shown promise, the challenge of robust surface reconstruction under such conditions is still being explored. To address this challenge, we propose a method called GS-3I. Specifically, to mitigate 3D Gaussian optimization bias caused by underexposed regions in single-view images, based on Convolutional Neural Network (CNN), a tone mapping correction framework is introduced. Furthermore, inconsistent lighting across multi-view images, resulting from variations in camera settings and complex scene illumination, often leads to geometric constraint mismatches and deviations in the reconstructed surface. To overcome this, we propose a normal compensation mechanism that integrates reference normals extracted from single-view image with normals computed from multi-view observations to effectively constrain geometric inconsistencies. Extensive experimental evaluations demonstrate that GS-3I can achieve robust and accurate surface reconstruction across complex illumination scenarios, highlighting its effectiveness and versatility in this critical challenge. https://github.com/TFwang-9527/GS-3I

Problem

• Robust surface reconstruction from inconsistent illumination images.
• Mitigate 3D Gaussian optimization bias in underexposed regions.
• Compensate for lighting inconsistencies in multi-view image reconstruction.

Innovation

• CNN-based tone mapping correction framework
• Normal compensation mechanism for geometric inconsistencies
• Robust surface reconstruction under inconsistent illumination