UniGaussian: Driving Scene Reconstruction from Multiple Camera Models via Unified Gaussian Representations

TL;DR

UniGaussian introduces a unified 3D Gaussian representation for urban scene reconstruction that effectively integrates multiple camera models, including fisheye and pinhole, enabling real-time, high-quality rendering for autonomous driving simulation.

Contribution

The paper presents a differentiable rendering method tailored for fisheye cameras and a framework that learns a unified Gaussian model from diverse sensor modalities, addressing a key gap in scene reconstruction.

Findings

Achieves superior rendering quality in driving scene simulation.

Maintains real-time rendering speed.

Successfully models multiple sensors and modalities.

Abstract

Urban scene reconstruction is crucial for real-world autonomous driving simulators. Although existing methods have achieved photorealistic reconstruction, they mostly focus on pinhole cameras and neglect fisheye cameras. In fact, how to effectively simulate fisheye cameras in driving scene remains an unsolved problem. In this work, we propose UniGaussian, a novel approach that learns a unified 3D Gaussian representation from multiple camera models for urban scene reconstruction in autonomous driving. Our contributions are two-fold. First, we propose a new differentiable rendering method that distorts 3D Gaussians using a series of affine transformations tailored to fisheye camera models. This addresses the compatibility issue of 3D Gaussian splatting with fisheye cameras, which is hindered by light ray distortion caused by lenses or mirrors. Besides, our method maintains real-time rendering while ensuring differentiability. Second, built on the differentiable rendering method, we design a new framework that learns a unified Gaussian representation from multiple camera models. By applying affine transformations to adapt different camera models and regularizing the shared Gaussians with supervision from different modalities, our framework learns a unified 3D Gaussian representation with input data from multiple sources and achieves holistic driving scene understanding. As a result, our approach models multiple sensors (pinhole and fisheye cameras) and modalities (depth, semantic, normal and LiDAR point clouds). Our experiments show that our method achieves superior rendering quality and fast rendering speed for driving scene simulation.