Robust and High-Fidelity 3D Gaussian Splatting: Fusing Pose Priors and Geometry Constraints for Texture-Deficient Outdoor Scenes

TL;DR

This paper enhances 3D Gaussian Splatting for outdoor scenes by integrating LiDAR-IMU pose priors and geometric constraints, resulting in more robust, accurate, and high-fidelity scene reconstructions especially in texture-deficient environments.

Contribution

It introduces a novel approach combining pose priors and geometric constraints into 3D Gaussian Splatting, improving robustness and quality in outdoor scene reconstruction.

Findings

Pose optimization is three times faster with maintained accuracy.

Scene representation significantly outperforms conventional pipelines.

Enhanced visualization in weak or repetitive texture environments.

Abstract

3D Gaussian Splatting (3DGS) has emerged as a key rendering pipeline for digital asset creation due to its balance between efficiency and visual quality. To address the issues of unstable pose estimation and scene representation distortion caused by geometric texture inconsistency in large outdoor scenes with weak or repetitive textures, we approach the problem from two aspects: pose estimation and scene representation. For pose estimation, we leverage LiDAR-IMU Odometry to provide prior poses for cameras in large-scale environments. These prior pose constraints are incorporated into COLMAP's triangulation process, with pose optimization performed via bundle adjustment. Ensuring consistency between pixel data association and prior poses helps maintain both robustness and accuracy. For scene representation, we introduce normal vector constraints and effective rank regularization to enforce consistency in the direction and shape of Gaussian primitives. These constraints are jointly optimized with the existing photometric loss to enhance the map quality. We evaluate our approach using both public and self-collected datasets. In terms of pose optimization, our method requires only one-third of the time while maintaining accuracy and robustness across both datasets. In terms of scene representation, the results show that our method significantly outperforms conventional 3DGS pipelines. Notably, on self-collected datasets characterized by weak or repetitive textures, our approach demonstrates enhanced visualization capabilities and achieves superior overall performance. Codes and data will be publicly available at https://github.com/justinyeah/normal_shape.git.