Large-Scale Scene Relocation via 3D Gaussian Splatting in Urban Enviroment

ZHANG Shoujian; LIU Haoze; LU Haitao; CHEN Haijier; XU Bo; ZHAO Lei

doi:10.13203/j.whugis20250276

您当前的位置：

首页 >

文章列表页 >

Large-Scale Scene Relocation via 3D Gaussian Splatting in Urban Enviroment

Geodesy and Survey Engineering | 更新时间：2026-11-05

- Large-Scale Scene Relocation via 3D Gaussian Splatting in Urban Enviroment
- Geomatics and Information Science of Wuhan University Vol. 51, Issue 4, Pages: 689-702(2026)
- 作者机构：
  
  1.武汉大学测绘学院，湖北武汉，430079
  2.武汉大学珞珈实验室，湖北武汉，430079
- 作者简介：
- 基金信息：
- DOI：10.13203/j.whugis20250276
  CLC： P228
- Received：20 September 2025，
  
  Published：05 April 2026
- 稿件说明：
移动端阅览
张守建,刘皓泽,陆海涛，等.基于3D高斯溅射的大规模场景重定位方法[J].武汉大学学报（信息科学版）,2026,51(4):689-702.

ZHANG Shoujian,LIU Haoze,LU Haitao,et al.Large-Scale Scene Relocation via 3D Gaussian Splatting in Urban Enviroment[J].Geomatics and Information Science of Wuhan University,2026,51(4):689-702.
张守建,刘皓泽,陆海涛，等.基于3D高斯溅射的大规模场景重定位方法[J].武汉大学学报（信息科学版）,2026,51(4):689-702. DOI：,10.13203/j.whugis20250276

ZHANG Shoujian,LIU Haoze,LU Haitao,et al.Large-Scale Scene Relocation via 3D Gaussian Splatting in Urban Enviroment[J].Geomatics and Information Science of Wuhan University,2026,51(4):689-702. DOI：,10.13203/j.whugis20250276

摘要

为了获取复杂环境下精确位姿信息，提出一种利用3D高斯溅射技术（3D Gaussian splatting，3DGS）的大规模重定位框架，使用单目相机实现厘米级定位。首先，利用多传感器数据构建城市复杂环境高精度3DGS地图，然后，利用基于深度学习的特征点提取与匹配方法和的视觉前端结合，实现了高效、鲁棒的特征点提取和匹配，最后通过求解3D到2D点对的方法，迭代渲染优化得到精确的位姿信息。在开源街景数据集KITTI上选取3种场景道路对所提方法进行了测试，实验结果表明，在城镇道路、林荫大道和交通密集型公路上的平均定位精度分别为0.026 m、0.029 m和0.081 m，优于其他常规方法。

Abstract

Objectives

In the field of autonomous driving

acquiring high-precision pose information is a critical factor in ensuring its safety and reliability. Although global navigation satellite system (GNSS) can provide absolute pose references

its positioning accuracy significantly degrades in complex scenarios due to signal obstruction and multipath effects

making it difficult to meet the stringent requirements of autonomous driving systems. To address this challenge

we have conducted systematic exploration into precise pose acquisition in large-scale

complex

and dynamic environments.

Methods

First

based on multi-sensor fusion technology

we developed a reconstruction system tailored for dynamic urban environments

focusing on three representative large-scale outdoor scenarios: urban roads

tree-lined avenues

and high-traffic highways. A high-precision 3D Gaussian splatter (3DGS) map was constructed for these scenarios. In the visual front-end

we integrated the SuperPoint feature extraction algorithm with the SuperGlue feature matching algorithm to achieve high-quality feature point detection and matching. Additionally

we introduced the perspective-

-point (P

P) algorithm

which estimates the initial camera pose by matching feature points with corresponding points in the 3D map.To optimize positioning results and improve the system's real-time performance

we proposed an iterative optimization strategy based on progressive rendering. This strategy utilizes the geometric structure and radiometric field characteristics of the 3DGS map

combining step-by-step rendering and optimization to continuously adjust the camera pose estimation

ultimately achieving progressive optimization of real-time high-precision positioning results.

Results

The experimental validation further demonstrates the following results: (1) In dynamic urban environments

the reconstruction system exhibits exceptional performance

effectively rendering continuous and realistic lighting and texture effects in real-time. Particularly in complex urban settings

the system handles issues such as lighting changes and occlusions caused by dynamic objects effectively. (2) The feature extraction and matching algorithms

SuperPoint and SuperGlue

significantly outperform traditional methods in terms of both the quantity and quality of feature points. Compared to traditional methods

the number of extracted feature points has been greatly increased

with a feature point distribution rate reaching 70%

more than twice that of conventional methods. (3) The system demonstrates excellent positioning accuracy in three typical outdoor scenarios. On urban roads

tree-lined avenues

and high-traffic highways

the system achieves average positioning accuracies of 0.026 m

0.029 m

and 0.081 m

respectively

which are significantly better than other representative methods. Additionally

the system achieves a frequency of 96% for absolute displacement and rotation errors smaller than 10 cm and 1°

respectively

indicating high stability in positioning accuracy and the ability to provide reliable real-time positioning results in various environments. (4) Ablation experiments further validate the effectiveness of the system's design. The complete model achieves the best results in both typical scenarios

with average absolute translation and rotation errors of 3.22 cm and 0.69°

respectively. Compared to experiments where dynamic occlusion handling

SuperPoint feature extraction

and iterative optimization strategies were removed

the system's overall design demonstrates superior performance. (5) The system's real-time performance has been verified. Time analysis shows that the system processes each frame of images in approximately 1 s

which meets the real-time requirements and is suitable for real-time application scenarios.

Conclusions

In large-scale outdoor environments

the visual relocalization technology based on the 3DGS map enables high-precision location information retrieval

effectively compensating for the limitations of existing positioning methods under GNSS signal occlusion or in complex environments. Through multi-sensor fusion and precise map construction

the system performs exceptionally well in dynamic urban scenarios. Particularly in complex urban environments

the issues of lighting changes and dynamic object occlusion are effectively handled

significantly enhancing the system's robustness and adaptability. The front-end feature extraction and matching methods greatly improve the system's stability and accuracy. In complex dynamic scenes

the distribution of feature points is more uniform

further enhancing positioning precision. Compared to various methods

the system demonstrates optimal performance. Ablation experiments further validate the necessity and effectiveness of these techniques. The system's processing speed meets real-time requirements

ensuring efficient and reliable real-time localization.

关键词

Keywords

references

张小红，陶贤露，王颖喆，等 . 城市场景智能手机GNSS/MEMS融合车载高精度定位［J］. 武汉大学学报（信息科学版）， 2022 ， 47 （ 10 ）： 1740 - 1749 .

ZHANG Xiaohong ， TAO Xianlu ， WANG Yingzhe ， et al . MEMS-Enhanced Smartphone GNSS High-Precision Positioning for Vehicular Navigation in Urban Conditions ［J］. Geomatics and Information Science of Wuhan University ， 2022 ， 47 （ 10 ）： 1740 - 1749 .

KUUTTI S ， FALLAH S ， KATSAROS K ， et al . A Survey of the State-of-the-Art Localization Techniques and Their Potentials for Autonomous Vehicle Applications ［J］. IEEE Internet of Things Journal ， 2018 ， 5 （ 2 ）： 829 - 846 .

LU Y Q ， MA H J ， SMART E ， et al . Real-Time Performance-Focused Localization Techniques for Autonomous Vehicle： A Review ［J］. IEEE Transactions on Intelligent Transportation Systems ， 2022 ， 23 （ 7 ）： 6082 - 6100 .

崔晓珍，周琪，武东杰，等 . GNSS/IMU与里程计紧-松耦合的因子图融合定位方法［J］. 武汉大学学报（信息科学版）， 2024 ， 49 （ 10 ）： 1911 - 1921 .

CUI Xiaozhen ， ZHOU Qi ， WU Dongjie ， et al . Factor Graph Fusion Localization Method with Tight and Loose Coupling of GNSS/IMU and Odometry ［J］. Geomatics and Information Science of Wuhan University ， 2024 ， 49 （ 10 ）： 1911 - 1921 .

李磊磊，梁琳，钟傲，等 . 基于GPS的先验地图辅助视觉惯性定位方法［J］. 中国惯性技术学报， 2023 ， 31 （ 1 ）： 53 - 60 .

LI Leilei ， LIANG Lin ， ZHONG Ao ， et al . Visual-Inertial Positioning Method Based on GPS Prior Map Assisted ［J］. Journal of Chinese Inertial Technology ， 2023 ， 31 （ 1 ）： 53 - 60 .

BRACHMANN E ， KRULL A ， NOWOZIN S ， et al . DSAC Differentiable RANSAC for Camera Localization ［C］// IEEE Conference on Computer Vision and Pattern Recognition ， Honolulu， HI ， 2017 .

LASKAR Z ， MELEKHOV I ， KALIA S ， et al . Camera Relocalization by Computing Pairwise Relative Poses Using Convolutional Neural Network ［C］// IEEE International Conference on Computer Vision Workshops ， Venice， Italy ， 2018 .

SARLIN P E ， UNAGAR A ， LARSSON M ， et al . Back to the Feature： Learning Robust Camera Localization from Pixels to Pose ［C］//IEEE/CVF Conferen‑ce on Computer Vision and Pattern Recognition， Nashville， TN， USA ， 2021 .

MILDENHALL B ， SRINIVASAN P P ， TANCIK M ， et al . NeRF： Representing Scenes as Neural Radiance Fields for View Synthesis ［C］// Computer Vision ECCV ， Glasgow， UK ， 2020 .

YEN-CHEN L ， FLORENCE P ， BARRON J T ， et al . INeRF： Inverting Neural Radiance Fields for Pose Estimation ［C］// IEEE/RSJ International Conference on Intelligent Robots and Systems ， Prague， Czech Republic ， 2021 .

MAGGIO D ， ABATE M ， SHI J N ， et al . Loc-NeRF： Monte Carlo Localization Using Neural Radiance Fields ［PP/OL］. arXiv （ 2022-09-19 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2209.09050 https://arxiv.org/abs/2209.09050 .

赵文晔，高井祥，李增科，等 . 地图匹配辅助的KF-PF室内定位算法模型［J］. 武汉大学学报（信息科学版）， 2018 ， 43 （ 5 ）： 806 - 812 .

ZHAO Wenye ， GAO Jingxiang ， LI Zengke ， et al . An Indoor Positioning System Based on Map-Aided KF-PF Module ［J］. Geomatics and Information Science of Wuhan University ， 2018 ， 43 （ 5 ）： 806 - 812 .

MATSUKI H ， MURAI R ， KELLY P H J ， et al . Gaussian Splatting SLAM ［C］//IEEE/CVF Conferen‑ce on Computer Vision and Pattern Recognition ， Seattle， WA， USA ， 2024 .

SUN Y ， WANG X ， ZHANG Y F ， et al . IComMa： Inverting 3D Gaussian Splatting for Camera Pose Estimation via Comparing and Matching ［PP/OL］. arXiv （ 2023-12-14 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2312.09031 https://arxiv.org/abs/2312.09031 .

ZHAI H J ， ZHANG X Y ， ZHAO B M ， et al . SplatLoc： 3D Gaussian Splatting-Based Visual Localization for Augmented Reality ［J］. IEEE Transactions on Visualization and Computer Graphics ， 2025 ， 31 （ 5 ）： 3591 - 3601 .

XIN Z ， DAI C K ， LI Y ， et al . GauLoc： 3D Gaussian Splatting-Based Camera Relocalization ［J］. Computer Graphics Forum ， 2024 ， 43 （ 7 ）： e15256 .

BOTASHEV K ， PYATOV V ， FERRER G ， et al . GSLoc： Visual Localization with 3D Gaussian Splatting ［C］// IEEE/RSJ International Conference on Intelligent Robots and Systems ， Abu Dhabi， United Arab Emirates ， 2024 .

ZELLER A J ， WU H J . GSplatLoc： Ultra-Precise Camera Localization via 3D Gaussian Splatting ［PP/OL］. arXiv （ 2024-12-28 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2412.20056 https://arxiv.org/abs/2412.20056 .

CHEN Z Y ， YANG J W ， HUANG J H ， et al . OmniRe： Omni Urban Scene Reconstruction ［PP/OL］. arXiv （ 2024-08-29 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2408.16760 https://arxiv.org/abs/2408.16760 .

CHEN Y R ， GU C ， JIANG J Z ， et al . Periodic Vibration Gaussian： Dynamic Urban Scene Reconstruction and Real-Time Rendering ［PP/OL］. arXiv （ 2023-11-30 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2311.18561 https://arxiv.org/abs/2311.18561 .

YANG J W ， IVANOVIC B ， LITANY O ， et al . EmerNeRF： Emergent Spatial-Temporal Scene Decomposition via Self-Supervision ［PP/OL］. arXiv （ 2023-11-03 ）［ 2025-08-20 ］. https：//arxiv.org/abs/2311.02077 https://arxiv.org/abs/2311.02077 .

王铉彬，李星星，廖健驰，等 . 基于图优化的紧耦合双目视觉/惯性/激光雷达SLAM方法［J］. 测绘学报， 2022 ， 51 （ 8 ）： 1744 - 1756 .

WANG Xuanbin ， LI Xingxing ， LIAO Jianchi ， et al . Tightly-Coupled Stereo Visual-Inertial-LiDAR SLAM Based on Graph Optimization ［J］. Acta Geodaetica et Cartographica Sinica ， 2022 ， 51 （ 8 ）： 1744 - 1756 .

NOBLE F K . Comparison of OpenCV’s Feature Detectors and Feature Matchers ［C］// The 23rd International Conference on Mechatronics and Machine Vision in Practice ， Nanjing， China ， 2017 .

DETONE D ， MALISIEWICZ T ， RABINOVICH A . SuperPoint： Self-Supervised Interest Point Detection and Description ［C］// IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops ， Salt Lake City， UT， USA， 2018 .

SARLIN P E ， DETONE D ， MALISIEWICZ T ， et al . SuperGlue： Learning Feature Matching with Graph Neural Networks ［C］//IEEE/CVF Conferen‑ce on Computer Vision and Pattern Recognition，Seattle， WA， USA ， 2020 .

SHI Y ， CAI J X ， SHAVIT Y ， et al . ClusterGNN： Cluster-Based Coarse-to-Fine Graph Neural Network for Efficient Feature Matching ［C］// IEEE/CVF Conference on Computer Vision and Pattern Recognition ， New Orleans， LA， USA， 2022 .

MA J Y ， JIANG X Y ， FAN A X ， et al . Image Matching from Handcrafted to Deep Features： A Survey ［J］. International Journal of Computer Vision ， 2021 ， 129 （ 1 ）： 23 - 79 .

WANG A ， PRUKSACHATKUN Y ， NANGIA N ， et al . SuperGLUE： A Stickier Benchmark for General-Purpose Language Understanding Systems ［PP/OL］. ［ 2025-08-20 ］. https：//arxiv.org/abs/1905.00537 https://arxiv.org/abs/1905.00537 .

叶沅鑫，单杰，熊金鑫，等 . 一种结合SIFT和边缘信息的多源遥感影像匹配方法［J］. 武汉大学学报（信息科学版）， 2013 ， 38 （ 10 ）： 1148 - 1151 .

YE Yuanxin ， SHAN Jie ， XIONG Jinxin ， et al . A Matching Method Combining SIFT and Edge Information for Multi-Source Remote Sensing Images ［J］. Geomatics and Information Science of Wuhan University ， 2013 ， 38 （ 10 ）： 1148 - 1151 .

BESL P J ， MCKAY N D . A Method for Registration of 3-D Shapes ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence ， 1992 ， 14 （ 2 ）： 239 - 256 .

WANG J Y ， RUPPRECHT C ， NOVOTNY D . PoseDiffusion： Solving Pose Estimation via Diffusion-Aided Bundle Adjustment ［C］// IEEE/CVF International Conference on Computer Vision ， Paris， France ， 2024 .

SATTLER T ， LEIBE B ， KOBBELT L . Efficient & Effective Prioritized Matching for Large-Scale Image-Based Localization ［J］. IEEE Transactions on Pattern Analysis and Machine Intelligence ， 2017 ， 39 （ 9 ）： 1744 - 1756 .

BRACHMANN E ， ROTHER C . Neural-Guided RANSAC： Learning Where to Sample Model Hypotheses ［C］// IEEE/CVF International Conference on Computer Vision ， Seoul， Korea ， 2019 .

BRACHMANN E ， ROTHER C . Learning Less Is More：6D Camera Localization via 3D Surface Regression ［C］// IEEE/CVF Conference on Computer Vision and Pattern Recognition ， Salt Lake City， UT， USA， 2018 .

JIANG P ， BENDAPUDI N ， SARIPALLI S ， et al . 3DGS-ReLoc： 3D Gaussian Splatting for Map Representation and Visual Localization ［J］. Journal of Autonomous Vehicles and Systems ， 2025 ， 5 （ 3 ）： 031004 .

ZHANG S J ， LIU Z A ， XU B ， et al . Fusion GNSS/INS/Vision with Path Planning Prior for High Precision Navigation in Complex Environment ［J］. IEEE Sensors Journal ， 2025 ， 25 （ 5 ）： 9045 - 9055 .

XU B ， ZHANG S J ， KUANG K F ， et al . A Unified Cycle-Slip， Multipath Estimation， Detection and Mitigation Method for VIO-Aided PPP in Urban Environments ［J］. GPS Solutions ， 2023 ， 27 （ 2 ）： 59 .

WANG J R ， LIU J N ， ZHANG S J ， et al . Sky-View Images Aided NLOS Detection and Suppression for Tightly Coupled GNSS/INS System in Urban Canyon Areas ［J］. Measurement Science and Technology ， 2024 ， 35 （ 2 ）： 025112 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

High-Resolution L-Band InSAR Study of Intense Near-Field Deformation Associated with Shallow Moderate Earthquakes

Geometric and Frequency-Domain Constrained Single-Step Diffusion for Seabed Terrain Super-Resolution Reconstruction

Autonomous Robotic Mapping for Complex Spaces： Progress， Challenges and Future Perspectives

Three-State Fusion Support Problem of Marine Geographic Information for Ship Navigation

A Review of GNSS Intentional Interference Detection Technologies for Mass-Market Intelligent Terminals： Research Status and Prospects

Related Author

No data

Related Institution

No data

AI问答

⁰