集成技术2025，Vol.14Issue(4)：1-20,20.DOI:10.12146/j.issn.2095-3135.20241127002
三维高斯泼溅技术在场景重建中的研究现状与挑战
Research Status and Challenges of 3D Gaussian Splatting Technology in Scene Reconstruction
摘要
Abstract
3D scene reconstruction is a critical research topic in autonomous driving,robotics,and related fields,with extensive applications in navigation mapping,environmental interaction,and virtual/augmented reality tasks.Current deep learning-based 3D scene reconstruction methods can be primarily categorized into 5 groups from the perspectives of scene representation and core modeling techniques:cost volume-based depth estimation methods,truncated signed distance function-based voxel approaches,transformer architecture-based large-scale feedforward methods,multilayer perceptron-based neural radiance fields,and 3D Gaussian splatting.Each category exhibits unique strengths and limitations.The emerging 3D Gaussian splatting method distinguishes itself by explicitly representing scenes through Gaussian functions while achieving rapid scene rendering and novel view synthesis via efficient rasterization operations.3D Gaussian splatting diverges from the neural radiance fields-based scene representation paradigm.Its most significant advantage is that it ensures both efficient rendering and interpretable,editable scene modeling,thereby paving the way for accurate 3D scene reconstruction.However,3D Gaussian splatting still faces numerous challenges in practical scene reconstruction applications.Based on this analysis,this paper first provides a concise introduction to the fundamentals of 3D Gaussian splatting and conducts a comparative analysis with the aforementioned 4 categories.Following a systematic survey of existing 3D Gaussian splatting reconstruction algorithms,we summarize the key challenges addressed by these methods and review current research progress on core technical difficulties through representative case studies.Finally,we prospect potential future research directions worthy of exploration.关键词
深度估计/三维高斯泼溅/神经辐射场/三维场景重建Key words
depth estimation/3D Gaussian splatting/neural radiance field/3D scene reconstruction分类
信息技术与安全科学引用本文复制引用
朱东林,陈淼,毛宇岩,张峻豪,王忠立..三维高斯泼溅技术在场景重建中的研究现状与挑战[J].集成技术,2025,14(4):1-20,20.基金项目
国家科技创新2030—重大项目(2022ZD0205005) （2022ZD0205005）
北京交通大学基本科研业务费研究生创新项目(2023YJS142) This work is supported by National Science and Technology Innovation 2030 Major Project(2022ZD0205005),and Beijing Jiaotong University Fundamental Research Funds for the Central Universities(2023YJS142) （2023YJS142）
