BROWSE
Detail
Journal Article
An Efficient Distributed Pipeline for Large-Scale High-Resolution 3D Spatial Image Reconstruction
- Authors
- Sangwoo Ahn, Hanshin Lim, Hyun-Cheol Kim, Hyukmin Kwon, Hyon-Gon Choo
- Issue Date
- 2026-03
- Citation
- 방송공학회 논문지, v.31, no.2, pp.260-279
- ISSN
- 1226-7953
- Publisher
- 한국방송∙미디어공학회
- Language
- Korean
- Type
- Journal Article
- Abstract
- Three-dimensional (3D) spatial image services reconstruct realistic and high-resolution 3D environments from large-scale collections of two-dimensional (2D) images, with applications spanning the metaverse, cultural heritage preservation, and industrial monitoring. The generation of 3D content typically involves point cloud construction, mesh generation, and texture mapping, and has been the subject of extensive research. However, existing algorithms and processing pipelines often suffer from limited processing speed and image quality. To overcome these limitations, this paper presents an efficient pipeline for large-scale and high-resolution 3D spatial image reconstruction. The proposed system supports automated image grouping and distributed processing for dense point cloud generation, adaptive mesh reconstruction, and color mapping based on color continuity. This design enables faster processing of massive datasets while preserving or enhancing visual quality compared to conventional methods. Experimental results demonstrate that the proposed pipeline achieves superior image quality while significantly improving processing speed. This approach facilitates fast and accurate 3D data generation and is expected to be effectively utilized in various immersive services.
- Keyword
- Point cloud, mesh, Color mapping, Spatial image generation, 3D spatial image service
- KSP Keywords
- 3D content, 3D data, 3D environment, 3D spatial, Color Mapping, Conventional methods, Cultural heritage preservation, Data generation, Dense point cloud, Fast and accurate, High resolution
This work is distributed under the term of Creative Commons License (CCL)
(CC BY NC ND)
(CC BY NC ND)
