Octree-Integrated Density-Aware Hierarchical Compression for Web-Optimized 3D Building Point Cloud Models

Facing the challenges in web visualization of large architectural point cloud datasets, this paper proposes a novel lightweight algorithm integrating octree spatial partition and density-aware hierarchical simplification. The method adopts octree-based spatial partition for hierarchical data segmentation, and improves data quality by uniform downsampling and statistical outlier removal. Thereafter, a hierarchical simplification based on structural features guided by density is implemented. Subsequently, LOD-based segmentation and quantization compression are conducted. Evaluation on datasets containing over 1 billion points shows that the approach reduces data volume by 92.8% and enhances web loading speed by 71.4% while retaining significant geometric accuracy. The lightweighting method provides a scalable solution for BIM and smart city Web3D applications, which is computationally efficient and storage/transmission friendly.