DEGS: a dual-elastic grid system for efficient 3D urban spatial data management and computation

Managing heterogeneous 3D urban geospatial data requires compact representation and efficient computation. Existing DGGSs and uniform voxel structures suffer from scale rigidity and storage redundancy, while elastic grids often complicate indexing and topology-aware operations. This study proposes the Dual-Elastic Grid System (DEGS), which combines anisotropic partitioning with selected code-domain operations. DEGS uses a Morton-like interleaved code extended with dimension-specific separators and suffix-padding bits, so position and axis-specific scale information are embedded in one integer. Anchor-relative parity patterns further distinguish voxels, faces, edges, and vertices, while selected operators support scale extraction, element-type detection, and atomic adjacency updates without repeated coordinate reconstruction. Experiments on a representative Zhengzhou urban dataset show a 7.62 compression ratio relative to a standard octree representation for 2.5 m low-altitude free space. DEGS also improves 26-neighborhood traversal and online index-accelerated ray traversal at fine resolutions, although this benefit requires additional offline map construction. These results indicate that DEGS can support compact representation and repeated fine-resolution spatial queries in large-scale low-altitude urban environments.