A Multi-Resolution Pipeline for High-Fidelity Coastal Terrain Reconstruction for XR From Heterogeneous Geospatial Data

The reconstruction of digital heritage environments increasingly demands the coherent integration of heterogeneous spatial data as a foundation for visually immersive XR applications. However, existing terrain-generation pipelines are typically not designed to support the levels of geometric detail, continuity, and scale consistency required for high-fidelity first-person exploration in real-time XR. This paper presents a reproducible processing pipeline for constructing a physically and perceptually coherent coastal virtual environment from heterogeneous geospatial datasets of varying resolution and origin. The pipeline integrates high-resolution aerial point clouds, lower-resolution satellitebased elevation data, and global bathymetric data into a continuous, multi-resolution terrain suitable for real-time XR applications, such as high detail and resolution for first-person exploration. We describe data cleanup, multi-resolution fusion, procedural enhancement of underwater terrain, terrain texturing, vegetation generation, and background simplification. The resulting virtual coastal environment serves as a multimodal-ready spatial substrate that can support synchronized visual, acoustic, and physical simulations. We demonstrate the pipeline through a case study of a critical segment of the Caribbean coast of Costa Rica and discuss its implications for multisensory digital heritage reconstruction and environmental communication.