A Multi-Scale 3D Terrain Modeling and Visualization Method for Radiating Tidal Sand Ridges Based on DEM

Abstract. To address challenges in modeling radiating tidal sand ridge groups—such as weakened feature representation, insufficient semantic expression, and low visualization efficiency—this study proposes a "feature fusion–semantic constraint–spatial clustering" framework based on DEM data. A dual-window dynamic threshold model integrates a 9×9 relief window with a 3×3 coefficient of variation, enabling adaptive extraction of terrain features. Semantic constraints are introduced by using the D8 algorithm to extract tidal channels and constructing a constrained TIN, improving feature representation by 8.1% over the standard VIP method. A semantic-aware simplification mechanism assigns higher protection to major channels (λ = 10) and applies dynamic tolerance elsewhere, preserving morphological integrity at 96.86%. Furthermore, terrain zones are classified via landscape indices into core, transition, and background areas, supporting a three-level LOD strategy that enhances rendering efficiency while maintaining geometric fidelity in key regions. Cognitive evaluations confirm the method’s superiority in visual clarity and structural accuracy, providing an effective solution for coastal terrain analysis and planning.