To address the problems of geometric gaps and seam color distortion that commonly occur in the transition zones during the integration of terrestrial realistic 3D models and underwater terrain point clouds due to scale inconsistency,texture differences,and geometric topological discontinuities,a water-land integrated realistic 3D data fusion technique is investigated.A method combining multi-scale geometric seamless processing and face-primitive-based Poisson color blending is adopted.Through boundary detection and adaptive meshing,geometric cracks in the water-land junction areas are eliminated,and abrupt color differences are reduced.In demonstration applications,the proposed method effectively resolves geometric incompleteness in the junction regions and achieves continuous and smooth color optimization across seams while preserving the gradient structural characteristics of the source imagery,resulting in a geometrically continuous and visually coherent water-land integrated realistic 3D model.The results demonstrate that the fusion strategy is suitable for high-quality integration of multi-source water-land realistic 3D data in complex navigation waterway environments,and can provide reliable technical support for the refined construction and application of the realistic 3D China initiative in the field of smart waterway.