发表机构
School of Electronics and Information Engineering, Harbin Institute of Technology; th Research Institute, China Electronics Technology Group Corporation; State Key Laboratory of Comprehensive PNT Network and Equipment Technology; School of Electrical and Electronic Engineering, University of Manchester（哈尔滨工业大学电子与信息工程学院; 中国电子科技集团公司第五十四研究所; 综合PNT网络与装备技术国家重点实验室; 曼彻斯特大学电气与电子工程学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
GeoRay针对卫星三维重建问题，通过射线一致适配器、显式基准机制和逆方差融合技术，在绝对大地坐标系下实现了高精度、高覆盖率的前馈卫星三维重建，性能优于现有基线。
AI 中文摘要
前馈三维基础模型可一次重建透视场景。卫星摄影测量需要一种不同的产物，仅领域自适应无法提供：非中心有理多项式相机（RPC）下绝对大地坐标系中的密集表面高度。透视预训练特征无法沿RPC高度射线可靠观测，绝对高程包含与传感器偏差一阶可交换的低阶高程-基准量规，单目和多视图线索在不同区域失效。GeoRay（\texttt{\backslash method}）解决这三个问题：轻量型射线一致适配器使冻结骨干网络可沿原生RPC射线匹配；显式基准机制将地形与绝对高程分离，且构造上对垂直原点等变，因此一个训练好的模型可支持零控制、单控制和稀疏控制推理；校准的逆方差融合结合两个地形流。我们的绝对框架基准\bench{}包含18个系统，覆盖域内、跨数据集和跨城市层级，可在不进行配准或测试参考泄露的情况下评估绝对定位。在26个保留的US3D瓦片上，GeoRay在91.9%的覆盖率下达到2.99米的绝对平均绝对误差（MAE），比最强的符合要求的前馈基线提升46.4个百分点的完整性感知精度，在两种迁移偏移下仍是最准确的此类系统，每个瓦片的模型前向运行时间为24秒。代码和模型将在此https URL发布。
英文摘要
Feed-forward 3D foundation models reconstruct perspective scenes in one pass. Satellite photogrammetry needs a different product, one that domain adaptation alone does not deliver: dense surface height in an absolute geodetic frame under non-central rational polynomial cameras (RPCs). Perspective-pretrained features are not reliably observable along RPC height rays, absolute elevation carries a low-order height--datum gauge exchangeable with sensor bias to first order, and monocular and multi-view cues fail in different regions. \method{} treats all three. Lightweight ray-consistent adapters make a frozen backbone matchable along native RPC rays. An explicit datum mechanism separates relief from absolute level and is equivariant to the vertical origin by construction, so one trained model serves zero-, one-, and sparse-control inference. Calibrated inverse-variance fusion combines the two relief streams. \bench{}, our absolute-frame benchmark of eighteen systems across in-domain, cross-dataset, and cross-city tiers, scores absolute placement without registration or test-reference leakage. On 26 held-out US3D tiles, \method{} attains $2.99$\,m absolute MAE at $91.9\%$ coverage, improves completeness-aware accuracy by $46.4$ points over the strongest compliant feed-forward baseline, remains the most accurate such system under both transfer shifts, and runs in $24$\,s model-forward time per tile. Code and models will be released at https://github.com/HIT-SIRS/GeoRay
