发表机构
School of Geodesy and Geomatics, Wuhan University（武汉大学测绘学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对现有流式3R方法不适用于无人机跨航线无序图像流的问题，提出On-the-Fly3R框架，通过检索引导子集构建等技术实现大规模无人机场景的鲁棒在线3R，精度优于SOTA方法。
AI 中文摘要
前馈三维重建（3D Reconstruction，3R）虽能提供高效的端到端建模，但其在大规模无人机测绘中的应用受限于Transformer注意力机制过高的内存成本。当前可扩展的流式3R方法假设输入是时间和空间连续的，这使得它们在无人机跨航线作业中常见的弱有序或无序图像流中效果不佳。为解决该问题，我们提出On-the-Fly3R，这是一种面向大规模无人机图像的无训练渐进式在线三维重建框架，可将各类3R骨干模型升级以适配大规模无人机场景。我们的方法通过检索引导的动态子集构建实现从无序输入中重建，该过程会自适应选择空间相关的图像。为进一步提升鲁棒性，我们设计了验证-拒绝-重试机制以保证全局一致性，该机制执行预集成一致性检查，自动拒绝未对齐的图像并使用替代子集重试。最后，受视觉同步定位与建图（VSLAM）启发，采用基于检索回环的位姿图优化来缓解相机漂移。在多个无人机基准上的评估表明，我们的On-the-Fly3R成功将各类3R模型扩展至平方公里级无人机场景中超过5000张图像的规模，与多个SOTA流式3R方法相比，其精度显著更优。代码可在该https URL获取。
英文摘要
While feed-forward 3D reconstruction (3R) offers efficient end-to-end modeling, its application in large-scale UAV mapping is hindered by the prohibitive memory cost of Transformer attention. Current scalable streaming 3R methods assume temporally and spatially continuous inputs, rendering them ineffective for the weakly ordered or unordered image streams common in cross-strip UAV operations. To address this, we propose On-the-Fly3R, a training-free, progressive online 3D reconstruction framework for large-scale UAV images that upgrades various 3R backbones for large-scale UAV scenarios. Our method enables reconstruction from unordered inputs via retrieval-guided dynamic subset construction, which adaptively selects spatially relevant images. To further improve the robustness, a validation-rejection-retry mechanism is designed to guarantee global consistency, performing a pre-integration consistency check and automatically rejecting misaligned images and retrying with alternative subset. Finally, inspired by VSLAM, pose graph optimization based on the retrieval loop closure is employed to mitigate camera drift. Evaluations on several UAV benchmarks show that our On-the-Fly3R successfully scales various 3R models to over 5,000 images across square-kilometer UAV scenes, delivering substantially superior accuracy compared to several SOTA streaming 3R methods. Code is available at https://github.com/Sh1nZzz/On_the_Fly3R
CommentsThis paper was submitted to the ICRA 2027 for consideration. Copyright would be transferred if it got accepted
