GeoFF3D: Coordinate-Anchored Feed-Forward Reconstruction for Large-Scale UAV Mapping
Computer Vision and Pattern Recognition
Summary
The authors introduce GeoFF3D, a method that improves 3D mapping from drone images taken over large areas. Unlike previous methods that only work on small image sets, GeoFF3D splits images into overlapping chunks and combines them carefully to create accurate maps in real-world coordinate frames. Their approach handles challenges with camera alignment and produces more precise reconstructions quickly, outperforming existing methods on several tests. GeoFF3D can reconstruct thousands of images in minutes, showing it works well for large-scale drone mapping.
Authors
Xiang Yang, Yongli Wang, Yunsheng Zhang
Abstract
Existing feed-forward 3D reconstruction methods typically process a bounded number of images and recover cameras and geometry in local or internally normalized frames. Extending them to large-scale UAV mapping requires scalable multi-chunk processing and reliable aggregation, while full Sim(3) alignment can become unstable for near collinear trajectories. We present GeoFF3D, which combines a coordinate-anchored model with a spatial large-scale reconstruction framework (SLRF). The model uses georeferenced camera translations and optional geometric priors to predict camera poses and dense point maps directly in a gravity-aligned Z-up metric frame. SLRF partitions images into spatially overlapping chunks, propagates shared-view priors, and aggregates local reconstructions hierarchically, while remaining applicable to different bounded-view models. Across nine aerial mapping blocks, GeoFF3D achieves the best average reconstruction quality, improving F@5 from 0.829 for Pi3X + SLRF to 0.877. On long UAVScenes sequences, it reaches 0.848, compared with 0.687 for Pi3X + SLRF and 0.451 for the strongest evaluated SLAM/streaming baseline. GeoFF3D reconstructs 2,000 images in approximately five minutes, demonstrating scalable and robust large-scale UAV reconstruction.The code is available at https://github.com/yanxian-ll/GeoFF3D.