ABSTRACT
Large-scale 3D reconstruction for aerial photography is achallenging. For aerial image dataset, large scale means that the amount and resolution of images are enormous, which brings a huge amount of computation in Structure from Motion (SfM) pipeline, especially on the process of feature detection, feature matching and bundle adjustment (BA). In this paper, we present a novel method to solve the large-scale 3D reconstruction in parallel to accelerate the process. It could be generalized as the process of Divide-Reconstruct-Optimize-Fuse. We propose an effective graph-based framework that could robustly conduct aerial images grouping task and optimize parameters to fuse sub-models seamless. Experimental results on large-scale aerial datasets demonstrate the efficiency and robustness of the proposed method.
- Marc Pollefeys, Luc Van Gool, Maarten Vergauwen, Frank Verbiest, Kurt Cornelis, Jan Tops, and Reinhard Koch. Visual modeling with a hand-held camera. International Journal of Computer Vision, 59(3):207--232, 2004. Google ScholarDigital Library
- Sameer Agarwal, Yasutaka Furukawa, Noah Snavely, Ian Simon, Brian Curless, Steven M. Seitz, and Richard Szeliski. Building rome in a day. In IEEE International Conference on Computer Vision, pages 72--79, 2010.Google Scholar
- Johannes L. Schonberger, Filip Radenovic, Ondrej Chum, and Jan Michael Frahm. From single image query to detailed 3d reconstruction. In Computer Vision and Pattern Recognition, pages 5126--5134, 2015.Google ScholarCross Ref
- Changchang Wu. Towards linear-time incremental structure from motion. In International Conference on 3dtv-Conference, pages 127--134, 2013. Google ScholarDigital Library
- Johannes L. Sch¨ onberger and Jan Michael Frahm. Structure-from-motion revisited. In Computer Vision and Pattern Recognition, 2016.Google Scholar
- Kyle Wilson and Noah Snavely. Robust global translations with 1dsfm. In European Conference on Computer Vision, pages 61--75, 2014.Google ScholarCross Ref
- Chris Sweeney, Torsten Sattler, Tobias Hollerer, Matthew Turk, and Marc Pollefeys. Optimizing the viewing graph for structure-from-motion. In IEEE International Conference on Computer Vision, pages 801--809, 2015. Google ScholarDigital Library
- Xiaowei Li, Changchang Wu, Christopher Zach, Svetlana Lazebnik, and Jan Michael Frahm. Modeling and recognition of landmark image collections using iconic scene graphs. In European Conference on Computer Vision, pages 427--440, 2008. Google ScholarDigital Library
- Jan Michael Frahm, Pierre Fite-Georgel, David Gallup, Tim Johnson, Rahul Raguram, Changchang Wu, Yi Hung Jen, Enrique Dunn, Brian Clipp, and Svetlana Lazebnik. Building rome on a cloudless day. In European Conference on Computer Vision, pages 368--381, 2010. Google ScholarDigital Library
- Yueming Yang, Ming Ching Chang, Longyin Wen, Peter Tu, Honggang Qi, and Siwei Lyu. Efficient large-scale photometric reconstruction using divide-recon-fuse 3d structure from motion. In IEEE International Conference on Advanced Video and Signal Based Surveillance, pages 180--186, 2016.Google ScholarCross Ref
- Herbert Bay, Andreas Ess, Tinne Tuytelaars, and Luc Van Gool. Speeded-up robust features. Computer Vision and Image Understanding, 110(3):346--359, 2008. Google ScholarDigital Library
- David G Lowe. Distinctive image features from scale-invariant keypoints. International Journal of Computer Vision, 60(2):91--110, 2004. Google ScholarDigital Library
- Marius Muja and David G. Lowe. Scalable nearest neighbor algorithms for high dimensional data. Pattern Analysis and Machine Intelligence, IEEE Transactions on, 36, 2014.Google Scholar
- John E. Hopcroft and Robert E Tarjan. Efficient algorithms for graph manipulation. Stanford University, 1971.Google Scholar
- S Umeyama. Least-squares estimation of transformation parameters between two point patterns. Pattern Analysis and Machine Intelligence IEEE Transactions on, 13(4):376--380, 1991. Google ScholarDigital Library
- Rá lšl Mur-Artal, J. M. M. Montiel, and Juan D. Tardĺős. Orb-slam: A versatile and accurate monocular slam system. IEEE Transactions on Robotics, 31(5):1147--1163, 2017.Google ScholarDigital Library
- Raul Mur-Artal and Juan D. Tardĺős. Fast relocalisation and loop closing in keyframe-based slam. In IEEE International Conference on Robotics and Automation, pages 846--853, 2014.Google ScholarCross Ref
- Rainer Kuemmerle, Giorgio Grisetti and Others. g2o: http://openslam.org/g2o.html.Google Scholar
- Pierre Moulon, Pascal Monasse, Renaud Marlet, and Others. Openmvg. https://github.com/openMVG/openMVG.Google Scholar
- PIX4D. Pix4d dataset. https://support.pix4d.com.Google Scholar
- Y Furukawa and J Ponce. Accurate, dense, and robust multiview stereopsis. IEEE Transactions on Pattern Analysis and Machine Intelligence, 32(8):1362--1376, 2010. Google ScholarDigital Library
Index Terms
- Fast and Seamless Large-scale Aerial 3D Reconstruction using Graph Framework
Recommendations
GNN&GBDT-guided fast optimizing framework for large-scale integer programming
ICML'23: Proceedings of the 40th International Conference on Machine LearningThe latest two-stage optimization framework based on graph neural network (GNN) and large neighborhood search (LNS) is the most popular framework in solving large-scale integer programs (IPs). However, the framework can not effectively use the embedding ...
Large-scale outdoor 3D reconstruction on a mobile device
Approach for reconstructing large-scale outdoor scenes using monocular motion stereo.Runs at interactive frame rates on a Google Project Tango Development Kit Tablet.Can use strict outlier filtering, as the user can interactively record more ...
PRec-I-DCM3: a parallel framework for fast and accurate large-scale phylogeny reconstruction
Accurate reconstruction of phylogenetic trees often involves solving hard optimisation problems, particularly the Maximum Parsimony (MP) and Maximum Likelihood (ML) problems. Various heuristics yield good results for these problems within reasonable ...
Comments