TY - JOUR
T1 - Fast robust large-scale mapping from video and internet photo collections
AU - Frahm, Jan Michael
AU - Pollefeys, Marc
AU - Lazebnik, Svetlana
AU - Gallup, David
AU - Clipp, Brian
AU - Raguram, Rahul
AU - Wu, Changchang
AU - Zach, Christopher
AU - Johnson, Tim
N1 - Funding Information:
We would like to acknowledge the DARPA UrbanScape project, Department of Energy, Navy SPAWAR and NSF Grant IIS-0916829 , as well as our collaborators David Nister, Amir Akbarzadeh, Philippos Mordohai, Paul Merrell, Chris Engels, Henrik Stewenius, Brad Talton, Liang Wang, Qingxiong Yang, Ruigang Yang, Greg Welch, Herman Towles, Xiaowei Li.
PY - 2010/11
Y1 - 2010/11
N2 - This paper presents a system approaching fully automatic 3D modeling of large-scale environments. Our system takes as input either a video stream or collection of photographs obtained from Internet photo sharing web-sites such as Flickr. The system achieves high computational performance through algorithmic optimizations for efficient robust estimation, the use of image-based recognition for efficient grouping of similar images, and two-stage stereo estimation for video streams that reduces the computational cost while maintaining competitive modeling results. In addition to algorithmic advances, we achieve a major improvement in computational speed through parallelization and execution on commodity graphics hardware. These improvements lead to real-time video processing and to reconstruction from tens of thousands of images within the span of a day on a single commodity computer. We demonstrate modeling results on a variety of real-world video sequences and photo collections.
AB - This paper presents a system approaching fully automatic 3D modeling of large-scale environments. Our system takes as input either a video stream or collection of photographs obtained from Internet photo sharing web-sites such as Flickr. The system achieves high computational performance through algorithmic optimizations for efficient robust estimation, the use of image-based recognition for efficient grouping of similar images, and two-stage stereo estimation for video streams that reduces the computational cost while maintaining competitive modeling results. In addition to algorithmic advances, we achieve a major improvement in computational speed through parallelization and execution on commodity graphics hardware. These improvements lead to real-time video processing and to reconstruction from tens of thousands of images within the span of a day on a single commodity computer. We demonstrate modeling results on a variety of real-world video sequences and photo collections.
KW - 3D modeling from video
KW - 3D registration video
KW - Camera registration photo collections
UR - http://www.scopus.com/inward/record.url?scp=78149497497&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=78149497497&partnerID=8YFLogxK
U2 - 10.1016/j.isprsjprs.2010.08.009
DO - 10.1016/j.isprsjprs.2010.08.009
M3 - Article
AN - SCOPUS:78149497497
VL - 65
SP - 538
EP - 549
JO - ISPRS Journal of Photogrammetry and Remote Sensing
JF - ISPRS Journal of Photogrammetry and Remote Sensing
SN - 0924-2716
IS - 6
ER -