TY - JOUR
T1 - Detailed real-time urban 3D reconstruction from video
AU - Pollefeys, M.
AU - Nistér, D.
AU - Frahm, J. M.
AU - Akbarzadeh, A.
AU - Mordohai, P.
AU - Clipp, B.
AU - Engels, C.
AU - Gallup, D.
AU - Kim, S. J.
AU - Merrell, P.
AU - Salmi, C.
AU - Sinha, S.
AU - Talton, B.
AU - Wang, L.
AU - Yang, Q.
AU - Stewénius, H.
AU - Yang, R.
AU - Welch, G.
AU - Towles, H.
PY - 2008/7
Y1 - 2008/7
N2 - The paper presents a system for automatic, geo-registered, real-time 3D reconstruction from video of urban scenes. The system collects video streams, as well as GPS and inertia measurements in order to place the reconstructed models in geo-registered coordinates. It is designed using current state of the art real-time modules for all processing steps. It employs commodity graphics hardware and standard CPU's to achieve real-time performance. We present the main considerations in designing the system and the steps of the processing pipeline. Our system extends existing algorithms to meet the robustness and variability necessary to operate out of the lab. To account for the large dynamic range of outdoor videos the processing pipeline estimates global camera gain changes in the feature tracking stage and efficiently compensates for these in stereo estimation without impacting the real-time performance. The required accuracy for many applications is achieved with a two-step stereo reconstruction process exploiting the redundancy across frames. We show results on real video sequences comprising hundreds of thousands of frames.
AB - The paper presents a system for automatic, geo-registered, real-time 3D reconstruction from video of urban scenes. The system collects video streams, as well as GPS and inertia measurements in order to place the reconstructed models in geo-registered coordinates. It is designed using current state of the art real-time modules for all processing steps. It employs commodity graphics hardware and standard CPU's to achieve real-time performance. We present the main considerations in designing the system and the steps of the processing pipeline. Our system extends existing algorithms to meet the robustness and variability necessary to operate out of the lab. To account for the large dynamic range of outdoor videos the processing pipeline estimates global camera gain changes in the feature tracking stage and efficiently compensates for these in stereo estimation without impacting the real-time performance. The required accuracy for many applications is achieved with a two-step stereo reconstruction process exploiting the redundancy across frames. We show results on real video sequences comprising hundreds of thousands of frames.
KW - 3D reconstruction
KW - Depth map fusion
KW - Large scale modeling
KW - Plane sweeping
KW - Stereo vision
KW - Structure from motion
KW - Urban reconstruction
UR - http://www.scopus.com/inward/record.url?scp=40949155911&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=40949155911&partnerID=8YFLogxK
U2 - 10.1007/s11263-007-0086-4
DO - 10.1007/s11263-007-0086-4
M3 - Article
AN - SCOPUS:40949155911
SN - 0920-5691
VL - 78
SP - 143
EP - 167
JO - International Journal of Computer Vision
JF - International Journal of Computer Vision
IS - 2-3
ER -