AliceVision Meshroom: An open-source 3D reconstruction pipeline

被引:115
作者
Griwodz, Carsten [1 ,2 ]
Gasparini, Simone [3 ]
Calvet, Lilian [3 ]
Gurdjos, Pierre [3 ]
Castan, Fabien [4 ]
Maujean, Benoit [4 ]
De Lillo, Gregoire [4 ]
Lanthony, Yann [4 ]
机构
[1] Univ Oslo, Oslo, Norway
[2] SimulaMet, Oslo, Norway
[3] Univ Toulouse, IRIT, Toulouse, France
[4] Mikros, Technicolor Creat Studios, Paris, France
来源
MMSYS '21: PROCEEDINGS OF THE 2021 MULTIMEDIA SYSTEMS CONFERENCE | 2021年
关键词
open source; 3D computer vision; photogrammetry; nodal frame-work; camera tracking; HDR imaging; panorama; point clouds;
D O I
10.1145/3458305.3478443
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
This paper introduces the Meshroom software and its underlying 3D computer vision framework AliceVision. This solution provides a photogrammetry pipeline to reconstruct 3D scenes from a set of unordered images. It also features other pipelines for fusing multi-bracketing low dynamic range images into high dynamic range, stitching multiple images into a panorama and estimating the motion of a moving camera. Meshroom's nodal architecture allows the user to customize the different pipelines to adjust them to their domain specific needs. The user can interactively add other processing nodes to modify a pipeline, export intermediate data to analyze the result of the algorithms and easily compare the outputs given by different sets of parameters. The software package is released in open source and relies on open file formats. These features enable researchers to conveniently run the pipelines, access and visualize the data at each step, thus promoting the sharing and the reproducibility of the results.
引用
收藏
页码:241 / 247
页数:7
相关论文
共 51 条
[1]   Fast Explicit Diffusion for Accelerated Features in Nonlinear Scale Spaces [J].
Alcantarilla, Pablo F. ;
Nuevo, Jesus ;
Bartoli, Adrien .
PROCEEDINGS OF THE BRITISH MACHINE VISION CONFERENCE 2013, 2013,
[2]  
Allène C, 2008, INT C PATT RECOG, P2539
[3]  
Baumberg Adam, 2003, P BR MACH VIS C
[4]  
Blanco J. L., 2014, NANOFLANN C HEADER O
[5]   An experimental comparison of min-cut/max-flow algorithms for energy minimization in vision [J].
Boykov, Y ;
Kolmogorov, V .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2004, 26 (09) :1124-1137
[6]   Automatic panoramic image stitching using invariant features [J].
Brown, Matthew ;
Lowe, David G. .
INTERNATIONAL JOURNAL OF COMPUTER VISION, 2007, 74 (01) :59-73
[7]   A MULTIRESOLUTION SPLINE WITH APPLICATION TO IMAGE MOSAICS [J].
BURT, PJ ;
ADELSON, EH .
ACM TRANSACTIONS ON GRAPHICS, 1983, 2 (04) :217-236
[8]   Detection and Accurate Localization of Circular Fiducials under Highly Challenging Conditions [J].
Calvet, Lilian ;
Gurdjos, Pierre ;
Griwodz, Carsten ;
Gasparini, Simone .
2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, :562-570
[9]  
Calvet Lilian, 2014, THESIS I NATL POLYTE
[10]   Fast and Accurate Image Matching with Cascade Hashing for 3D Reconstruction [J].
Cheng, Jian ;
Leng, Cong ;
Wu, Jiaxiang ;
Cui, Hainan ;
Lu, Hanqing .
2014 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2014, :1-8