Feature-based quality evaluation of 3D heterogeneous data registration

被引:0
作者
Ridene, Taha [1 ]
Goulette, Francois [1 ]
机构
[1] Mines ParisTech, CAOR Ctr Robot Math & Syst, F-75272 Paris 06, France
来源
THREE-DIMENSIONAL IMAGE PROCESSING (3DIP) AND APPLICATIONS | 2010年 / 7526卷
关键词
3D Registration; ICP; RANSAC; FB-Evaluation; Point Clouds; DSM; Mobile Mapping Systems;
D O I
10.1117/12.847722
中图分类号
O43 [光学];
学科分类号
070207 ; 0803 ;
摘要
The production of 3D realistic map databases is a matter of increasing interest. An approach based on the fusion of heterogeneous 3D representations was studied in previous works. Three methods of rigid registration were tested as well as the behaviors of those algorithms. One of the most difficult problems is the quantitative evaluation of 3D registration. In this term we propose a new quantitative evaluation method based on reference features comparison to improve and compare the accuracy of the registration methods.
引用
收藏
页数:8
相关论文
共 12 条
  • [1] Besl P., 1992, TRANSACTIONS, V14, P239
  • [2] OBJECT MODELING BY REGISTRATION OF MULTIPLE RANGE IMAGES
    CHEN, Y
    MEDIONI, G
    [J]. IMAGE AND VISION COMPUTING, 1992, 10 (03) : 145 - 155
  • [3] Robust smooth feature extraction from point clouds
    Daniels, Joel, II
    Ha, Linh K.
    Ochotta, Tilo
    Silva, Claudio T.
    [J]. IEEE INTERNATIONAL CONFERENCE ON SHAPE MODELING AND APPLICATIONS 2007, PROCEEDINGS, 2007, : 123 - +
  • [4] GOULETTE F, 2005, P ISPRS COMM I S SEN, P3
  • [5] Gumhold S., 2001, P 10 INT MESH ROUNDT, P293
  • [6] LATREMOLIERE F, 2005, J FUNCT ANAL, P365
  • [7] MELONI F, 2004, P EUR ACM SIGGRAPH S, P40
  • [8] Multi-scale feature extraction on point-sampled surfaces
    Pauly, M
    Keiser, R
    Gross, M
    [J]. COMPUTER GRAPHICS FORUM, 2003, 22 (03) : 281 - 289
  • [9] RIDENE T, 2009, 8 IEEE INT S CIRA200
  • [10] Ridene T., 2009, MOBILE MAPPING TECHN