proposed approach, the two models can compensate for each
other. Specifically, the image-based point cloud model can be
used to extend the
SLAM
point clouds from short distances,
whereas the
SLAM
model can provide a scale for the image-
based model. Moreover, as discussed, the
SLAM
results can
provide additional depth constraints for the
SfM
process to
improve it. The designed scale-adaptive registration can then
merge those kinds of point clouds into a common coordinate
system to produce enhanced and extended 3D mapping re-
sults. Two challenging cases were used to evaluate the perfor-
mance of the proposed solution. The theoretical analysis and
experimental validation yield the following conclusions.
1. The incorporation of additional depth constraints from the
SLAM
results benefits the offline
SfM
; moreover, the data
collection can be completed at one time measurement.
2. The fusion of the distant point cloud model from the
RGB
image sequences to the short-range point clouds from the
depth sensors can significantly improve the coverage of 3D
mapping results (more than 50
3. The designed scale-adaptive regi
geometric accuracy of the struct
in distant ranges is 1% at 20 m i
point quality (i.e., the bias is lower than 10 cm for the
major surface of about 800 m
2
).
Although
RGB-D
sensors are rarely used in real mapping cases,
this paper shows the potential of such sensors to generate en-
hanced and extended 3D models with high mobility and low
cost. Such low-cost equipment could be used to quickly build
3D models in large indoor spaces, such as shopping malls,
hospitals, and airports, for a variety of indoor navigation ap-
plications. Thus, our future work will not only focus on the
methods of related technical aspects but also consider the ap-
plication of equipment in mapping and modeling projects.
Acknowledgments
This work was supported by grants from the Hong Kong Poly-
technic University (Project Nos. 1-ZEAB and 1-ZVN6) and
grants from the National Natural Science Foundation of China
(Project Nos. 41671426 and 41471345).
References
Aiger, D., N. J. Mitra and D. Cohen-Or. 2008. 4-points congruent sets
for robust pairwise surface registration.
ACM Transactions on
Graphics
(
TOG
): 85.
Besl, P. J. and N. D. McKay. 1992. Method for registration of 3-D
shapes, sensor fusion IV: Control paradigms and data structures.
International Society for Optics and Photonics
: 586–607.
Bolles, R. C. and M. A. Fischler. 1981. A RANSAC-based approach
to model fitting and its application to finding cylinders in range
data.
IJCAI
: 637–643.
Byravan, A. and D. Fox. 2017. Se3-nets: Learning rigid body motion
using deep neural networks. Pages 173–180 in
2017 IEEE
International Conference on Robotics and Automation
(
ICRA
).
Chow, J. C., D. D. Lichti, J. D. Hol, G. Bellusci and H. Luinge. 2014.
IMU and multiple RGB-D camera fusion for assisting indoor
stop-and-go 3D terrestrial laser scanning.
Robotics
3 (3):247–280.
Comport, A. I., E. Malis and P. Rives. 2007. Accurate quadrifocal
tracking for robust 3d visual odometry.
ICRA
, Citeseer: 40–45.
Crandall, D., A. Owens, N. Snavely and D. Huttenlocher. 2011.
Discrete-continuous optimization for large-scale structure from
motion. Pages 3001–3008 in
2011 IEEE Conference on Computer
Vision and Pattern Recognition
(
CVPR
).
Dai, A., M. Nießner, M. Zollhöfer, S. Izadi and C. Theobalt. 2017.
Bundlefusion: Real-time globally consistent 3d reconstruction
using on-the-fly surface reintegration.
ACM Transactions on
Graphics
(
TOG
) 36 (4):76a.
Dryanovski, I., R. G. Valenti and J. Xiao. 2013. Fast visual odometry
and mapping from RGB-D data. Pages 2305–2310 in
2013 IEEE
International Conference on Robotics and Automation
(
ICRA
).
Du, S., N. Zheng, S. Ying, Q. You and Y. Wu. 2007. An extension of
the ICP algorithm considering scale factor. Pages V-193-V-196 in
IEEE International Conference on Image Processing
(
ICIP
), 2007.
Engel, J., T. Schöps and D. Cremers. 2014. LSD-SLAM: Large-scale
direct monocular SLAM. Pages 834–849 in
European Conference
on Computer Vision
. Springer.
Frahm, J. M., P. Fite-Georgel, D. Gallup, T. Johnson, R. Raguram, C.
Wu and M. Pollefeys. 2010. Building Rome on a cloudless day.
Pages 368–381 in
European Conference on Computer Vision
.
Springer.
Gao, X., R. Wang, N. Demmel and D. Cremers. 2018. LDSO:
Direct sparse odometry with loop closure.
arXiv
preprint:
iro and I. Reid. 2016. Unsupervised
th estimation: Geometry to the rescue.
an Conference on Computer Vision
.
Ge, X. 2017. Automatic markerless registration of point clouds with
semantic-keypoint-based 4-points congruent sets.
ISPRS Journal
of Photogrammetry and Remote Sensing
130:344–357.
Ge, X. and T. Wunderlich. 2015. Target identification in terrestrial
laser scanning.
Survey Review
47 (341):129–140.
Ge, X. and T. Wunderlich. 2016. Surface-based matching of 3D point
clouds with variable coordinates in source and target system.
ISPRS Journal of Photogrammetry and Remote Sensing
111:
1–12.
Ge, X. 2016. Terrestrial laser scanning technology from calibration to
registration with respect to deformation monitoring. Dissertation,
Technische Universität München.
Gherardi, R., M. Farenzena and A. Fusiello. 2010. Improving the
efficiency of hierarchical structure-and-motion. Pages 1594–1600
in
CVPR
.
Havlena, M. and K. Schindler. 2014. Vocmatch: Efficient multiview
correspondence for structure from motion. Pages 46–60 in
European Conference on Computer Vision
. Springer.
Heinly, J., E. Dunn and J. Frahm. 2012. Comparative evaluation of
binary features. Pages 759–773 in
Computer Vision–ECCV 2012
.
Springer.
Henry, P., M. Krainin, E. Herbst, X. Ren and D. Fox. 2010. RGB-D
mapping: Using depth cameras for dense 3D modeling of
indoor environments. In the
12th International Symposium on
Experimental Robotics
(
ISER
). Citeseer.
Hesch, J. A., D. G. Kottas, S. L. Bowman and S. I. Roumeliotis. 2014.
Camera-IMU-based localization: Observability analysis and
consistency improvement.
The International Journal of Robotics
Research
33 (1):182–201.
Huttenlocher, D. 1991. Fast affine point matching: An output-
sensitive method. Pages 263–268 in
Proceedings IEEE Computer
Society Conference on Computer Vision and Pattern Recognition
(
CVPR
), 1991.
Johnson, R. and T. Zhang. 2014. Effective use of word order for
text categorization with convolutional neural networks.
arXiv
preprint: arXiv:1412.1058.
Kerl, C., J. Sturm and D. Cremers. 2013. Dense visual SLAM for
RGB-D cameras. Pages 2100–2106 in
2013 IEEE/RSJ International
Conference on Intelligent Robots and Systems
(
IROS
). Citeseer.
Kerl, C., J. Sturm and D. Cremers. 2013. Robust odometry estimation
for RGB-D cameras. Pages 3748–3754 in
2013 IEEE International
Conference on Robotics and Automation
(
ICRA
)..
Kümmerle, R., G. Grisetti, H. Strasdat, K. Konolige and W. Burgard.
2011. g 2 o: A general framework for graph optimization. Pages
3607–3613 in
2011 IEEE International Conference on Robotics
and Automation
(
ICRA
). .
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
September 2019
641
