supervoxels are, the more details of the scene will be blurred
in the segmentation result. However, one has to find a trade-
off between the accuracy and the efficiency. In fact, this is
also a common problem for the majority of the voxel-based
segmentation methods. In the general case, the sizes of voxels
and supervoxels are identified empirically, largely depending
on the requirement within the accuracy of segments and the
data quality. The selection of voxel sizes should be designed
as an adaptive step in our future work.
The manually-segmented ground truth should be refined
for the evaluation in our following work. Although the
rule for manual segmentation is fixed, namely each seg-
ment should correspond to a semantic object of the building
component, personal preferences of people will affect the
reliability of ground truth when they manually segment the
dataset. In some recent work (Vo
et al.
, 2015), multiple refer-
ence datasets are generated, and each of them is segmented
independently by point cloud processing experts. Then,
automatic segmentation results are compared against at least
two reference datasets individually. Moreover, when using
different manually-segmented reference dataset as ground
truth, evaluation results will also different due to the quality
of the manual segmentation process. To make the evaluation
more convincing, we need uniform criteria considering the
influence caused by the manual generation of the reference
dataset used as ground truth.
Acknowledgments
The author would like to express gratitude to the anonymous
reviewers and editor whose valuable comments and sugges-
tions have improved the quality of this paper. The authors
would like to thank Dr. Przemyslaw Polewski for his valu-
able help and support to this work. This work was carried
out within the frame of Leonhard Obermeyer Center (
LOC
) at
Technische Universitaet Muenchen (TUM) [
].
References
Aldoma, A., Z.-C. Marton, F. Tombari, W. Wohlkinger, C. Potthast,
B. Zeisl, R.B. Rusu, S. Gedikli, and M. Vincze, 2012. Tutorial:
Point cloud library: Three-dimensional object recognition and
6 dof pose estimation,
IEEE Robotics & Automation Magazine
,
19(3):80–91.
Aljumaily, H., D.F. Laefer, and D. Cuadra, 2017. Urban point cloud
mining based on density clustering and mapreduce,
Journal of
Computing in Civil Engineering
, 31(5), pp. 04017021.
Awrangjeb, M., and C.S. Fraser, 2014. Automatic segmentation of
raw lidar data for extraction of building roofs,
Remote Sensing
,
6(5):3716–3751.
Ballard, D.H., 1981. Generalizing the Hough transform to detect
arbitrary shapes,
Pattern Recognition
, 13(2):111–122.
Besl, P.J. and R.C. Jain, 1988. Segmentation through variable-order
surface fitting,
IEEE Transactions on Pattern Analysis and
Machine Intelligence
, 10(2):167–192.
Boerner, R., L. Hoegner, and U. Stilla, 2017. Voxel based segmentation
of large airborne topobathymetric lidar data,
ISPRS International
Archives of the Photogrammetry, Remote Sensing and Spatial
Information Sciences
, XLII-1/W1, pp. 107–114.
Cour, T., F. Benezit, and J. Shi, 2005. Spectral segmentation with
multiscale graph decomposition,
Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
, Vol. 2,
IEEE, pp. 1124–1131.
Felzenszwalb, P.F. and D.P. Huttenlocher, 2004. Efficient graph-based
image segmentation,
International Journal of Computer Vision
,
59(2):167–181.
Golovinskiy, A., and T. Funkhouser, 2009. Min-cut based
segmentation of point clouds,
Proceedings of the IEEE
International Conference on Computer Vision Workshop
, IEEE,
pp. 39–46.
Green, W.R. and H. Grobler, 2015. Normal distribution transform
graph-based point cloud segmentation,
Proceedings of the
Pattern Recognition Association of South Africa and Robotics
and Mechatronics International Conference
, IEEE, pp. 54–59.
Grilli, E., F. Menna, and F. Remondino, 2017. A review of point
clouds segmentation and classification algorithms,
ISPRS-
International Archives of the Photogrammetry, Remote Sensing
and Spatial Information Sciences
, XLII(2/W3), pp. 339–344.
Hackel, T., N. Savinov, L. Ladicky, J.D. Wegner, K. Schindler, and M.
Pollefeys, 2017. Semantic3d. net: A new large-scale point cloud
classification benchmark,
ISPRS Annals of Photogrammetry,
Remote Sensing and Spatial Information Sciences
, IV-1/W1, pp.
91– 98.
Hackel, T., J.D. Wegner, and K. Schindler, 2016. Contour detection
in unstructured 3D point clouds,
Proceedings of the IEEE
Conference on Computer Vision and Pattern Recognition
, pp.
1610–1618.
Ioannou, Y., B. Taati, R. Harrap, and M. Greenspan, 2012. Difference
of normals as a multi- scale operator in unorganized point
clouds,
Proceedings of International Conference on 3D Imaging,
Modeling, Processing, Visualization & Transmission
, IEEE
Computer Society, pp. 501–508.
Landrieu, L., and Simonovsky, M., 2017. Large-scale point cloud
semantic segmentation with superpoint graphs,
arXiv preprint
arXiv
:1711.09869.
Lin, Y., C, Wang, B. Chen, D. Zai, and J. Li, 2017. Facet segmentation-
based line segment extraction for large-scale point clouds,
IEEE
Transactions on Geoscience and Remote Sensing
, 55(9):4839–
4854.
Lu, X., J. Yao, J. Tu, K. Li, L. Li, and Y. Liu, 2016. Pairwise linkage for
point cloud segmentation,
ISPRS Annals of Photogrammetry,
Remote Sensing and Spatial Information Sciences
, III(3):201–
208.
Morsdorf, F.,E. Meier, B. Allgöwer, and D. Nüesch, 2003. Clustering in
airborne laser scanning raw data for segmentation of single trees,
ISPRS International Archives of the Photogrammetry, Remote
Sensing and Spatial Information Sciences
, 34(3):W13.
Nurunnabi, A., D. Belton, and G. West, 2012. Robust segmentation
for multiple planar surface extraction in laser scanning 3D point
cloud data,
Proceedings of the International Conference on
Pattern Recognition
, IEEE, pp. 1367–1370.
Nurunnabi, A., D. Belton, and G. West, 2016. Robust segmentation for
large volumes of laser scanning three-dimensional point cloud
data,
IEEE Transactions on Geoscience and Remote Sensing
,
54(8):4790–4805.
Nurunnabi, A., G. West, and D. Belton, 2015. Outlier detection and
robust normal-curvature estimation in mobile laser scanning 3D
point cloud data,
Pattern Recognition
, 48(4):1404–1419.
Papon, J., A, Abramov, M. Schoeler, and F. Worgotter, 2013. Voxel
cloud connectivity segmentation-supervoxels for point clouds,
Proceedings of the IEEE Conference on Computer Vision and
Pattern Recognition
, pp. 2027–2034.
Peng, B., L. Zhang, and D. Zhang, 2013. A survey of graph theoretical
approaches to image segmentation,
Pattern Recognition
,
46(3):1020–1038.
Pham, T.T., M. Eich, I. Reid, and G. Wyeth, 2016a. Geometrically
consistent plane extraction for dense indoor 3d maps
segmentation,
Proceedings of the IEEE/RSJ International
Conference on Intelligent Robots and Systems
, IEEE, pp.
4199–4204.
Pham, T.T., S. Hamid Rezatofighi, I. Reid, and T.-J.Chin, 2016b.
Efficient point process inference for large-scale object detection,
Proceedings of the IEEE Conference on Computer Vision and
Pattern Recognition
, pp. 2837–2845.
Rabbani, T., F. Van Den Heuvel, and C. Vosselmann, 2006.
Segmentation of point clouds using smoothness constraint,
ISPRS International Archives of the Photogrammetry, Remote
Sensing and Spatial Information Sciences
, 36(5):248–253.
Ramiya, A.M., R.R. Nidamanuri, and K. Ramakrishnan, 2016. A
supervoxel-based spectro-spatial approach for 3D urban point
cloud labelling,
International Journal of Remote Sensing
,
37(17):4172–4200.
390
June 2018
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
