e XXXIX-B3, 2012
b
c denotes the translation
^s. For this, it is assumed
the center between both
jon
n calculated for the Ieg-
zed for detecting special
Iculating the coarse area
ng features which should
ss as they arise from ob-
hese features have to be
ickground being relevant
e representations of sev-
ibase before starting the
images contains a tem-
ne, but from a different
ice and at a different sea-
intensity images show a
e detected SIFT features
> to the object templates
1 maximum similarity to
transformation based on
template is transformed.
emplate is then assumed
lure allows for detecting
ne as well as for decou-
the presented approach
ase of dynamic environ-
ly known.
IMAGING SYSTEMS
1e scene monitoring with
à future operational sys-
ces fairly realistically, a
owever, due to the large
le system, mounting the
1 and data storage on an
impracticable. Hence,
active multi-view range
long a rope is used as
ire 3. The components
ision] CamCube 2.0) for
sk for efficiently storing
pendent power supply.
ge imaging devices can
variable multi-view op-
or divergent data acqui-
nce of noise effects aris-
liation in comparison to
ipath scattering, the uti-
range accuracy of a few
expected. Furthermore,
time-of-flight cameras,
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
Figure 3: PMD[vision] CamCube 2.0 and model of a cable car
equipped with two range imaging devices.
depends on the modulation frequency fm, where co denotes the
speed of light. A modulation frequency of 20 MHz thus corre-
sponds to à non-ambiguous range of 7.5 m. In order to overcome
this range measurement restriction, image- or hardware-based un-
wrapping procedures have been introduced (Jutzi, 2009; Jutzi,
2012). When dealing with multiple range imaging devices, it also
has to be taken into account that these may influence each other
and that interferences are likely to occur. This can be overcome
by choosing different modulation frequencies.
4 EXPERIMENTAL RESULTS
The estimation of the flight trajectory of a sensor platform re-
quires the definition of a global world coordinate frame. This
world coordinate frame is assumed to equal the local coordinate
frame of the sensor platform at the beginning. The local coor-
dinate frame has a fixed orientation with respect to the sensor
platform. It is oriented with the X -direction in forward direction
tangential to the rope, the Y -direction to the right and the Z-
direction downwards. For evaluating the proposed methodology,
a successive pairwise registration is performed. The threshold for
the matching of 2D features is selected as tdes = 0.7. The result-
ing 2D/2D correspondences are projected into 3D space which
yields 3D/3D correspondences. Including the weights in the esti-
mation of the rigid transformation yields position estimates and,
finally, an estimated trajectory which is shown in Figure 4 in nadir
view and in Figure 5 from the side. The green and blue points
describe thinned point clouds captured with both range imaging
devices and transformed to the global world coordinate frame.
Y [m]
X [m]
Figure 4: Projection of the estimated trajectory and thinned point
cloud data onto the X Y -plane.
A limitation of the experimental setup seems to be the fact that
no reference values are available for checking the deviation of
the position estimates from the real positions. However, due to
the relative orientation of the sensor platform to the rope, the
projection of the real trajectory onto the X Y -plane should ap-
proximately be a straight line. Additionally, the length of the real
trajectory projected onto the ground plane can be estimated from
aerial images or simply be measured. Here, the distance Aground
between the projections of the end points onto the ground plane
has been measured as well as the difference Aaititude between
maximum and minimum altitude. From the measured values of
Âground = T m and Acititude = 1.25 m, a total distance of
Z [m]
X [m]
Figure 5: Projection of the estimated trajectory and thinned point
cloud data onto the X Z-plane.
approximately 7.11 m can be assumed. A comparison between
the start position and the point with the maximum distance on the
estimated trajectory results in a distance of 6.90 m. As a con-
sequence, the estimated trajectory can be assumed to be of rela-
tively high quality. The results for a subsequent object detection
and segmentation is illustrated for an example frame in Figure 6.
Figure 6: SIFT-based object detection and segmentation: normal-
ized active intensity image, template and transformed template
(upper row, from left to right). The corresponding point cloud for
the area of the transformed template and the sensor position (red
dot) are shown below.
5 DISCUSSION
The presented methodology is well-suited for dynamic environ-
ments. Instead of considering the whole point clouds, the prob-
lem of registration is reduced on sparse point clouds of physically
almost identical 3D points. Due to this fact and the non-iterative
processing scheme, the proposed algorithm for point cloud reg-
istration is very fast which is required for monitoring in such
demanding environments. Although the current Matlab imple-
mentation is not fully optimized with respect to parallelization of
tasks, a total time of approximately 1.63 s is required for pre-
processing, point quality assessment, feature extraction and point
projection. Further 0.46 s are required for feature matching, cal-
culation of weights and point cloud registration. This can signif-
icantly be reduced with a GPU-implementation of SIFT, as the
calculation of SIFT features already takes approximately 1.54 s.
Furthermore, the simple estimation of a rigid transformation is
not sufficient, as used 3D/3D correspondences have the same
weight, even if the uncertainty of the respective 3D points is very
high or if outlier correspondences not fitting to the transforma-
tion have been detected. Hence, a quality measure for 3D/3D
correspondences has been introduced which is based on the qual-
ity of the respective 3D points. This quality measure is used for
