International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B3, 2012
ume XXXIX-B3, 2012 |
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
d 71 known high precision RMS coordinates of image points extracted and matched by the
alidation. The efficiency of y y y y proposed new method reaches a subpixel level and the accuracy
ompared to the commercial New 0.8 12 SS > of aerial triangulation is +1.5um = £0.21 pixels. The skip rate is
010. The latter provides method : S s 1.0876.
LPS 2010 1.8 17 7.2 7.5
(a) residuals of photo coordinates ( x , y )
sed on least-squared image
x 5 standard tie points per Also, the efficiency and benefit of AFTP and QF processing
ber of N-fold tie points
[FT-supported dense point
S/ERDAS Imagine 2010.
rovide denser cloud of tie
S/ERDAS Imagine 2010.
4 5 6
712 | 3415 | 1211
230.1: 231 | 238
RMS (mm) MAX(mm)
VX Vy Vz | Yx | | vv | | vz |
New
Soie 11 8 112 17 11 142
LPS 2010 21 2 188 32 3 142
(b) residuals of ground coordinates (X, Y, Z) of 6 GCPs
10.] 11:|.12 |. 13 1 14
164 | 98 | 80 | 63 | 24
RMS (mm) RMS (1/1000 gon)
Gx óy, Gu 65 ó Gy
New
method 70 67 103 4.1 3.9 0.4
661 371221127 8
al number of tie points
89210
1764
old tie points
nber of Skip Rate
ipped (Unit: %)
oints
2548 1.08
174 1.79
Total
Value Redundancy
1.05 193965
1.00 14371
stments done by the new
2010
by means of free network
in Table 6, where the test
'acy divided by posteriori
od provides the denser tie
curacy à, = +1.5um =
t with control data is
ints and 65 independent
:d. The statistic figures of
he tie points measured by
and by the commercial
7. The root mean square
s on all 65 check points
he new method have the
rtical accuracy +11.9cm.
thod is available to aerial
and good efficiency.
eed any information on
v method don’t need the
ters and strip parameters
nmercial AT softwares.
LPS 2010 123 121 166 7.9 7.2 0.9
(c) posteriori standard deviations of exterior orientation data
RMS (mm)
X Y Z
New
method 28 19 119
LPS 2010 31 20 214
(d) ground coordinate differences on 65 check points
Table 7. Results of bundle block adjustments done by the new
method and LPS2010
4. CONCLUSION
Although the commercial solutions (e.g. Bundler, pix4d.com.
acute3d.com) are out there which do almost exactly what is
proposed in this paper, namely they use SIFT or an equivalent
detector/descriptor scheme to match an unorganized set of
images in unknown configuration, and orient them, this paper
was developed in the work of a master thesis in NCKU, and
proposes an alternative method for dense point cloud matching
and aero triangulation based on the well-known scale invariant
feature transform (SIFT) technique. Test results verify some
advantages of this method. The proposed QF operation reserves
those key points with best image quality, e.g. in the top 16% of
all key points. This method uses AFTP to estimate the overlap
area, and to predict the location of searching window. In order
to increase the computational efficiency, image pyramid with a
top level image of about 700 x 700 pixels is adopted. The
automatic extraction, selection and measurement of
corresponding tie points is done efficiently, especially without
the need on block parameters and strip parameters for providing
priori knowledge on image overlap information. Moreover, the
density of key points depends not only on the distance ratio
threshold of SIFT but also on the amount of image information
as well as image quality.
Tests are done by using 108 aerial images. They show that
operation with both QF and AFTP provides better matching
efficiency and best matching with lowest rate of skipped points,
when distance ratio is set to the threshold less than 0.3. In
general, the larger the distance ratio is, more points are matched
per second, but the more the skipped points become. Due to the
robustness of bundle block adjustment with data snooping
operation, the skipped, namely inaccurate, points can be
detected and eliminated. Furthermore, the accuracy of photo
73
have been verified. These two pre-process procedures still can
be further improved on efficiency and feasibility, especially the
quality indicator for QF. Moreover, some extension versions of
the semiglobal matching (SGM) will be further developed in the
future to perform a robust pixelwise matching with quality
figures. Their applications such as on high resolution true ortho
image generation, high resolution and high precision digital
surface model generation as well as 3D cyber city modelling
with a good LOD (level of detail) will also be studied and
developed.
5. REFERENCE
Baarda, W., 1968. A Testing Procedure for Use in Geodetic
Networks. Netherlands Geodetic Commission, Vol. 2, No. 5.,
Delft.
Computerworld, 2012. Microsoft combines Photosynth, Virtual
Earth. Downloaded on April 22, 2012 from http://www.
computerworld.com/s/article/9132675/Microsoft combines Ph
otosynth Virtual Earth.
Heipke, C., 1997. Special issue: Automatic image orientation -
Preface. ISPRS Journal of Photogrammetry and Remote
Sensing, 52(3), 101-102.
Hirschmueller, H., 2008. Stereo Processing by Semiglobal
Matching and Mutual Information. /EEE Transactions on
Pattern Analysis and Machine Intelligence, 30(2), pp. 328-341.
Hirschmueller, H., 2011. Semi-Global Matching — Motivation,
Developments and Applications, 53" Photogrammetric Week,
pp. 173-184, University of Stuttgart, Germany.
Kruck, E., 1984. Ordering and Solution of Large Normal
Equation Systems for Simultaneous Geodetic and
Photogrammetric Adjustment. 15th ISPRS Congress, 17-29
June, 1984, Rio de Janeiro, BRASIL.
Lowe, D. G., 2004. Distinctive image features from scale-
invariant keypoints. International Journal of Computer Vision,
60(2), pp. 91-110.
RealityMaps, 2012. 3D RealityMaps - Faszinierende
Landschaften in der dritten Dimension, downloaded on April 22,
2012 from http://www.reality-maps.de/.
Scharstein, D. and Szeliski, R., 2011. Middlebury Stereo
Evaluation. http://vision.middlebury.edu/stereo.
Siegert, F., 2011. Very Advanced 3D Landscape Models, 53'7
Photogrammetric Week, pp. 295-303, University of Stuttgart,
Germany, September 5-9, 2011.
