one defines the reliability percentages of the matches (Equa-
tion 4). The ability of the model to find inliers is defined as
the capacity-factor (Equation 5).
validity factor
No of Correct Matched Lines
No of Match
−
=
−
.
.
ed Lines
−
*100 (4)
capacity factor
No of Correct Matched Lines
Total No of
−
=
−
.
.
*
Inliers
100 (5)
• For the
HQPS
-phase, the proposed impact-factors (
p
i
)
are tested over the range of different values of {3, 2.5,
2, and 1.5}. These impact-factors show the importance
of different quality-elements introduced for
HQPS
-
phase (Figure 7). Figure 7 demonstrates the number
of matched-lines (vertical axis) over the first dataset
for some different combinations (cases 1 to 24) of the
above proposed values (horizontal axis). Also, the case
25 shows the situation when all the impact-factors
have a similar value (
p
1
=
p
2
=
p
3
=
p
4
).
Among the above results (Figure 7), it can be observed
that except cases 3 and 4, approximately, similar number of
matched-lines are found in most cases. The capacity-factors
in most cases are up to 90 percent. It means, the selected
triangles based on the proposed
HQPS
strategy are more likely
to have correspondence in object space.
Figure 5. Flowchart of the proposed SLIM.
T
able
1. T
he
S
pecifications
of
the
F
irst
and
S
econd
D
atasets
Datasets
sun angle
elevation (degree)
scan azimuth
(degree)
Image size
information (pix.)
Date
GPS
(m)
NA
(m)
No. of extracted lines in
image and object spaces
Total no.
of inliers
1) GeoEye
63.26
270.83
19220*30784
8.8.2010 0.5
5
175*276
75
2) IKONOS
47.4
20.4
15108*10976 10.7.2000 1
15
126*144
61
T
able
2. T
he
S
pecifications
of
the
T
hird
D
ataset
Datasets
Image size information (pix.)
GPS (m)
NA
(m) No. of extracted lines in image and object spaces Total no. of inliers
3- UltraCam-D
2300*1500
0.2
0.2
44*241
31
370
May 2016
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
