The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2, 2018 
ISPRS TC II Mid-term Symposium “Towards Photogrammetry 2020”, 4-7 June 2018, Riva del Garda, Italy 
This contribution has been peer-reviewed. 
https://doi.org/10.5194/isprs-archives-XLII-2-961-2018 | ©Authors 2018. CC BY4.0 License. 
965 
(a) 
(b) 
Time step 
(c) 
Figure 5. Coordinate displacement error (percentage of water 
depth) obtained from 100 runs of the simulation tool with a 
flying height of 500 m, (a) horizontal water surface, (b) tilted 
water surface point density 3 points per square meter (p/m 2 ), (c) 
tilted water surface point density 10 p/m 2 . 
less prominent than the lateral component. For a flying height of 
500 m, the RMS depth errors vary in a range of 0.12 % to 0.28 % 
(0.2 cm to 0.4 cm) depending on the chosen correction method. 
The variations between the different flying heights are less dis 
tinctive. Furthermore, the water surface complexity does not nec 
essarily affect the depth errors (cf. hz6oo/tl60o). 
5.2 Experimental Validation 
In contrast to the simulation approach, the local wave-induced 
water surface (Fig. 1, blue) is unknown in the experimental val 
idation, inhibiting the calculation of dXYhz, dZh z , dXY t n t and 
dZtnt. Therefore, the discrepancies between the coordinates ob 
tained from the different correction methods are analyzed. The 
results presented in Table 2 show that the planimetrie coordinates 
are mainly influenced by the local surface tilt. At a flying height 
of 500 m the consideration of the local height (Mi - Mi ) reveals a 
coordinate difference of 2.29%. Taking into account the local tilt 
of the water surface results in a difference of 6.67 % (Mi - M3 ) 
respectively 6.17 % (M2 - M3). The comparatively small depth 
min. 
dXY 
max. 
RM SE 
min. 
dZ 
max. 
RMSE 
hz 5 oo 
0.14 
3.12 
1.40 
-0.62 
0.66 
0.28 
H500 
0.14 
2.45 
1.02 
-0.73 
0.64 
0.21 
HO500 
0.02 
0.64 
0.23 
-0.38 
0.43 
0.12 
hz6oo 
0.15 
2.33 
1.16 
-0.69 
0.72 
0.23 
tlôOO 
0.09 
1.85 
0.93 
-0.98 
0.72 
0.31 
tlCLoo 
0.04 
1.44 
0.60 
-0.87 
0.87 
0.25 
hz 7 oo 
0.16 
3.45 
1.19 
-0.5 
0.48 
0.22 
H700 
0.10 
2.71 
0.85 
-0.60 
1.06 
0.25 
UO700 
0.02 
1.63 
0.51 
-0.63 
0.90 
0.24 
Table 1. Planimetric and depth coordinate displacements (in 
percent of the water depth) for different flying heights (500 m, 
600 m, 700 m) with respect to horizontal water surface (hz), 
locally tilted water surface with 1 p/m 2 (tl) and 10 p/m 2 (tlO). 
dXY RMSE 
dZ RMSE 
M 
500 m 
600 m 
700 m 
500 m 
600 m 
700 m 
1-2 
2.29 
2.19 
2.48 
2.82 
2.51 
2.83 
1-3 
6.57 
5.57 
6.53 
3.18 
2.76 
3.10 
2-3 
6.17 
5.17 
5.94 
1.30 
0.97 
1.08 
Table 2. Discrepancies between the coordinates obtained from 
the different correction methods M ; (Mi - M 2 , Mi - M3, 
M 2 - M3) in percent of the water depth for different flying 
heights (500 m, 600 m, 700 m). 
coordinate differences of 1.30% for the comparison of method 
M 2 and M3 show that the local surface tilt is less important for 
the depth coordinate whereas the local height is more essential 
(Mi - M 2 and Mi - M3). The results are similar for different fly 
ing heights. 
The comparison of refraction corrected data and terrestrial ref 
erence data is based on the investigation of deviations between 
ALS and TLS point cloud at the pool bottom (depth displace 
ment) and at the pool wall (planimetric displacement). Due to the 
oblique incidence angle in combination with the highly reflecting 
material, the point coordinates at the pool wall are affected by 
noise which superimposes the geometric effects. Therefore, only 
a small number of points at the concrete base of the water slide 
are available for the planimetry effect evaluation. The relevant 
region is marked in figure 6. The deviations initially contain sys 
tematic errors due to the limited registration accuracy as well as 
effects of the beam divergence and incidence angle. In order to 
achieve the random part of the deviations, the systematic errors 
have to be eliminated. For that purpose, we estimate the system 
atic errors based on the results of the third refraction correction 
method M3, which provides the best available correction results 
(cf. table 2). 
Figure 6. TLS reference point cloud with location of the 
concrete base (black rectangle).