Skip to main content

Full text: Automatic detection of boulders by neural networks

3oulder detection | 
A 
sS55 
VBES BS 
MBES Slope 
bh 
As 
a 
BES Depth 
B SsQ 
ABES Depnth 
BES combined 
APES BS 
ABES combined 
= m A-A 
E vs 
— -18.0 50 5 
& S 
Ö -18.5 m 
— 0 
0 1 2 3 
Distance [m] 
— MN B-B' 
E -17.75 40 5 
— GI 
& 18,00 205 
En A 
—18.25 > + 
3 
2 
Distance [m] 
MBES Slopr 
E -19.50 I AN C-C'F 75 _ 
zZ 1: Y 
E Fr 5.0 5 
& 19,55 ; 2 
X . 2.5 
— 
1 2 
Distance [m] 
Pe D-D' | — 
45 
2 
E -19.550 
S _19,575 
WU 
o 
—19.600 
ü 
Z 
Distance [m] 
OWw RS hi. *- 
iR ar. 
m“ 
F 
012m 
3 
Fig. 3: The appearance of boulders in the different data sets. A) At a distance of 45 m to the nadir individual boulders are recognised in S5S backscatter. 
The same boulders (27 m to the nadir) are more difficult to recognise in MBES backscatter. The boulders are visible in bathymetry, slope, and combined 
data sets. B) Small boulders as imaged in the outer part (75 m to nadir) of a side-scan sonar swath. The characteristic boulder pattern is hard to recognise 
and appears smeared in the along-track direction, due to yaw movements or decreasing along-track resolution. The appearance of boulders is difficult to 
nterpret in MBES (20 m to nadir) backscatter, but the objects are recognised in slope, bathymetry and combined data sets. The position of SSS images was 
shifted bv several metres to account for positional differences to the MBFS. The green arrow points to the nadir. SSS data was recorded with a (MSS-annr 
Hydrographische Nachrichten
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.