VERIFICATION OF THE DIGITAL SEA BOTTOM MODEL

BUILT BY BATHYMETRIC DATA DEEP WATER STUDY

Assoc. Prof. Dr. Artur Makar1

1
Institute of Navigation and Hydrography, Polish Naval Academy, Gdynia,
Poland

ABSTRACT
Data that come from multi- and singlebeam echosounders make a basic source
for reliable and complex displaying of the seabed profile. Their quality and
reliability have significant impact on safe shipping and navigation. These are main
targets and purposes of the measurements. Thanks to detection and description of
many underwater objects, shallows or other dangerous places, the hydrographical
measurements do guarantee improvement of safe human activities at sea.
The measurement dynamics and variation of factors disturbing in time mean
that the bathymetric measurements are especially difficult and they require
experience in the process of planning, performing and elaborating the
measurements. The geospatial data obtained for the need of creating a numerical
model of the seabed must be reliable because it is not possible to compare them
with any standard.
There is a method of verification of the geospatial data obtained by means of
an MBES multibeam echosounder used in bathymetric deep-water maritime
measurements presented in the article. Dynamic calibration, being a preliminary
stage of the echosounding works, has been applied in the process of the model
verification. Reference to DTM (Digital Terrain Model) verification in
photogrammetry has been done where ground points obtained in land survey make
references.
Keywords: Digital Terrain Model, Digital Sea Bottom Model, multibeam
echosounder, calibration, verification

INTRODUCTION
Hydrographical survey is of great significance in securing underwater works
among which the following ones may be underlined: routing submarine cables,
construction of pipelines or dredging of waterways. The seabed relief may change
due to inflow of sediments. It is necessary to make systematic surveys on sea routes,
fairways and other sea areas.
Numerical Terrain Models (NTMs) make one of the most important subject
matters of Geographical Information Systems these days. They are applicable in
rendering two or three dimensions of the seabed and in checking its shape. Methods
of creating the NMTs are greatly various hence they differ in accuracy. This is to
note that the seabed is an environment changing dynamically, so it is recommended
to elaborate methods allowing verification of the models already existing.
 Character of performance of bathymetric survey (dynamic measurements,
waving, variability of conditions for acoustic wave propagation in water) makes
execution of many measurements or a long-term measurement in the same place, in
order to strengthen accuracy or to verify the model, impossible [8], [9], [10]. As far
as the method is concerned, the measurements are similar to photogrammetric
estimations (dynamics, spatial orientation of an airplane) but they differ in respect
to verification, due to lack of reference. In photogrammetry, it is possible to refer
to characteristic terrain points determined with geodesic methods applied.

METHODS OF DATA ACQUISITION AND SOURCES OF

ERRORS
Aerial photogrammetry and low bank UAV (unmanned aerial vehicles) are the
main data sources for building DTM (digital terrain model) of terrestrial areas. The
photogrammetric works mainly consist in measuring homologous points on two or
more pictures (photos). In digital photogrammetry, algorithms of automatic
measurements on digital images are mostly based on methods of correlation.
Correlation coefficient is a basic benchmark of similarity (goal function) [11], [13].
Having the homologous points of pixel coordinates been determined, it is possible
to translate them into photogrammetric coordinate system and then, based on the
known elements of reciprocal and absolute orientations, to determine terrestrial
[7].
The following factors, among the others, have impact on accuracy of the
photogrammetric elaboration [1], [11], [13]:
proper calibration of photogrammetric cameras,
atmospheric refraction,
contraction of film in the period between exposition and scanning,
mechanical and optical errors of photogrammetric scanners,
precision in identification of photo-points and accuracy in measuring
them on pictures and in the field,
pure correlation of the picture
process of automatic measurement of the points.
In the hydrographical survey, where bathymetric measurements make primary

elaboration, the main factors that influence accuracy of the determination are as
follow:
positioning of the echosounding ship,
considering vertical distribution of sound velocity in water,
sonar calibration, mainly dynamic calibration of multibeam
echosounder,
 disruptions of movement of the echosounding ship compensated by
measurements of her roll, heave and pitch motions by means of
instruments serving determination of spatial orientation,
considering variations in water level.
With a use of the multibeam echosounder, the least distortions resulting from
the acoustic wave propagation in water from occurrence of refraction
phenomenon, appear under converter for beams the least tilted from the vertical [8,
9, 10]. The distortions arise for the beams having larger and larger output angle and
they are the most significant on an edge of the stream [4], [5], [12]what has an

VERIFICATION OF THE DIGITAL TERRAIN MODEL

Surface described by the DTM is only an approximate of the Earth real surface.
Difference between these two surfaces is defined as [11], [13]:
(1)
and the mean error of the difference as [11], [13]:

(2)

determine difference between the generated DTM (approximate surface of the

terrain) and the real area of the surface. Different, more accurate (by an order of
magnitude) model of surface of the same area must be used in order to verify quality
of the given DTM.
The proposed [11], [13]

a model we want to test there and a difference of elevation between the two models,
defined for enough number of points.
The first step is to check whether the random variable (difference between the
models on the selected area of the elaboration) for this sample represents normal
distribution.
Then, elements of the statistical test are being defined [11], [13]:
We hypothesize that the parameter m amounts to the value of m0.
Thereby, we shall verify the hypothesis H0 in the form m = m0.
We also define an alternative hypothesis m>m0 (or HA: m<m0)
We perform the statistical test determination of
which allows assessing whether we may reject a null hypothesis.
The real surface on ground may be determined based on land measurements,
with an accuracy of 1 cm, executed by means of satellite measuring systems. They
are also used in securing measurements being taken by UAVs [2], [14] to determine
georeferenced points.
In hydrography, determination of the sea-

difficult and even not possible. Therefore, there are geodesic positioning systems
and methods of echo- to obtain the best possible

MBES CALIBRATION AS A TOOL INCREASING

RELIABILITY OF DTM DETERMINATION

time before and after execution of the bathymetric survey in order to check
correctness of operation of the system and devises. The cross check consists in
performing two measurements on profiles perpendicular to one another.
Measurements on extreme outer beams, burdened with the greatest error (red
points), with vertical indications (green points) vitiated by the smallest error (Fig.
1). Difference in the indications

Fig. 1. Calibration of multibeam echosounder

uncertainty of measurement taken by means of MBES. This is a result of sound

velocity changes in water. Field of acoustic wave emitted in water environment by
the echo- -domain may be described
mathematically with solving wave equation [3], [6], which in hydroacoustics is a
basis for descriptions of small amplitude waves. This solution does exist, however
only for simple or idealized cases. In general, a point source of the acoustic wave
and horizontal stratification of water in which the sound speed c(x,y,h) = c(h)
are assumed.
 In short, dete

the sound wave propagation [4], [5]. The sound radius shall deflect towards lower
velocity of sound. The acoustic ray deflection phenomenon is the most visible for

radiation characteristics implement the greatest errors in depth measurement.

Fig. 2. Influence of the refraction phenomena (sound speed in water) on

determination the seabed.

RESEARCH
Geospatial data recorded during calibration of the multibeam echosounder for
rolling correction on waters of Gdansk Bay (Fig. 3)were used to verify the Digital
Sea Bottom Model. The calibration preceded hydrographical soundings executed
by hydrographic ship by means of R2Sonic 2022 multibeam echosounder, operating
at the 300 kHz frequency. The hydrographic system was also constituted by: two-
antenna Applanix POS MV precise positioning system, Applanix detector of roll
and heave motions, Valeport Mini sound velocity profiler, DESO-30 singlebeam
echosounder. They all were integrated by QPS QINSy hydrographic system.
 Fig. 3. Area of the research

Fig. 4. 3D and cross-section visualizations of crossed profile and the base

In the process of the seabed model verification, the depths were compared at
the stage of raw data for the same narrow area of the seabed. The depths obtained
from the survey on the entire width of the swatch with a sector searching one head
in sector 87 were analysed with depths obtained by means of the echo-
vertical beam during passage of the vessel perpendicular to the given profiles.
Values of the depths were acquired in VALIDATOR application, with a 2m
interval, what as a consequence provided over 100 measurement points for each
profile.

30
Difference between

25 Profil
e1
depths [cm]

20
15
10
5
0
-120 -100 -80 -60 -40 -20 0 20 40 60 80 100 120
Linear deviation of the beam [m]

Fig. 5. Error determination of the depths with relations to linear deviation od

the acoustic beam
Tab. 1. Differences between depths in crossed profiles

Number of Differences between depths

profile min [cm] max [cm] average [cm]
Profile 1 1 28 13
Profile 2 1 28 12
Profile 3 1 27 12

CONCLUSIONS
These days, systems of multibeam echosounder pay a serious role in the field

The presented results regarding evaluation of the model quality have been obtained
based on data collected from deepwater survey this reduces, in significant way,
the requirements concerning accuracy imposed by IHO, as well as the requirements
resulting from principles for charting depths onto marine navigational maps. This
criterion has been met in all measurement points of the test on the entire width of
the swatch. This has been possible thanks to the method of executing the
measurements, adjustment of the ship speed to the conditions, skills of the
helmsman, consideration of the sound velocity in water, as well as calibration of the
shipboard hydrographical system, including: the positioning system, detector of the
roll and heave motions and others used in the test.
Findings obtained as a result of the performed analysis show that the topic
regarding verification of DTM, built as a result of charting bathymetric data coming
from multi-beam sonar, is broad. Accuracy of those measurements greatly depends
on accuracy of additional sensors and Motion Reference Units. The test results do
show that the angular coverage sector is limited due to serious errors being
generated by extreme outer beams even with a use of two heads to map the sea-

maximal and effective width of the seabed coverage sector.

ACKNOWLEDGEMENTS
The author would like to thank the headquarter and officers of Hydrographic
Support Squadron of the Polish Navy for permitting participation in soundings and
using hydrographic equipment and recorded data.

REFERENCES
[1] Ackermann F., Technique and strategies for DEM generation. Digital
photogrammetry: an addendum to the manual photogrammetry. American Society
for Photogrammetry and Remote Sensing, USA, s. 135-141, 1996.
[2] Banaszek A., Zarnowski A., Cellmer A., Banaszek S., Application of New
Technology Data Acquisition Using Aerial (UAV) Digital Images for the Needs of
Urban Revitalization.
Vilnius Gediminas Technical University, Lithuania, 2017.
[3] Brekhovskikh L.M., Lysanov Yu.P., Fundamentals of Ocean Acoustics.
Springer-Verlag, New York Berlin Heidelberg, 2014.
, Depth measurement accuracy for outer part of the
multibeam echosounder swath width, Logistics, pp. 2246-2251, 2014.
[5] Estimation of effective swath width for dual-head
multibeam echosounder, Annual of Navigation, vol. 23, pp. 173-183, 2016.
[6] Klusek Z., Propagation Conditions of the Sound Speed in the Southern
Baltic. Polish Academy of Science, 1990.
[7] Krzystek P., Generation of Digital Elevation Models. In Second Course in
Digital Photogrammetry, Bonn, Germany. Institute for Photogrammetry at Bonn
University and Landesvermessungsamt Nordrhein-Westfalen, 1995.
[8] Makar A. Sea bottom surface described by Coons pieces, Scientific Journals
of the Maritime University of Szczecin, Poland, 45 (117), 2016, pp. 187-190
[9] Makar A. Cleaning of hydrographic data for determination the seabed of
waterways. 7th International Conference on Maritime Transport Technological,
Innovation and Research Barcelona, Spain, pp. 213-219, 2016.
[10] Makar A. Cleaning of MBES Data Using CUBE Algorithm. 17th
International Multidisciplinary Scientific GeoConference SGEM2017 Albena,
Bulgaria, Volume: 17 (21), pp. 841-847, 2017.
[11] of the Digital Terrain Model,
Archives of Photogrammetry, Cartography and Remote Sensing, vol.16, 2006, pp.
457-466.
 [12] Snellen M., Siemes K., Simons D., An efficient method for reducing the
sound speed induced errors in multibeam echosounder bathymetric measurements,
Delf University of Technology, The Netherlands, 2009.

numerycznego modelu terenu, Wydawnictwo Polskiego Internetowego Informatora

Geodezyjnego, seria GEOMATYKA, 2016.
[14] Zarnowski A., Banaszek A., Banaszek S., Application of Technical
Measures and Software in Constructing Photorealistic 3D Models of Historical
Building Using Ground-Based and Aerial (UAV) Digital Images. Reports on
Geodesy and Geoinformatics vol. 99 /2015; pages 54-63, 2015.