Three dimensional mapping of aquatic plants at shallow lakes using
1.8 MHz high-resolution acoustic imaging sonar and image processing technology.
Katsunori Mizuno and Akira Asada （IIS. The University of Tokyo）
1. Motivation
2. Survey site
Aquatic plants play an important role in underwater ecosystems
and have an impact on the biological diversity of the world.
However, many aquatic species are decreasing in number [1].
A very accurate mapping and monitoring system to assess the
well-being and distribution of aquatic plants is needed.
Table 1. Current survey methods.
Methods
Diver
Satellite
Acoustic
Efficiency
☓
◎
◯
Turbidity
☓
☓
◎
Classification
◎
△
☓
In this study, we proposed a new observation method and an
image processing technique to assess the individual spatial
distribution of the aquatic plants in mixed community.
The field experiment was
conducted at the lake
Yunoko (36°47’ N, 139°
25’ E) in Japan from July
23 to July 25, 2013. At the
lake,
foreign
species
Elodea
nuttallii
and
endangered species Chara
globularis var. globularis
and Nitella flexilis var.
flexilis were seen. At some
areas, endangered species
and foreign ones made
mixed communities. The
habitat mapping of them
are needed.
3. Experimental settings
36˚48’ 18.86”N
139˚25’ 08.48”E
N
A survey boat was equipped with the
standard DIDSON (Sound Metrics, Bellevue,
WA, USA), a motion sensor (OS-5000US,
Ocean Server Technology, Massachusetts,
SA), and a DGPS (A100, Hemisphere,
Alberta, Canada).
Lake
Yunoko
Fig. 3 Survey boat
250 m
Fig. 1 Observation site
Fig. 2 Experimental set-up
Frame rate: 8 fps
Range: 5-10 m
Frequency: 1.8 MHz
Tilt angle: around 40 °
Pixel count: 96(H)✕512(V)
Beam width: 29(H)✕3(V)
4. Acoustic image processing
5. Classification and 3D mapping in mixed community area
6. Conclusions
Characteristics of leaves and rods of aquatic plants were
different between the species. It was seen in the acoustic images
as the differences of intensity and shape. The classification and 3D [2] mapping of aquatic plants methods based on spectrum
analysis with DoG (Difference of Gaussian) filtering and
difference of intensity were developed and tested in this study.
We applied the proposed acoustic image processing to high-resolution acoustic image (resolution: 5 mm ~ 2 cm)
and 3-D mapping of aquatic plants was reconstructed. The distribution of three types of aquatic plants including
two endangered species was visualized with the values of water depth and volume. Chara globularis was
distributed at shallow area (1 - 3 m), Elodea nuttallii was mainly middle area, and Nitella flexilis was deep area (5
- 7 m) but competing with Elodea nuttallii. Thus, this measurement system, based on high-resolution acoustic
imaging, can be useful for assessing the status of lakes and the distribution of aquatic plants.
The results of our study
demonstrate that imaging sonar can
be used to reconstruct the
underwater status of lakes. The
high-resolution acoustic images
obtained allowed us to identify
individual aquatic plants with high
accuracy. At the mixed community
area, our proposed processing
method could classify two species
of aquatic plants and help us to
understand the spatial distribution of
them.
DoG  G (u , v )  G 1  G 2 
1
2
 1  ( u 2  v 2 ) / 2 2 1  ( u 2  v 2 ) / 2 2  u, v: position
1
2 

 2e
e
  12
 σ: dispersion
2


Raw data
Background
subtraction
Bottom detection
(a) Before
2D FFT
(a) Chara globularis
Original data
References
(b) After
[1] Hooper, D. U, et al., Ecological Monographs, 75,
3–35.(2005)
[2] Chunhui, X. U, et al., J. Marine Acoust. Soc. Jpn.
40, 14-26. (2013)
Fig. 6 2D acoustic images
DoG filtering
2D iFFT
+
Classification
with threshold
Acknowledgement
✕w
w: weighting
factor
Motion correction
3D mosaicking
(b) Elode nuttallii
Fig. 4 Optical and Acoustic images
(a) Before
Fig. 5 Image processing flow
(b) After
Fig. 7 3D acoustic images
Fig. 8 Classification and 3D acoustic image
Part of this work was supported by the River Fund in
charge of River Foundation, Japan 251263006 and
JSPS KAKENHI Grant Number 25870153.
The authors gratefully acknowledge Tochigi
Prefectural Institute of Public Health and
Environmental Science, National Research Institute of
Aquaculture, National Federation of Inlandwater
Fisheries Cooperatives, and Nikko Yumoto Rest
House for their understanding in the significance of
this work and providing necessary collaborations in
the Lake Yunoko.