International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 6

Number 1

March 2012

149

1 INTRODUCTION

Concerning the problem of collision accident be-

tween large cetaceans and vessels, it is expected to

identify the critical season and species of large ceta-

cean in some courses of hydrofoils off Japan water

and also to technically improve the Under Water

Speaker (UWS) for repellent of cetaceans. However,

it is necessary to take measures of not only repellent

of cetacean but also actively cetacean avoidance.

The UWS is one of the countermeasures for ceta-

ceans to give way. For the safety navigation of hy-

drofoils, as part of these studies, we are envisaging

an alert system triggered by the whale detection us-

ing infrared camera (thermography). Because any

cetaceans are of air-breathing animals, they come to

the surface inevitably in regular interval and their

blow or body parts appear at the sea surface. In this

system, we aim at these objects as detection cue by

using infrared camera.

In this study, we examine the feasibility of the in-

frared camera on detecting large cetaceans. On that

basis, we intended to form the foundation of early

cetacean detection and alert system using infrared

camera by formulating the necessary conditions or

the method of infrared camera operation for detect-

ing whales.

2 MATERIAL AND METHOD

2.1 Equipments

In this experiments, we used a Infrared Thermal

camera; Thermo Tracer TH9260 made by NEC/Avio

Infrared Technologies (Fig. 1). Its detector is un-

cooled focal plane array (microbolometer), and

thermal image pixels is 640(H) × 480(V). The oper-

ating band is 8 to 13 μm, and it can measure objects

from -40°C to 500°C. Optical field of view of this

camera is fixed (21.7°(H)×16.4°(V)) . This camera

obtains the images 30frame per second.

Figure 1 Infrared camera used in this study (TH9260)

Feasibility on Infrared Detection of Cetaceans

for Avoiding Collision with Hydrofoil

Y. Yonehara, L. Kagami, H. Yamada & H. Kato

Tokyo University of Marine Science and Technology, Tokyo, Japan

M. Terada & S. Okada

KHI JPS Co., Ltd., Kobe, Japan

ABSTRACT: To achieve safer navigation without sudden collisions with large cetaceans at high speed boats

such as the hydrofoil, we examined its feasibility of an installation of the infrared camera. Because any ceta-

ceans are of air-breathing animals, it is theoretically expected that they can be potentially detected through

imaging of the infrared cameras. Thus, we examined the feasibility of detection with aiming at sperm whales

in waters off Chichijima Islands (27°4'N, 142°13'E), Japan. Through the experiment, it was revealed that

sperm whales could be detected stably within 200m, and detectable cue were blow, back body and fluke tails.

However, boats and waves were also detected as noise images. Especially, waves greatly resemble the whale

back bodies. Although potential of the infrared camera was confirmed, there are still necessities of further ex-

periments including ones conducting at different temperate waters, to successfully install the infrared camera

for earlier finding of large cetaceans.

150

Additionally, we used HD video camera (Sony

Digital HD video camera recorder HDR-SR1) for

recording normal images and thermal images in the

same time as a control of the experiment. For meas-

uring accurate distance between cetaceans and us,

we used the Laser range finder (Bushnell Laser

Rangefinder Elite 1500).

2.2 Experiment of infrared detection on sperm

whales in Bonin Water

An infrared recording experiment of sperm whales

was conducted off the South-East water off Chichi-

jima island, Ogasawara Islands (also known as the

Bonin Islands, a subtropical archipelago located ca.

1000 km directly south of central Tokyo in the west-

ern Pacific Ocean) from September 12 to 14, 2009

(Fig. 2).

We conducted this survey by chartered small fish-

ing boat ”Shoeimaru” (Fig. 3). At eye level was ap-

proximately 2m since TH9260 was held at the bow.

Captain and three researchers got on the boat and

participate in this survey. We intended to investigate

sperm whales mainly and record boats or sea surface

at the same time. We left Futami port every morning

and returned before sunset. The investigation time

was less than 10 hours a day.

Figure 2 A survey area off Chichijima island (Bonin Islands).

Gray square indicates survey area (26°55′ to 27°05′ N, 142°11′

to 142° 24′ E). Cross marks indicates the position sperm whale

sighted.

Figure 3 Syoeimaru; Length:12m,

After leaving port, we began a search of sperm

whales by eyes. When we found the sperm whale,

we tried to approach them and began the IR record-

ing experiment in diverse distance. A researcher

held the IR camera. Others served the distance

measurement and data notation, or filming the be-

havior of sperm whale and the research status by the

HD camera. When the blow or body of sperm

whales appeared above sea surface, we noted wheth-

er researcher could perceive it on the display of IR

camera (Success or Failure) with visual assistance of

sperm whale behavior made by other researchers and

an experienced operator of “Shoeimaru” in sperm

whale sightings around the waters. And we also not-

ed the time, type of the object and distance from the

object at the same time. We saved a thermal image

or thermal movie if at all possible. The distance

from the object was measured by use of Laser range

finder. If it couldn’t, researcher measured the dis-

tance with the eye.

Acquired infrared data were conducted an analy-

sis of temperature measurement using dedicated

software (NS9205Viewewr program made by

NEC/Avio infrared technology). We measured tem-

peratures of objects as T

and temperatures of sea

surface around the objects as T

sea

(Fig. 4) and calcu-

lated the difference in temperature (

⊿

T = T

- T

sea

Figure 4 The example of measuring method that of T

and T

sea

by using software (This is the magnified figure). In this figure,

point ‘a’ indicates the temperature of T

and ’b’ indicates that

of T

sea

3 RESULTS

We could conduct the experiment in three days. Ta-

ble1 shows the summary of the experiments. In this

experiment, we could observe and obtain data from

Sperm whale, small boat and sea surface. Show the

result of observation below.

Number of IR detection indicates the number of

data that we noted whether or not to perceive sperm

whales on IR camera in diverse distances. We ex-

cluded the data that noted same result (S or F), same

distance and within three minutes to avoid data bias.

However, it has potentially observed same individu-

151

al, other cases of data were dealt with the different

results since it is very hard to identify that the whale

examined previously or not. Number of thermal im-

ages indicates the whole number of saved thermal

images through the experiment.

Table 1 Summary of the experiment in waters off Ogasawara,

2010.

___________________________________________________

Date 12 Sept. 13 Sept. 14 Sept.

___________________________________________________

Observed whale groups 2 2 2

Number of IR detection 21 8 33

Number of thermal images 184 794 1371

___________________________________________________

3.1 Sperm whale

In this experiment, sperm whale’s blow, back body

and tail flukes were detected by using TH9260 (Fig.

5). As for object range, we could always detect these

sperm whale objects on the display of TH9260 with-

in a range of 150m. However, we never detect them

out of a range of 350m. Within a range of 160m to

300m, we could detect or not (Table 2 and Fig. 5).

The maximum distance of laser range finder was

118m.

Table 2 The result of IR sperm whale detection in each dis-

tance.

___________________________________________________

Distance (m) Success Failure Sum

___________________________________________________

～50 6 0 6

51～100 12 0 12

101～150 12 0 12

151～200 11 6 17

201～250 2 7 9

251～300 3 1 4

301～ 0 2 2

___________________________________________________

46 16 62

___________________________________________________

Figure 6 Numbers of success or failure in IR detection experi-

ment each distances.

Figure 5 Thermal images of sperm whale’s objects detected by

TH9260. Objects were indicated by arrows.

Top: blow (100m away), Middle: dorsal part of body (117m

away) and Bottom: tail fluke (200m away).

Concerning measured temperature, Median T

and

⊿

T of sperm whale obtained by TH9260 was

29.2°C and 1.30°C (Table 3 and Fig. 7). Median

⊿

T of each sperm whale’s detected objects were

1.10°C (blow), 1.40°C (back body), and 1.30°C (tail

fluke) (Table 4, Fig.8). As a result,

⊿

T of sperm

whale’s body parts indicates high value compared

with that of their blow, however statistical signifi-

cant difference didn’t shown in Tukey-Kramer’s

multiple comparisons (P=0.185) between

⊿

T of

sperm whale objects.

Distance (m)

failure

success

152

Table 3 T

and

⊿

T estimated in each objects.

___________________________________________________

of each objects (°C)

___________________________________________________

Whale (n=41) Boat (n=8) Wave (n=13)

Median 29.2 33.9 29.4

Max 32.2 38.8 30.3

Min 25.8 32.7 27.1

___________________________________________________

⊿

T of each objects (°C)

___________________________________________________

Whale (n=41) Boat (n=8) Wave (n=13)

Median 1.3 6.3 1.2

Max 2 11.3 1.5

Min -0.3 5.1 0.4

___________________________________________________

Figure 7 Box-plot of T

and

⊿

T of each objects

Vertical bar indicates the maximum and minimum value. Box

indicates lower and upper quartile point. Short crossbar indi-

cates median value.

Table 4 T

and

⊿

T estimated in each objects

___________________________________________________

of sperm whale's objects (°C)

___________________________________________________

Blow (n=19) Back (n=16) Fluke (n=6)

___________________________________________________

Median 28.8 29.15 29.65

Max 30.7 32.2 30.2

Min 25.9 27.1 25.8

___________________________________________________

⊿

T of sperm whale's objects (°C)

___________________________________________________

Blow (n=19) Back (n=16) Fluke (n=6)

___________________________________________________

Median 1.1 1.4 1.3

Max 1.9 2 2

Min -0.3 0.5 1.1

___________________________________________________

Figure 8 Box-plot of T

and

⊿

T of sperm whale’s objects

Vertical bar indicates the maximum and minimum value. Box

indicates lower and upper quartile point. Short crossbar indi-

cates median value.

3.2 Small boat

We could observe a small boat employed for whale

watching. The boat appeared very strong on the dis-

play of TH9260 since it has high rate of infrared ra-

diation (Fig. 9). Median T

and

⊿

T of small boat

was 33.4°C and 6.30°C (Table 3 and Fig. 7). It was

highest temperature in this study, and statistical sig-

nificant difference was showed in Tukey-Kramer’s

multiple comparisons (P<0.01) between

⊿

T of boat

and sperm whale or wave.

25,0

27,0

29,0

31,0

33,0

35,0

37,0

39,0

Whale Boat Wave

Temperature of T

(

℃

)

Objects

-2,0

0,0

2,0

4,0

6,0

8,0

10,0

12,0

Whale Boat Wave

Temperature of

⊿

T (

℃

)

Objects

25,0

27,0

29,0

31,0

33,0

35,0

37,0

39,0

Blow Back Fluke

Temperature of Ta (

℃

)

Objects

-2,0

0,0

2,0

4,0

6,0

8,0

10,0

12,0

Blow Back Fluke

Temperature of

⊿

T (

℃

)

Objects

153

Figure 9 A small boat detected by TH9260.

3.3 Sea surface

In the display of TH9260, concentric temperature

gradient always appeared on the sea surface (e.g.

Fig. 7 and 11). Waves such as white caps were also

detected on the display of TH9260 (e.g. Fig. 7). As

for temperature, median T

and

⊿

T of waves was

29.10°C and 1.10°C (Table3 and Fig. 7).

4 DISCUSSION

4.1 Detectable range

In this experiment, it is suggested that maximum

surely detectable range of TH9260 was 150m and

stable detectable range of it seems to be approxi-

mately 200m (Table 2 and Fig. 6). However, it is

considered that the measured instances aren’t

enough to conclude it. In early studies, McCafferty

et al. (2007) introduced that Infrared camera has the

potential for measuring mammals as far as 1000m or

more. Barber et al. (1991) could detect walruses on

the sea surface away from more than 2000m above

in the sky. Thus, infrared camera might be able to

detect large cetaceans at more long distance. The

possible way to improve the detectable range of in-

frared cameras is using telephotographic lens and

higher resolution infrared camera.

4.2 Ta and

⊿

T estimated in each objects

Cuyler et al. (1992) made a survey of large ceta-

ceans and taking their temperature using infrared

camera in the northern Norway water (68~80°N).

They resulted that

⊿

T of sperm whale’s blow and

tail fluke was +3.00°C and +6.00°C. The tendency

that body parts exceed blow in

⊿

T corresponds to

Cuyler et al. (1992). However, the

⊿

T of our sur-

vey showed lower temperature than that of Cuyler et

al. (1992). It is suggested that this caused by the dif-

ference of the regions that Cuyler et al. (1992) in-

vestigated in high latitude water (air temperature

was 2.5 ~13.0°C and water temperature was

2.7~10.1°C during investigation period), in contrast,

we investigated in subtropical water (air temperature

was 28.1~33.5°C and water temperature was

27.0~30.5°C during investigation period). Thus, it is

considered that surrounding temperature make a

large effect on cetacean detection using infrared

camera. If this idea is correct, it is suggested the ce-

tacean detection using infrared camera will be more

effective in cool winter season at the sea off Japan.

Especially, early detection of large cetaceans and

alert system using infrared camera is expected to be

more effective in the winter Sea of Japan which is

hard to find cetaceans by the naked eye.

In our experiment, we also could detect a small

boat, sea surface and waves through the thermal im-

ages as a noise of cetacean detection. The boat is

thought to be able to distinguish from cetacean ob-

jects easily on the thermal images since it was indi-

cated that the boat has large difference of tempera-

ture between sea surface and it emitted definitely

high thermal energy than other objects (Table 3, and

Fig. 5 and 9). Concerning concentric temperature

gradient on sea surface and the temperature indicat-

ed by waves, it is inconsiderable that they reflect the

real variation in water temperature. Though Infrared

has directionality, it is suggested that apparent varia-

tion in temperature was occurred by changing the

shooting angle between infrared camera and objects

of shooting. Additionally, the configuration and the

composition of waves were similar to these of the

sperm whale’s back body in thermal images (Figs. 5,

7 and 8, and Table 3, 4). Therefore, it seems to be

difficult to distinguish sperm whale’s back body

from waves in thermal images. It is suggested that

cetacean detection using infrared camera will be

hard to operate in heavy weather which waves ap-

pear a lot on the sea surface.

5 CONCLUSION

From this study, it was revealed that the infrared

camera has the feasibility to detect large cetaceans

on the sea around Japan. However, it is difficult to

say that we obtained a sure result for the detectable

range since the shortage of data. Therefore, it is nec-

essary to accumulate more data of detectable range.

It is suggested that oceanic condition and air or

water temperature affect the result of cetacean detec-

tion using infrared camera. Thus, we should conduct

more elaborate survey of environmental factor that

infrared cetacean detection works effectively. It is

also concluded that sperm whales are difficult to dis-

tinguish from waves in thermal images. Therefore,

we should devise the technique to distinguish them

not only by temperature.

154

In addition, early study, Cuyler et al. (1992) con-

ducted cool water and obtained more pronounced

difference in temperature between cetaceans and sea

surface than our study that conducted subtropical

water. Thus, summer Bonin island water might be

too warm to investigate the feasibility of infrared

camera to detect large cetaceans, and it is necessary

to make more practical verification surveys in an

environment similar to the course of hydrofoils off

Japan.

ACKNOWLEDGEMENT

This study was supported by KHI JPS Co., Ltd. and

Kawasaki Shipbuilding Corporation. We wish to

thank M. Yamaguchi and A. Nakajima for getting on

the boat and carrying out the experiment with us.

We are grateful to Ogasawara Marine Center and

Ogasawara Whale Watching Association for support

conducting the survey off Chichijima, Ogasawara

(Bonin Islands).

REFERENCES

Barber, D. G., Richard, P. R., Hochheim, K. P., Orr J. 1991.

Calibration of aerial thermal infrared imagery for walrus

population assessment. Arctic, 44, 58-65.

Cuyler, L. C., Wiolsrod, R., Oritsland, N. A. 1992. Thermal in-

frared radiation from free living whales. Marine Mammal

Science, 8(2): 120-134.

McCafferty, D. J. 2007. The value of infrared thermography

for research on mammals: previous applications and future

direction. Mammal Rev. 2007, Volume 37, No. 3, 207-233.

Shakata, K., Odagawa, A., Yamada, H., Matsunaga, H., Kato,

H. 2008. Toward to avoiding ship strike of cetaceans with

the high-speed HF (1) Identifying expected cetacean spe-

cies on the track lines of HF. 2008 Annual meeting of M. S.

J. summary: 135