Underwater three-dimensional imaging
with an amplitude-modulated laser radar
at a 405 nm wavelength
Luciano Bartolini, Luigi De Dominicis, Mario Ferri de Collibus, Giorgio Fornetti,
Massimiliano Guarneri, Emiliano Paglia, Claudio Poggi, and Roberto Ricci
We report the results of underwater imaging with an amplitude-modulated single-mode laser beam and
miniaturized piezoactuator-based scanning system. The basic elements of the device are a diode laser
source at 405 nm with digital amplitude modulation and a microscanning system realized with a smallaperture aspheric lens mounted on a pair of piezoelectric translators driven by sawtooth waveforms. The
system has been designed to be a low-weight and rugged imaging device suitable to operate at medium
range 共⬃10 m兲 in clear seawater as also demonstrated by computer simulation of layout performance. In
the controlled laboratory conditions a submillimeter range accuracy has been obtained at a laser amplitude modulation frequency of 36.7 MHz. © 2005 Optical Society of America
OCIS codes: 110.0110, 280.340, 010.3920.
1. Introduction
The development of laser scanning systems for underwater imaging is a subject of remarkable interest
in view of their potential applications in several fields
ranging from submarine archaeological sites visualization to background inspection for industrial and
scientific purposes. Nevertheless, the task is challenging because light absorption and scattering, in
undersea applications, act to degrade the image quality. The effects of light absorption can be in principle
minimized by selecting the laser wavelength in a
transmission spectral window of water. Nevertheless, seawater composition, as determined by dissolved contaminants, strongly affects its absorption
spectrum making it difficult to unequivocally determine an optimal working wavelength. Currently,
most of the underwater laser imagers are based on a
source-emitting radiation in the green region of the
spectrum 共⬃532 nm兲, corresponding to a minimum of
absorption of turbid seawater characterized by a rel-
The authors are with the Ente Nazionale Energie Alternative
Advanced Technology Division, Via Enrico Fermi 45, 00044 Frascati(Rome),Italy.Thee-mailaddressforL.DeDominicisisdedominicis
@frascati.enea.it.
Received 15 June 2005; revised manuscript received 25 July
2005; accepted 26 July 2005.
0003-6935/05/337130-06$15.00/0
© 2005 Optical Society of America
7130
APPLIED OPTICS 兾 Vol. 44, No. 33 兾 20 November 2005
ative chlorophyll abundance. On the other hand, several experimental schemes have been proposed and
realized to minimize the detrimental effects of scattered laser light. Use of pulsed lasers allows both for
temporal discrimination of scattered light and for
target range information by time-of-flight measurements.1,2 In a system with cw and amplitudemodulated (AM) laser sources, the bistatic optical
layout, which limits the transmitter and receiver
common field of view, is the most adopted geometry to
reduce the stray light associated with backscattering
in water.3 Although in AM devices the target range
measurement is straightforward, because the information is directly stored in the reflected signal phase,
they have attracted poor attention in underwater applications because of the need for mirror-based scanning cameras that are usually too slow and heavy.
The potential of a bistatic AM device for underwater imaging has been explored and demonstrated by
Mullen et al.4 in turbid water (beam attenuation coefficient ⬎1 m⫺1) at short ranges 共⬃3 m兲 with an electrooptic-modulated Nd:YAG laser and with a bistatic
configuration. Seawater is turbid and has a high concentration of chlorophyll only in proximity of coasts,
but it approaches clean water features in open sea.5 It
follows that in open sea conditions both a considerable
reduction of light backscattering and a blueshift of
water absorption minimum are expected. For an AM
imager suitable to operate in clean seawater, the
bistatic geometry condition can thus be in principle
weakened without severely compromising the device
performance. In addition, the recent availability of a
compact diode at 405 nm with the possibility of digital
modulation at radio frequency (RF) allows for the best
match of the absorption minimum of clear seawater.
These considerations open up a question about the
possibility of developing compact, near-bistatic AM devices for quantitative imaging in clean seawater conditions and operating at medium range 共⬃10 m兲.
In this paper we describe a new AM miniaturized
underwater optical radar equipped with a piezotranslator-based scanning system and very small
transmitting and receiving optics. The layout of the
apparatus results in a low-weight and compact threedimensional (3D) viewing system suitable for metrology and inspection of underwater archaeological
reports in environments with hostile access. The basic element of the AM imaging device is a diode laser
emitting 20 mW of radiation at 405 nm, near the
minimum of absorption of pure water, driven with a
RF that modulates the intensity up to 100 MHz. To
keep the system compact, in view of future applications in environments with hostile access, the AM
imaging device has been equipped with a microscanning system based on a short focal-length lens
mounted on piezoelectric translators. In the following
sections the basic principle of operation of an AM
laser radar together with the experimental apparatus will be described. The performance of the device,
in terms of maximum operative range (MOR), is then
theoretically investigated for clean seawater by
means of a properly developed calculation code. In
the experimental results section a statistics analysis
of device stability and accuracy is reported together
with the demonstration of short-range underwater
imaging.
2. Underwater Amplitude- and Phase-Modulated
Laser Radar
The working principle of an AM laser radar is based
on the indirect determination of the round-trip time
delay of the AM laser beam through the demodulation of the phase signal ⌬␾ of the carrier with respect
to a reference signal. The target range d is simply
determined by the formula
d⫽
c⌬␾
,
4␲n␯m
(1)
where ␯m is the modulation frequency, c is the velocity
of light in vacuum, and n is the index of refraction of
water. The range noise ⌺ affecting a measurement
with a scanning laser radar depends on the modulation depth m and the current signal-to-noise ratio
SNRi through the following formula derived by
Nitzan et al.6:
⌺⫽
c
冑2nm2␲␯mSNRi
,
(2)
Fig. 1. Scheme of the AM laser radar. MMS, microscanning system; L1, lens ( f ⫽ 2 mm); L2, lens ( f ⫽ 2.5 mm); PMT, photomultiplier tube.
which shows the dependence of the image resolution
on the modulation frequency ␯m. The scheme of the
underwater AM laser radar realized in the laboratory of Ente Nazionale Energie Alternative (ENEA)
Frascati is shown in Fig. 1. The sounding element of
the device is a diode laser (LG Laser Technologies,
Blu Photon) emitting 20 mW of radiation at ␭ ⫽
405 nm. By driving the bias current with a pulse
train, the laser amplitude can be modulated up to
␯m ⫽ 100 MHz. Laser output is coupled to a 3 m long
single-mode fiber with a core diameter of 4 ␮m. Insertion losses and fiber attenuation results in 11 mW
of peak power at the fiber output. The operation of the
scanning system is based on the dynamic displacement of the output lens in front of the fiber core set in
its focal plane. The elements of the scanning system
are a short focal-length aspheric lens 共 f ⫽ 2 mm兲,
with 2 mm of aperture, mounted on a pair of mutually
orthogonal linear piezoelectric translators 共LINOS
PX5-400兲. The lens acts simultaneously as a focusing
element of the sounding beam and the scanning device as the lens translation results in off-axis operation. The piezotranslator travel distance is 400 ␮m
corresponding to a full scanning angle of 10°. The
maximum scanning rate attainable with this
system is 4 ⫻ 10⫺3 deg s⫺1. The target to be visualized
is immersed in a 1.58 m long water tank equipped
with an antireflection-coated optical window and
filled with water from the distribution network. The
reflected signal from the target is collected by a short
focal-length lens 共 f ⫽ 3 mm兲 with a diameter of 5 mm
and focused onto the photocathode of a fast photomultiplier (Hamamatsu HP8153). The separation between the transmitter and the receiver optics is 3 cm,
resulting in a near-bistatic configuration that is effective in rejecting the optical noise coming from the
laser reflection at the tank entrance window. A
lock-in amplifier (Stanford Research) provides both
the modulation frequency ␯m to the diode laser and
the signal phase delay measurement. Synchronization between scan and acquisition is obtained by PC
control of the piezoelectric driver and lock-in amplifier by means of a general-purpose interface bus communication protocol.
20 November 2005 兾 Vol. 44, No. 33 兾 APPLIED OPTICS
7131
3. Modeling Device Performance
The performances of a near-bistatic AM underwater
imaging device have been simulated by means of a
calculation code called RDRSB, developed by the authors, which evaluates the angular backscattering in
the Henyey–Greenstein approximation.7 After having properly set the optical parameters characterizing the device shown in Fig. 1 and the water optical
properties, the code calculates the signal S and straylight N optical power at the detector photocathode as
a function of target range. The MOR of the device is
then defined as the target range, yielding S兾N ⫽ 1.
The model does not include multiple scattering and
interference effects.4 The extinction effect of absorption and scattering processes on a laser beam of intensity I0 are gathered in the coefficients a and b,
respectively. The laser beam intensity after a path of
length l in water is given by
I共l兲 ⫽ I0 exp关⫺共a ⫹ b兲l兴.
(3)
The dependence of the a and b coefficients on the
wavelength is governed by several factors, such as
the abundance of contaminants, water temperature,
and salinity. Contaminants diluted in water do not
exhibit absorption in the spectral region of interest;
the parameter a for clean seawater can in principle
approach the value of 0.006 m⫺1 at 405 nm as reported for pure water absorption in Ref. 8. This assumption holds mainly for off-coastal seawater,
where the absence of diluted organic materials does
not promote chlorophyll proliferation9 with a subsequent greenshift of the water absorption minimum.
Seawater composition also affects the value of b.
Among the enormous amount of data available in the
literature, it has been possible to identify in 3 ⫻
10⫺2–1.0 m⫺1 the range of values for b at 405 nm
characterizing clean open ocean seawater.10,11 To account for backscattering contribution to noise, the
water column illuminated by the laser beam has been
divided into infinitesimal volumes, each one acting as
a scattering source. The scattering intensity due to
the infinitesimal volumes, as viewed by the receiving
optics, has been integrated along the laser beam path
to the target. The optical noise contribution due to
optical window scattering has been neglected in view
of the assumed near-bistatic configuration. The receiving optics aperture and focal length have been set
equal to 5 mm and 3.1 mm, respectively. The target
is assumed to be a Lambertian plane surface with a
reflectance of 0.35%, the diameter of the focused laser
beam on target is set to 4 mm, and the detector–
window distance is Rf ⫽ 20 cm. The scattering of the
light in the seawater is assumed to follow a simplified
Henyey–Greenstein11 model with a phase function
P共␪兲 for backscattering given by
P共␲兲 ⫽
7132
1 1⫺g
,
4␲ 共1 ⫹ g兲2
(4)
APPLIED OPTICS 兾 Vol. 44, No. 33 兾 20 November 2005
Table 1. Summary of System Parameters for Performance
Device Simulation
Parameter
Laser wavelength
Laser power
Modulation frequency
Spot size on target
Source–receiver separation
Aperture of receiving optics
Target reflectivity
Pixel sampling time
Water extinction coefficient
Value
405 nm
20 mW
36.7 MHz
4 mm
30 mm
5 mm
0.35%
10 ms
0.0571–0.95 m⫺1
where g is the asymmetry parameter describing the dependence of the backscattered power on the particle size
共0 ⱕ g ⱕ 1兲. The results of the device performance
simulation (Table 1) for nearly pure seawater,11,13
with refractive index n ⫽ 1.33, b ⫽ 0.0426 m⫺1, a
⫽ 0.0145 m⫺1, particulate volumetric fraction Fv ⫽
10⫺6, and g ⫽ 0.98, are shown in Fig. 2. In the simulation the laser intensity has been assumed as I0
⫽ 20 mW before single-mode coupling. The simulation indicates that most of the optical noise due to
water backscattering comes from the initial stage of
the illuminated water column while the signal intensity is a decreasing function of target distance with
an estimated attenuation coefficient of 1.9 dB兾m. The
RDRSB code indicates that the backscattered optical
power on the detector equates the signal optical
power at a target distance MOR of 34 m, which corresponds to a field of view of 5° for the apparatus
described in Fig. 1. At short range 共⬃2 m兲, which is a
typical target distance in our experimental setup, the
model predicts a signal-to-noise ratio of nearly 104.
The developed model demonstrates that, because of
the low values assumed for water absorption and
scattering coefficients, the near-bistatic configuration
allows for an efficient backscattering noise rejection
Fig. 2. Dependence of the signal (solid curve) and the stray-light
(dashed curve) intensity as received by the detector on the target
range. The MOR is defined as the range at which the curves
intersect.
Fig. 3. Dependence of the MOR on the water scattering coefficient
for a ⫽ 0.0145 m⫺1.
from the receiving optic field of view. The scattering
coefficient b is effective in drastically reducing the
MOR of the simulated device as shown in Fig. 3. In
this simulation the MOR dependence on the water
scattering coefficient has been calculated for b in the
range of 0.0426–0.95 m⫺1.
4. Imaging Experimental Results
Laboratory tank experiments were conducted to demonstrate the possibility of underwater imaging with
the AM laser radar scanning device. The tank was
filled with water from the distribution network, and
the target used for 3D visualization was a steel surface with well-calibrated graduated 共1 cm兲 steps. The
target was uniformly coated with a high-reflectivity
paint resulting in a constant reflectance over the target
surface. Starting from long target sampling times necessary to obtain good images with a relatively low laser
beam power, it was found that the stability of the
overall apparatus, including electrical drifts and thermal fluctuations of the water sample, plays a key role
in the determination of the image quality. The best
performance of the apparatus is obtained if the fieldof-view scanning time T does not exceed the maximum
value Tmax as given by the Allan variance statistic.14
To determine Tmax for our apparatus, we acquired
time-series data consisting of phase measurements
taken at a fixed range without scanning the target
with a pixel acquisition time ␶pix ⫽ 10 ms and with a
laser modulation frequency ␯m ⫽ 36.7 MHz. The results of the Allan variance test on time-series data
are shown in Fig. 4. It can be seen that up to Tmax
⬃ 80 s the Allan variance decreases almost linearly.
Beyond this optimum acquisition time, the Allan
variance deteriorates with a linear slope, thus denoting a time regime dominated by white-noise fluctuations. As mentioned before, the Allan plot allows one
to set an upper limit to the image acquisition time. In
our controlled laboratory conditions, an acquisition
time, as determined by the system scanning rate,
considerably longer than 80 s is expected to introduce
Fig. 4. Allan variance plot of time-series data collected with ␶pix
⫽ 10 ms.
significant errors in image reconstruction. It is worthwhile to note that in a real situation, Tmax is mostly
determined by water temperature and density fluctuations instead of experimental device electrical
drifts. In particular, a significant decrease for Tmax is
expected as temperature gradients and water currents dominate the seawater thermodynamic conditions. In this sense Tmax, as determined by real
conditions, imposes strict limits to the system resolution (i.e., number of pixels) and scanning rate.
In a subsequent preliminary laboratory experiment, the precision and the accuracy of our apparatus
were studied by collecting time-series data of phase
measurements at a fixed range for a time Tmax. The
target was located at the bottom of the water tank at
a distance from the optical window inner face of
1.52 m (position 1) as determined with a ruler. The
time-series data were taken with a pixel acquisition
time ␶pix ⫽ 10 ms. A calibration procedure consisting
of phase measurement with the target next to the
optical window inner face (position 2) allows for target range estimation. The statistics on the timeseries data provided a mean value of 179.010° for the
phase difference between positions 1 and 2, with a
standard deviation of 0.03°. Target range d estimation by use of Eq. (1) with n ⫽ 1.33 gives the result
d ⫽ 1.526900 m ⫾ 253 ␮m. With a differential
method, the device estimates the target distance
within the precision of the ruler and with an accuracy
of nearly half of a millimeter.
To explore the potentiality of the device for underwater quantitative imaging in our controlled laboratory conditions, a surface of 10 cm ⫻ 3.8 cm of the
immersed target was scanned after suitable focusing.
The investigated field of view corresponds to a full
angle of a scan of ␪ ⫽ 3.73° and ␸ ⫽ 1.14° for horizontal and vertical directions, respectively. The target was scanned with steps of 0.046° for both angles,
resulting in a frame of 80 ⫻ 40 pixels. The full image
acquisition time was T ⫽ 32 s, less than Tmax as determined with the Allan variance test. In Fig. 5 we
show the results of a 3D image reconstruction of the
20 November 2005 兾 Vol. 44, No. 33 兾 APPLIED OPTICS
7133
Fig. 5. Results of the 3D image reconstruction of the target immersed in the water tank. The dimensions of the field of view are
10 cm ⫻ 3.8 cm.
target scanned at ␯m ⫽ 36.7 MHz. Different angles of
view of the postprocessed data are reported. The software for data processing and image reconstruction
has been properly developed under an Interactive
Data Language environment and its main features
are described in Ref. 15.
The device reproduces the target topology with
good accuracy despite smoothing effects due to the
spot-size diameter that are clearly visible mainly at
the different steps of the steel staircase. Because the
whole image was recorded with an acquisition time
T ⬍ Tmax, the collected data are expected to provide a
range estimation within the accuracy of the apparatus as determined with the Allan variance test. The
quantitative analysis of a subset of data with ␪ ranging in the field of view at constant ␸ is reported here.
The subset of data is visualized in Fig. 6. Each step in
Fig. 6 is a constant range line after the polar reformatting of raw data.16 The standard deviation of
phase measurements within a step is nearly equal to
Fig. 6. Subset of data at fixed ␸ taken for ␪ ranging in the field of
view. The different steps are numbered for comprehension.
7134
APPLIED OPTICS 兾 Vol. 44, No. 33 兾 20 November 2005
0.03°, in agreement with the accuracy analysis reported above. Following the notation for the step labeling in Fig. 6, if Pj is the averaged phase for step j,
the quantity Aj ⫽ Pj ⫺ Pj⫹1 should be equal to 1.184°,
corresponding to a distance of 1 cm at our operative conditions, for each j. Experimental findings
furnished for 兵Aj其j⫽1.4 the values 1.25°, 1.17°, 1.24°,
and 1.15°, respectively. The experimental findings
indicate that two measurements overestimate the
step of 590 and 500 ␮m, respectively, whereas the
remaining data show an underestimation of 118
and 287 ␮m. All the step width estimations, based on
a differential method, are within the accuracy range
of the apparatus as determined by the previously
described statistical analysis. The optical properties
of the water from the distribution network used in
the present experiment are not available and have
not been investigated in our laboratory; nevertheless
it has been certified to contain a certain amount of
organic and inorganic materials as well as metal particles coming from pipes. All these constituents act as
scattering centers thus making plausible the assumption that the scattering coefficient of water from
the distribution network approaches a value close to
clean seawater rather then to pure water.
5. Conclusions
An AM single-mode laser-based optical radar has
been developed for underwater 3D imaging. The utilization of a diode laser at 405 nm, with RF amplitude
modulation carried through the control of the polarizing current, constitutes an innovative aspect in the
field of AM underwater imaging. The diode laser wavelength matches the minimum of the pure water absorption spectrum, thus making the apparatus well
suited for image reconstruction at clean seawater conditions as also demonstrated by a properly developed
computer simulation of device performance.
Laboratory tests in the controlled environment
conditions and on water from the distribution network allowed us to estimate in 500 ␮m the accuracy
of a differential target distance measurement at a
range of nearly 1.5 m. Nevertheless, quantitative
range determination in real situations (i.e., submarine operation) strongly relies on the exact knowledge
of the water scattering and absorption coefficients
and the index of refraction, which for seawater is a
function of pressure, temperature, and salinity. The
role played by the uncertainty of the exact thermodynamic and chemical seawater parameters deserves
further discussion. The variation of n with pressure,
and hence with depth, has been demonstrated to be
negligible (less than 10⫺4) up to a depth of 100 m
(pressure ⬃10 kg兾cm2). Data available in the literature16,17 concerning variation of the index of refraction with temperature and salinity allow one to
estimate their influence on range determination
accuracy.
By using Eq. (1), with ␯m ⫽ 36.7 MHz and assuming zero salinity, it is possible to quantify, at 15 °C,
in 0.7 ␮m °C⫺1 deg⫺1 the range indetermination due
to temperature uncertainty. For salinity, the calculations at a temperature of 15 °C estimate the error introduced by salinity uncertainty in 1.5 ␮m deg⫺1 ‰,
for a variation of a part per thousand in salinity. It
turns out that for our controlled laboratory conditions
the error in range estimation introduced by temperature uncertainty is much less than the error
introduced by electrical drifts. Nevertheless, at a target distance of nearly 1.5 m, corresponding to ⌬␾
⬃179° at ␯m ⫽ 36.7 MHz, the error introduced by an
uncertainty of 1 °C for temperature and of 1‰ for
salinity is nearly 400 ␮m, less than, but of the same
order of, the limits in system accuracy due to intrinsic
instabilities. In real operative conditions, the performances of the device, in terms of range accuracy, are
then expected to be governed both by local fluctuations of seawater properties and by intrinsic device
instabilities.
To demonstrate the potentiality of the apparatus to
record a 3D image, a structured field of view of
10 cm ⫻ 3.8 cm has been recorded with an acquisition time T ⫽ 32 s. The 3D image reconstruction
performed by means of suitable developed software
has been demonstrated to reproduce target topology
with good fidelity and resolution. Nevertheless, the
accuracy and the spatial resolution of the apparatus
in real situations are expected to be limited by seawater local thermodynamic fluctuations. In fact, as
discussed previously, local fluctuations act to reduce
the maximum acquisition time Tmax as given by the
product of the number of image pixels Npix and the
pixel acquisition time ␶pix. Because it is not possible to
reduce the pixel acquisition time under the signal
detection limit, the only way to maintain T less than
or equal to Tmax is to reduce Npix.
It must be noted that in view of Eq. (2) the range
noise ⌺ is inversely proportional to ␯m, so a better
range accuracy can be obtained by increasing the
laser modulation frequency. This is the subject of
present investigations in our laboratory in view of the
recent availability of diodes at 405 nm with the possibility of analog modulation up to 350 MHz and
power up to 60 mW. Future experimental investigations will also be addressed toward studies of device
performance at medium 共⬃7 m兲 and long 共⬃20 m兲
range in a 25 m length test tank equipped with optical windows and operating at ENEA Frascati. For
long-range determination, to avoid range uncertainity due to phase periodicity, the system will be
upgraded to operate with a double-modulation
configuration.
References
1. G. R. Fournier, D. Bonnier, J. L. Forand, and P. W. Pace,
“Range-gated underwater laser imaging system,” Opt. Eng. 32,
2185–2190 (1993).
2. J. W. McLean, “High-resolution 3D underwater imaging,” in
Airborne and In-Water Underwater Imaging, G. D. Gilbert, ed.,
Proc. SPIE 3761, 10 –19 (1999).
3. P. Strand, “Underwater electro-optical system for mine identification,” in Detection Technologies for Mines and Mine-like
Targets, A. C. Dubey, I. Cindrich, J. M. Ralston, and K. A.
Rigano, eds., Proc. SPIE 2496, 487– 497 (1995).
4. L. Mullen, A. Laux, B. Concannon, E. P. Zege, I. L. Katzev, and
A. S. Prikhach, “Amplitude-modulated laser imager,” Appl.
Opt. 43, 3874 –3892 (2004).
5. D. Spitzer and M. R. Wernand, “In situ measurements of absorption spectra in the sea,” in Selected Papers on Underwater
Optics, F. M. Caini, ed., vol. MS118 of the SPIE Milestone
Series (SPIE, 1996), pp. 138 –142.
6. D. Nitzan, A. E. Brain, and R. O. Duda, “The measurements
and use of registered reflectance and range data in scene
analysis,” Proc. IEEE 65, 206 –220 (1977).
7. L. Henyey and J. Greenstein, “Diffuse radiation in the galaxy,”
Astrophys. J. 93, 70 – 83 (1941).
8. R. M. Pope and E. S. Fry, “Absorption spectrum (380 –700 nm)
of pure water. II. Integrating cavity measurements,” Appl.
Opt. 36, 8710 – 8723 (1997).
9. T. E. Jordan, D. L. Cornel, J. Miklas, and E. D. Weller,
“Nutrients and clorophyll at the interface of a watershed and
an estuary,” Limnol. Oceanogr. 36, 251–267 (1991).
10. R. E. Walker and J. W. McLean, “Lidar equations for turbid
media with pulse stretching,” Appl. Opt. 38, 2384 –2397
(1999).
11. R. H. Stavn and A. D. Weidemann, “Optical modeling of clear
ocean light fields: Raman scattering effects,” Appl. Opt. 27,
4002– 4011 (1988).
12. A. Morel and L. Prieur, “Analysis of variations in ocean color,”
Limnol. Oceanogr. 22, 709 –722 (1977).
13. I. L. Katsev, E. P. Zege, A. S. Prikhach, and I. N. Polonski,
“Efficient technique to determine backscattered light power for
various atmospheric and oceanic sounding and imaging systems,” J. Opt. Soc. Am. A 14, 1338 –1346 (1997).
14. P. Werle, R. Mücke, and F. Slemr, “The limits of signal averaging in atmospheric trace gas monitoring by tunable diodelaser absorption spectroscopy,” Appl. Phys. B 57, 131–139
(1993).
15. R. Ricci, R. Fantoni, M. Ferri de Collibus, G. G. Fornetti, M.
Guarneri, and C. Poggi, “High resolution radar for 3D imaging
in artwork cataloguing, reproduction and restoration,” in
Optical Metrology for Arts and Multimedia, R. Salimbeni, ed.,
Proc. SPIE 5146, 62–73 (2003).
16. P. D. T. Huibers, “Models for the wavelength dependence of the
index of refraction of water,” Appl. Opt. 36, 3785–3787 (1997).
17. X. Quan and E. S. Fry, “Empirical equation for the index of
refraction of seawater,” Appl. Opt. 34, 3477–3480 (1995).
20 November 2005 兾 Vol. 44, No. 33 兾 APPLIED OPTICS
7135