89

1 Introduction

Accelerated decay of ice-sheets in Antarc-tica and retreat of the Himalayan glacier havebeen noted as effects of global warming. Areport issued by the Intergovernmental Panelon Climatic Change (ICPP) in 1995 predictedthat global atmospheric temperature will riseby about two degrees by the end of the 21stcentury, the sea will rise by about 50 cm, andthe climate will tend toward greater extremes,including remarkably high temperatures. TheUnited Nations Environment Programme(UNEP) adopted the Framework Treaty onClimate Change at the global environmentalsummit in 1992 (in Rio de Janeiro) for thepurpose of stabilizing the concentration ofgreenhouse gases in the atmosphere to effectcountermeasures against global warming. TheKyoto Protocol, adopted at the third confer-ence in 1997 by signatory nations to thistreaty, has made it obligatory for advancedcountries to reduce their exhausts of green-house gases in the five years after 2008 rela-tive to 1990 levels (including 6% reduction forJapan, 7% reduction for the U.S., and 8%

reduction for the EU). In the sixth conferenceof signatory nations in July 2001, and despitethe departure of the U.S., administrativeguidelines for these reductions were finallyadopted. The signatory states are now prepar-ing respective domestic systems to put theKyoto Protocol into effect as of 2002.

Along with these countermeasures it isneeded to observe phenomena that serve asindices of global warming, together withobservations of greenhouse gases. For thispurpose, mass balance of ice sheets and heightof sea level, which is relatively immune toannual climate variations, are well suited forsuch observations. However, the balancebetween snowfall and melting of ice sheets inpolar regions can presently be estimated withaccuracy of only 50%, since there are fewsites for land-based observations. This accu-racy is equivalent to a sea level variation of0.3 cm or an ice-sheet height of 10 cm annual-ly. Annual increases in the altitudes of Green-land and the central part of Antarctica are esti-mated at 2 to 5 cm, and the uncertainty men-tioned above is estimated to be 1 to 2 cmannually. To measure these predicted values

ISHIZU Mitsuo

3-4 Airborne/Spaceborne Laser Altimeter

ISHIZU Mitsuo

Topography of the global ground surface with very high vertical accuracy and fine lateralresolution enables to observe important indicators related to the global climate change suchas the decay of polar ice sheets or the growth of rain forests, and to measure land activity inagriculture and at urban areas. Satellite-borne laser altimeter is expected to measure landsurface with 10 cm accuracy and 100 m resolution, which is sufficient to observe these indi-cators or activities. Communications Research Laboratory has been studying the laseraltimeter as a valuable space sensor and constructed an airborne laser altimeter for thetests of an availability flying over the sea ice off the Okhotsk coasts of Hokkaido Island. Thispaper reports the results of this observation as well as the recent progress of our study of asatellite-borne laser altimeter.

Keywords Nd:YAG laser, Diode-pumped laser, Heterodyne, Sea ice, Ice concentration

90

at this level of accuracy over the next 5 to 10years, it is necessary to conduct observationsover a region of 10,000 km2 or more, using analtimeter accurate up to 10 cm and performsubsequent averaging. Only a satellite-bornelaser altimeter is capable of meeting theserequirements[1].

A laser altimeter transmits intense pulsedlaser light to the ground from an airplane or asatellite, receives the scattered light, thenmeasures the distance to the surface based onthe round-trip delay time. Simultaneousmeasurement of the height and attitude of thevehicle carrying the laser altimeter enablesmeasurements of surface elevations. Sinceshort pulsed laser light with high energy and apulse width of about 6 ns is used, measure-ments with an accuracy within 10 cm can beperformed only by one laser pulse. In addi-tion, the laser spot (or “footprint”) on theground can be focused as small as 30 m indiameter when using a satellite-borne altime-ter, and it thus provides very high accuracyeven with sloped surfaces.

Finally, a laser with a high repetition rateenables the land topography with high densityand accuracy, even at complex geographicalfeatures. The satellite-borne laser altimeter ismost suitable for the observations of ice sheetsand sea ice in arctic regions where observa-tions cannot be performed from airplanes.The efficacy of satellite-borne laser altimetersin such observations has been the subject ofstudy in various foreign countries [2][3].Another significant advantage of the laseraltimeter lies in its ability to observe the entireearth because it is capable of collecting andprocessing data on the satellite. The laseraltimeter will thus be incorporated into explo-ration programs for the moon and Mars, andwill be used to make comprehensive topo-graphical maps of these celestial bodies.

Measurements by a satellite-borne laseraltimeter will also permit the creation of free-board maps of ice sheets, which will in turnallow observation of changes in ice sheets dueto global warming. In observation of landareas by a laser altimeter, the influence of

warming upon vegetation can be observed bysurveying the vertical distribution of forestsand crops, while volcanic eruptions and move-ments of the continental plates may also beobserved by surveying crustal movements. Ascanning laser altimeter can also make mapsof areas in floods and landslides, thus helpingto contribute to disaster-relief efforts. For thisreason, laser altimeter is also expected to con-tribute to the progress of aerial surveys.

With respect to the past results of satellite-borne laser altimeters, NASA has taken thelead. NASA’s first achievement was the MarsObserver Laser Altimeter (MOLA) installedon the Mars Observer spacecraft. However,the spacecraft collided with Mars in August1993 as it attempted to enter into orbit aroundthe planet. The primary goal was to observethe altitude (accuracy < 30 m) for the entiresurface of Mars on a grid with a 0.2-degreemesh, and laser pulses were scheduled at 60× 106 shots over the expected life of twoyears. Subsequently, the Mars Global Survey-or, launched in November 1996, was placed inorbit around Mars at an altitude of 400 km.The installed Mars Orbiter Laser Altimeter(MOLA-2) performed altitude measurement640 × 106 times until a transmitting laserstopped responding to emission command inJune 2001[4]. The laser maintained a specifiedoutput of 20 mJ/pulse until it stopped, and per-formed oscillations of 671 × 106 shots. TheMOLA-2 had a range accuracy of 37 cm andhorizontal resolution of 300 m, but the stan-dard accuracy of altitude data was limited toabout 5 m, due to the limitations on orbitalaccuracy. The MOLA-2 discovered a uniquegeoid and detailed maps of Mars, andobserved clouds, snow, and such geographicalfeatures as volcanoes[5].

In terms of observations of earth, the Shut-tle Laser Altimeters 1 and 2 (SLA-01, SLA-02) were launched on the Space Shuttles byNASA in 1996 and 1997[6]. These altimeterswere developed using spare parts from theMOLA, modified to provide the altimeter withthe ability to record laser echoes, to observevegetation and other physical features. Sur-

Journal of the Communications Research Laboratory Vol.49 No.2 2002

91

face areas between the latitudes of 28.5degrees north and 58 degrees south wereobserved, due to the inclination of the orbit.In particular, the SLA-2 emitted 3 × 106 laserpulses and received 900 × 103 echoes fromocean and land. In this way the land surfaceelevation, sea level, cloud-top height, surfaceroughness, and similar features were observed.Future plans include a scheduled GeoscienceLaser Altimeter System (GLAS) and a Vegeta-tion Canopy Lidar (VCL) (to be launched onan undecided date). GLAS will be installedon the ICESAT whose inclination will be setto 94 degrees, and be launched in the secondhalf of 2002. GLAS is to observe the massbalance of ice sheets, vertical distribution andoptical density of clouds and aerosols, andsurvey vegetation and land topography. VCLis to observe vertical height of vegetation onthe land and land topography. Table 1 lists thespecifications of these laser altimeters.

To demonstrate the effectiveness of thelaser altimeter, the Communications ResearchLaboratory has developed a laser altimeteraboard a small airplane. Observations of free-boards of drifting ice off the Okhotsk coast ofHokkaido were made using this system. Drift-ing ice was selected as an observation targetbecause laser emissions toward the surfaceposed no safety risks to humans (i.e., no riskof eye injury while standing on the surface),and because the sea level were used as a refer-ence plane. As a result, the measurementaccuracy of the altimeter was fully demon-strated.

Observation equipments to be installed onsatellites are subject to extreme restrictions on

available electric power and mass, and mustbe resistant to vibrations and sounds duringlaunch. The equipments must also be capableof operations without being affected by tem-perature variations in space. Although a laserdiode (LD)-pumped solid-state laser used as atransmitter of the laser altimeter is designedfor installation on a satellite, it consumes agreat deal of electric power and its operatingtemperature range is only half of that of otherelectronic instruments. Accordingly, we areconducting basic research on a satellite-bornesystem, including research on laser design fordeployment on a satellite. This reportdescribes the development of the airbornelaser altimeter, the results of observations ofdrifting ice, and basic research on a space-borne laser altimeter system.

2 Development of Airborne LaserAltimeter

ISHIZU Mitsuo

Examples of space-borne laser altimeters Table 1

LD-pumped Q-switched Nd:YAG

(Laser Diode Inc.)

1064nm, 10mj, 20pps, multi-mode

Transmitting laser

532nm, 2mj, 7nsTransmitting pulse

Newtonian, D=72mm, 10xTransmitting telescope

100μradBeam divergence

Schmidt-Cassegranian,

D=203mm, f=1280mm

Receiving telescope

4mradReceiving FOV

B=1nm, transmittance 40%Optical filter

Photomultiplier with micro-channel plate

(Hamamatsu Photonics, R3809U)

η=7%, rise time 150ps

Detector

50psrmsElectronics jitter

50mmrms @4352mRange resolution

Specifications of airborne laseraltimeter

Table 2

The airborne laser altimeter was construct-ed as a system to be used in observations totest the availability of the laser altimetry. Thissystem was basically the same as usual atmos-pheric lidars and consisted of a photomultipli-er to detect the receiving laser echoes. Thealtimeter can be divided into three blockswhich are a laser transmitter for generatingand sending the laser light, an optical receiverfor collecting scattered light on the surfaceand converting it into electric signals, and adata processing block to measure and recordthe propagation delay and waveforms of thereceived laser pulses. In addition, a GPSreceiver is installed on the airplane to measurethe altitude of the altimeter. Fig.1 shows thesystem block diagram; Table 2 lists the mainspecifications. The following sections discusseach of the block mentioned above in further

detail.

2.1 Laser Transmitter/Receiver Fig.2 shows a photograph of the laser

transmitter/receiver. Fig.3 shows the viewfrom the transmission/reception apertures.The transmitting laser at the upper right inFig.2 is a diode-pumped Q-switched Nd:YAGlaser. Since this laser is air-cooled and smallin size, it is suitable for installing in an air-plane, and will be free from problems of freez-ing coolant water even in very cold environ-ments. The laser pulse has an output energyof 10 mJ, a pulse width of 7 ns, and a repeti-tion frequency of 20 Hz. The pulses are sentthrough a KTP crystal to obtain a second har-monic of 2 mJ with a wavelength of 532 nm.The laser pulses are expanded and transmittedto the ground through a transmitting telescope

92 Journal of the Communications Research Laboratory Vol.49 No.2 2002

System block diagram of airborne laser altimeterFig.1

93

(72 mm in diameter) shown at the lower left inFig.2 to reduce the beam divergence to about100μrad. Assuming that the airplane is flyingat an altitude of 1,000 m, the footprint of thelaser pulse becomes approximately 10 cm.The transmitting time signal is obtained fromthe Si avalanche photodiode shown at thelower left in Fig.3 detecting the light leakedfrom the folding mirror which leads the laserpulse to the transmitting telescope from thelaser.

The laser light scattered on the ground iscollected by a Schmidt-Cassegrain telescope(203 mm in diameter) which is mounted asidethe transmitting telescope (as shown in Fig.3).An interference filter, with a transmissionbandwidth of 1 nm, reduces background lightsfrom the ground. The laser echoes are thendetected by the photomultiplier housed in thecylindrical aluminum case shown in the uppermiddle of Fig.2. To suppress backgroundnoise further, a narrower receiving field ofview provides better performance. However,if this field is too narrow, optical-axis adjust-ments will become difficult. In this opticalsystem, the receiving field of view is designedto cover 4 mrad, by placing an iris (5 mm indiameter) on the focal plane of the telescope.The photomultiplier is placed behind this iris,and has a rise time of 150 ps. The receivedsignal is monitored by an oscilloscope andprocessed by the data processing block togeth-er with the transmitting time signal.

2.2 Data Processing BlockSince laser pulse energy changes in each

pulses and also the reflectivity of the groundchanges according to locations, the receivedpulse intensity fluctuates from pulse to pulse.Therefore, a waveform analyzer measures theprecise time intervals between the leadingedges of the transmitting time pulses andreceived pulses. Fig.4 shows the diagram ofthe data processing block constructed usingconstant fraction discriminators (CFD) as thewaveform analyzer.

The CFD detects the time at which eachpulse rises to a pre-fixed level relative to each

ISHIZU Mitsuo

Transmitter/receiver of airborne laseraltimeter

Fig.2

Transmitter/receiver viewed fromtransmission/reception apertures

Fig.3

94

peak of the pulse not affected by the magni-tude or rising speed of the pulse. To preventthe CFD to detect signals by the scattered lightfrom aerosols, ice or snow in short distancesbelow the altimeter, the circuit is gated so asnot to operate within a distance of 170 m. Thetime difference between the transmission andreception signals from the CFD is measuredby an interval counter to obtain the propaga-tion delay time.

The waveform of the scattered light under-goes digital conversion at a rate of 2 G sam-ple/sec by an A/D converter installed in acommercial high-speed digital oscilloscope.If the sampling interval is 0.5 ns and the fullwidth of the half-maximum of the laser wave-form is 7 ns, the scattered laser light can berecorded in a strongly reproducible mannerexcept for high-speed fluctuations caused bylongitudinal mode-hopping in the laser pulse.This sampling period corresponds to a height

resolution of 7.5 cm. As mentioned before,the laser light is emitted with a beam diver-gence of 100 μrad and the footprint willbecome 10 cm or less. Even when both drift-ing ice and sea surface are included in thisspot, the laser altimeter can distinguish scat-tering targets of drifting ice, sea surface or icesubmerged in seawater. In order to regulatethe data transfer of waveform and height andto controle the electronics, the circuits areconnected to a computer through GPIB.

3 Altitude Accuracy of LaserAltimeter

To enable observation by a laser altimeteraboard a satellite at lower orbit (300 to 500km in altitude), it is necessary to ensure theeye safety of individuals on the ground. Thepower consumption, mass, operating tempera-ture range, and other specifications of the

Journal of the Communications Research Laboratory Vol.49 No.2 2002

Diagram of data processing blockFig.4

equipment must also fall within the relevantpermissible ranges of the satellite. The safetystandards of laser irradiation in Japan are pre-scribed by the Japanese Industrial Standards,which conform to standards stipulated by theInternational Electrotechnical Commission(IEC)[7]. According to these standards, themaximum permissible exposure (MPE) of asingle laser pulse to the eyes with a pulsewidth of 10 ns is MPE532 = 5 mJ/m2 at 532 nm(i.e., second harmonic of Nd:YAG laser), andMPE1064 = 50 mJ/m2 atλ= 1.064μm, the fun-damental wave.

With respect to the satellite-borne altime-ter, assuming that an observer on the ground isviewing through a telescope (30 cm in diame-ter) and all the laser light that enter the tele-scope aperture passes through his pupil (7 mmin diameter), then the permissible laser inten-sity becomes MPE532 = 2.7μJ/m2 and MPE1064

= 27μJ/m2. Assuming that the footprint of thelaser is 30 m, the allowable laser energybecome 5.3 mJ and 53 mJ, respectively, whichare quite small for laser altimetory from space.With an airborne altimeter in the daytime(viewable by the naked eye), the permissiblelaser energies are 0.04 mJ and 0.4 mJ, respec-tively, assuming a footprint of 10 cm. In thewavelength region of 1,400 nm to 2,600 nm(i.e., “optically safe” wavelength), however,MPE increases to between 103 and 104 J/m2.Therefore, the transmitting energy of the satel-lite-borne altimeter may be permitted up to aminimum of 385 J. The laser output is notlimited by the eye safety condition but byother factors as consuming electrical power,mass and cost depending on a satellite.

A technical problem for the laser altimeterin the eye-safe wavelength is that the high-speed photomultiplier with quantum-limitedsensitivity operates only at a wavelengthbelow 1μm. The fast photodiode that is sen-sitive in these regions has high quantum effi-ciency but no amplification capability, result-ing its sensitivity being limited by dark currentand amplifier noise. Such photodiodes, how-ever, will reach to the quantum-limited sensi-tivity if they are used in heterodyne detections

where a local light is superposed with a signallight, and it is possible to maintain the highrange accuracy[8].

The height accuracy of surface topographyof the earth by the laser altimeter is syntheti-cally determined by the range error of thelaser altimeter itself, range error due to thesurface slope, and accuracy of the satelliteposition and attitude. These errors will bedescribed below.

For a comparison of the light detectionmethods the range accuracy of each altimetersystem was estimated with the same transmit-ted energy and same diameters of the receiv-ing telescopes. The compared detection meth-ods consisted of (1) direct detection using aphotomultiplier (PMT) for a wavelength of532 nm, (2) direct detection using a fast pho-todiode (APD) for a wavelength of 1,064 nm,and (3) heterodyne detection (HET). Rangeaccuracy was estimated for both airborne andsatellite-borne altimeters. Since Q-switchedpulsed lasers in the eye safe wavelength arecurrently under development worldwide, HETaccuracy was specified for a wavelength of1,064 nm. The diameter of the receiving tele-scope was set to 1 cm for the airborne altime-ter, and to 15 cm for the satellite-borne altime-ter. Table 4 lists the optical system parametersused in the calculation.

One significant difference between thedirect detection and HET lies in the size of thetransmitting/receiving field of view. The for-mer uses discrete transmitting and receivingtelescopes with different field of view, where-as the latter uses a same telescope for trans-mission and reception, and they have the samediffraction-limited field of view.

3.1 Range Accuracy of AltimeterRange accuracyΔR is given using the S/N

ratio (SNR) by

assuming that the laser pulse widthΔT is 7 nsin any case, where c is the speed of light.SNR can be determined from received signal

95ISHIZU Mitsuo

96

energy Er and the noise. Er is given for dis-tance Z by

where Et: transmitted laser energy, Ar: receiv-ing aperture, To: transmittance of the opticalsystem, Ta: atmospheric transmittance, r: sur-face albedo, andΩ: surface scattering solid-angle. A standard ice/snow surface isassumed[9]. In this case,Ω= πstands for ver-tical incidence. For atmospheric transmit-tance, hazy conditions[10] (as described in theAFGL Tropospheric model) are considered.With respect to noise, background noise ener-gy can be expressed by

where Rt: intensity of scattered sunlight on thesurface, Fv: receiving field of view (sr), Fb:receiving optical filter bandwidth, and Tg:receiving gate time. For the optical filter, Fb= 0.5 nm, and transmittance = 0.6, and thegate time was set as Tg = 4ΔT. Table 3 liststhe parameters for scattered sunlight on thesurface and atmospheric absorption.

By designating the number of receivedphotons, number of background photons,detector dark current, and amplifier noise cur-rent by Nr, Nb, Id, and Ia, respectively, theSNR for respective detection systems become

where Nr =ηEr/hν, Nb = ηEb/hν,η: quan-

tum efficiency of the detector, h: Planck’s con-stant,ν: laser light frequency, Ilo: local lightcurrent for heterodyne detection, e: electriccharge, G: amplification factor of PMT, M:amplification factor of APD, and F: excessnoise factor. In HET, the diffraction-limittedfield of view was assumed, and the optical fil-ter bandwidth is set as equal to the amplifierbandwidth (Fb = 0.94 pm). Table 4 and 5 listthe parameters of each detector and the opticalsystem. Obtained SNR was 11, 2.8, and 900for PMT, APD, and HET for the 1,000-m air-borne altimeter, and 3.9, 0.43, and 125 for the450-km satellite-borne altimeter, respectively.HET provides an overwhelmingly high SNR.

Fig.5 and 6 show the variations in rangeaccuracy estimated from SNR as a function offlight height. In the case of direct detectionwith APD or PMT at low heights, the rangeaccuracy approaches a limit determined bystatistical fluctuations of the received lightbecause the received intensity is sufficientlystrong. As the height increases, the accuracyof APD tends to decrease due to the relativeincrease of dark current, whereas the accuracyof PMT gradually approaches the line of HET.The equivalent optical bandwidth determinedby the amplifier is very narrow and the field-of-view is very small in HET detection, whichleads to a lower background noise than that inthe other detections. The accuracy of HET is

Journal of the Communications Research Laboratory Vol.49 No.2 2002

1000nm500nmWavelength

r=0.5r=0.7Surface albedo

Rt=0.016 (W/m2 srÅ )Rt=0.025 (W/m2 srÅ )Solar background radiation

Ta=0.889Ta=0.678Atmospheric transmission

Parameters of surface scatteringand atmospheric absorption

Table 3

InGaP HeterodyneSi APDPMTSpecifications

G3476-01 (HPK*)C30954E (EG&G)R3809U (HPK*)Detector

1M=120G=2x105Gain

η=0.76η=0.36η=0.047Quantum efficiency

0.8nA50nA2nA(anode)Dark current

-F=4.0-Excess noise factor

0.5mW--Local power

250MHzAmplifier bandwidth

6.5pA/HZ1/2Amplifier noise

10mmφ(airborne)/150mmφ(space-borne)Receiving aperture

Detector parametersTable 4

Optical-system parameters for cal-culation of height accuracy

Table 5

97

ten times higher than PMT because of this lowbackground noise as well as the high quantumefficiency of the detector. For a requiredaccuracy of 0.1 m, neither APD nor PMTwould work according to this preliminary cal-culation. HET could attain this accuracy forflight altitudes below 2,900 m and for orbitalheights lower than 480 km. When a lasertransmitts pulses in the eye-safe region, eyesafety is ensured, and if the output isincreased, the transmitting/receiving telescopediameter can be made even smaller. Althoughthe receiving telescopes of the laser altimetersfor earth observation shown in Table 1 have adiameter of 900 mm or more, this is extremelysmall as a receiving telescope and offers sig-nificant advantages in satellite deployment interms of development schedule, manufactur-ing technology, and constructing cost.

3.2 Height Error Based on SurfaceSlope Angle

The transmitted laser beam has a beamdivergence as in a radar altimeter. Conse-quently, height error occurs. This will be dis-cussed for the HET laser altimeter mentionedin the previous paragraph. Let us assume thatthe laser altimeter located at an altitude Z (asin Fig.7) emits laser light with a beam diver-gence (half width)δdownward at an angleφrelative to the vertical line to the surface, withslope angle S. Because there is a differencebetween the distances to the altimeter at both

ends of the diameter of the footprint, the widthof the received pulse increases. This timewidthτcan be obtained by

IfΔT in equation (1) is substituted byτ,the height error based on the slope angle canbe found. Note that 82 % of the ice sheets inAntarctica has slope angle S = 0.01 rad (0.57degree) or less; if a surface with this slopeangle is observed by the airborne altimeterfrom a height of Z = 1,000 m,τ = 85 ps isobtained, assumingφ = 5 deg. From SNR =900 in this case, the error becomesΔR = 0.4mm.

In the satellite-borne altimeter, Z = 450 kmandφ = 5 deg. are assumed. Since the trans-mitting/receiving telescope diameter is 15 cm,the transmitting beam divergence is as smallasδ = 8.7μrad. The error becomesτ = 2.5nsΔ(R = 34 mm) from SNR = 125. This isthe advantage of HET with a small footprint,whereas in the PMT system the error increasestoτ= 29 nsΔ(R = 2.2 m) becauseδ= 100μrad and SNR = 3.9.

3.3 Height Error Based on AtitudeThe errorΔφ included in attitude angleφ

of the airplane expands the propagation delaytime of the scattered laser light and generatesthe error. To find the error it is necessary to

ISHIZU Mitsuo

Range error of airborne laser altimeterFig.5 Range error of satellite-borne laseraltimeter

Fig.6

98

set S = 0 andδ=Δφ in equation (5); the errormultiplied by c/2 becomes the height error. Ifthere is an error ofΔφ = 0.01 rad (0.57 deg.)in the attitude of the airplane at a height of1,000 m withφ= 5 deg. assumed, it will gen-erate an errorΔR = 0.87 m. To reduce this to0.1 m, the attitude accuracy must satisfyΔφ<1.1 mrad (0.065 deg.). When the altimeter isaboard the satellite at an altitude of 450 km,the error becomes as large asΔR = 390 m,and to obtain an accuracy of 0.1 m, an attitudeaccuracy of 2.5μrad is required. Therefore, itis necessary to measure the attitude with highaccuracy and maintain laser emission angleφat zero.

Based on the foregoing, it is determinedthat height error due to attitude error of theflying object is large in observations with thelaser altimeter. Further, also given the heightvariation of the aircraft or spacecraft, itbecomes important to measure its position andattitude with high accuracy. In this context,devices for measuring satellite attitude withreference to fixed stars are already in use. Theuse of such a device will resolve this problem.

4 Accuracy of Airborne LaserAltimeter

In observations of drifting ice with the air-borne laser altimeter, the received strengthbecomes 5.6 × 10-12 J at a flight altitude of 1

km based on the altimeter parametersdescribed in Table 2, the snow and icereflectance[11], and an optical-system trans-mittance of To = 0.4 (ignoring absorption byclouds and the atmosphere). Based on thisstrength, the number of photons obtained inthe photomultiplier becomes 1.0 × 106. Thisvalue of received photons is sufficiently large,and the detector can thus operate in photocur-rent mode during flight survey. The numberof dark current electrons is 0.002 and may bedisregarded. The background noise from theice and snow surface is 3.5 × 10-17 j, and thenumber of background photoelectrons, Nb, is6.5. The range error becomes 33 mm fromequations (1) and (4), and so the height errorcaused by background noise is thus extremelysmall in the flight survey of the ice and snowsurface.

For a performance test of the laser altime-ter, the laser was emitted from the third floorof a building in the Communications ResearchLaboratory toward a television- transmittingtower 4,352 m away at night. The results indi-cated a distance fluctuation of 200 mmpp and astandard deviation of 50 mmrms. Assuming thatthere was no background noise (as it wasnight) and that the laser light reflectivity of theiron tower was 0.5 to 0.8, the received photonsbecomes 7.5 × 105 to 1.2 × 106 based on thealtimeter specifications. It was expected thatthe range error was 74 to 66 mm, based onequations (1) and (4). The standard deviationof measured values was close to the estimatedrange of error when reflectivity was assumedto be 0.8. Since the surface of the transmittingtower is zinc-plated, the reflectivity was con-sidered to be consistent. From the results ofthis study it is concluded that the equipmentperformance is as expected, based on the mainspecifications.

5 Flight Observations

The laser altimeter was deployed in asmall turbo prop airplane (CESSNA 208),which flew above the sea near Abashiri-cityand above frozen Saroma Lake for observa-

Journal of the Communications Research Laboratory Vol.49 No.2 2002

Height error due to surface slopeangle

Fig.7

99ISHIZU Mitsuo

tions of the ice and snow surface. Fig.8 showsa photograph of the airplane at Memanbetsuairport. It has an aperture (50 cm in diameter)for photographic topography on the inboardback deck, where the altimeter was installed.Transmission and reception of laser light toand from the surface were performed throughthis aperture. Moreover, a compact videocammera was attached to the altimeter to pho-tograph the surface and permit identificationof the topography and height data. The flyingspeed during the experiment was approximate-ly 260 km/h, and the horizontal spacing of thelaser spot on the surface was 3.6 m, based on alaser-pulse frequency of 20 Hz.

Fig.8 shows the sample observation flightpath (Feb. 20, 1993). The airplane took a leftturn from Abashiri harbor above the sea (seaice 1-4), flew over the sand bank, thenapproached Saroma Lake (Saroma 1). Thenthe airplane made a round-trip over SaromaLake, changing its direction from east to west(Saroma 2-5). Circles plotted on the path indi-cate the mean freeboard of the drifting ice,calculated every two seconds. Although a dif-ferential GPS base station was installed at

Memanbetsu airport, it failed to function, sothe data from the single GPS on board the air-plane was used for analysis.

Fig.10 shows the data obtained in thispath, arranged chronologically in two rows.The upper plotted points of each row representthe distribution of height data from the altime-ter (left vertical axis). The line plotted on thelower edge of the distribution represents thedistance from the airplane to the sea surface,estimated based on the height data (left verti-cal axis).

The lower plotted point for each row in

Altimeter-bearing CESSNA 208 aircraftFig.8

Flight path for observation of drifting iceFig.9

100 Journal of the Communications Research Laboratory Vol.49 No.2 2002

Fig.10 represents the freeboards of drifting icemeasured from this estimated sea surface(right vertical axis). As shown in this graph,drifting ice data is characterized by uniformdistribution within a certain freeboard width,or is distributed near the upper and lowerboundaries, and a maximum freeboard of thedrifting ice reaches as high as 50 cm. In sig-nificant contradiction to this data, it wasreported that in a freeboard distribution ofdrifting ice in the Arctic Sea, the distributionof the drifting ice data decreases exponentiallywhen the freeboard becomes higher[12]. It isinferred that this contradiction results from thefact that the sea ice of the Okhotsk Sea is first-year ice, since the coast of Hokkaido is locat-ed south of the boundary of freezing, whereasthe sea ice in the Arctic Sea is multi-year ice,subject to repeated cycles of growth, melting,collection, and distribution. No drifting iceexisted along or near the sea coast close toSaroma Lake during the day of observation; asa result, waves reached the seashore. Theheights of these waves are shown in the graph;it can be seen that the height distribution isclose to the normal distribution. In the sand-bar, there were double cross-shaped structuresto prevent erosion by the ocean current, inaddition to some plant life. The altitude varia-tion was thus large, due to vegetation. SinceSaroma Lake was frozen over (except for thecentral portion), its surface was very smooth

and altitude variation small. The drifting ice off the coast of Hokkaido

is thin relative to arctic sea ice, and unstabledue to ocean current and winds. Further, thefreeboard distribution (as shown in Fig.10)differs clearly from the exponential freeboarddistribution of the arctic sea ice, and the shapeof the distribution varies from place to place.Therefore, it is not appropriate to study thedrifting ice off the Okhotsk coast by means offreeboard distribution. We analyzed the rela-tionship between the ice concentration andfreeboard of the drifting ice. If the measuredfreeboard was no higher than 10 cm, it wasassumed that the laser pulse had struck the seasurface; the ice concentration was thus definedevery 2 seconds (every 40 pulses) as a per-centage of freeboards above 10 cm. In Fig.11,the ice concentration and mean freeboard areplotted for the drifting ice, obtained by obser-vation from 1993 to 1995. The correlationbetween both sets of data was excellentregardless of observation point, time, andyear, and was found to distribute along a sameline. Therefore, if the freeboard distributionof the drifting ice is observed with the laseraltimeter, the ice concentration can be derivedor estimated. Further, it is concluded that, ifthe specific gravities of the drifting ice aboveand below the sea surface are known, itbecomes possible to estimate the amount ofdrifting ice. Details of the observations of

Height data and freeboard of drifting ice by flight surveyFig.10

101

drifting ice and of Saroma Lake using thealtimeter are included in a separate paper[13].

6 Basic Development of Space-Borne Laser Altimeter

As described in Chapter 3, the heterodynelaser altimeter is advantageous because it canemploy low laser-transmission energy and thelaser light can be transmitted in the eye-safewavelength region. Moreover, this systempossesses higher sensitivity than the directdetection system by approximately one orderof magnitude. However, in heterodyne detec-tion higher performance and stability (exceed-

ing those required in direct detection systems)are required for the laser and receiving opticalsystem. First, the transmitting pulsed laserand a local cw laser must both be single-fre-quency lasers, and their frequencies must besynchronized; second, light wave collectedonto the detector through the receiving opticalsystem must be in phase. In other words,these lasers must be of a single longitudinalmode and fundamental transverse mode andthe optical collecting system must be diffrac-tion-limited, and these characteristics must bemaintained throughout the period of observa-tion.

To realize these capabilities, we are devel-oping a heterodyne laser altimeter system, ini-tially for ground experiments. The mostimportant component of this system is thelasers. Since the reflectivity of ice and snowsurface decreases at wavelengths of 1 μm andlonger, it is necessary to carefully determinethe wavelength. Since the flight survey hasconfirmed a measurement accuracy of 10 cmand eye-safety even at a wavelength of 1 μm,the altimeter is being developed with theNd:YAG laser. As its configuration is basical-ly the same as the airborne altimeter, we willdescribe only the transmitter/receiver blockand the laser which are quite different fromour airborne system.

ISHIZU Mitsuo

Mean freeboard and concentrationof drifting ice

Fig.11

Block diagram of transmitter/receiver of heterodyne laser altimeterFig.12

102 Journal of the Communications Research Laboratory Vol.49 No.2 2002

6.1 Transmitter/Receiver BlockFig.12 shows the block diagram of the

transmitter/receiver of the altimeter system.The local laser is a single-frequencymicrochip Nd:YAG laser (CrystaLaser, IRCL-300-1064) delivering cw output of 460 mW.The output beam divergence is 4 mrad andbeam diameter is 0.4 mm. This beam passesthrough an isolator and a beam expander, anda portion of this beam is mixed with thereceived light and focused onto the mixer, anInGaAs PIN photodiode (Hamamatsu Photon-ics K.K., G3476-01) made for the optical-communication band. The quantum efficiencyat 1,064 nm is 80% and the cut-off frequencyis 2 GHz. The isolator is used to cut offreflected beam from the transmitting pulsedlaser, and the expander was inserted for modematching of the local laser beam with those ofthe received light and the transmitting pulsedlaser which is injection-locked by the locallaser.

Fig.13 shows the assembled transmitter/re-ceiver of the altimeter. The transmitting laseris under development, so only a case is shown.To adjust the receiver optics, a 240-mW por-tion of the local laser light was made to irradi-ate the surface of a rotating disk made of awhite plate located 3.5 m away, and its scat-tered light was received to obtain the beat sig-nal shown in Fig.14. This scattered light cor-responds to scattered light from an object at adistance of 2,400 m exposed to a pulsed laserhaving an energy of 4 mJ and pulse width of20 ns. Although the signal level was belowthe theoretical value by approximately 10 dBdue to an insufficient adjustment of the localpower level and the collecting optics, passableheterodyne detection was conducted. Aftercompletion of the transmitting laser, appropri-ate adjustments will be made.

6.2 Transmitting Pulse LaserFor the transmitting laser, it is indispensa-

Transmitter/receiver of heterodyne laser altimeterFig.13

103

ble to use a compact laser with a pulse rate of200 pps or more that oscillates in the funda-mental transverse and single longitudinalmode, and that can be injection-locked for fre-quency locking. This pulse rate falls betweenthat of a flush lamp-pumped pulse laser andthat of a cw-pumped pulse laser, and noappropriate commercial product is available.In addition, the configuration of the space-borne laser differs from those of commercialproducts. Therefore, an appropriate transmit-ting laser is currently under development inour laboratory. Fig.15 shows the basic struc-ture of this laser[14]. Two opposing Porroprisms constitute a resonator, which includes adiode-laser-pumped YAG rod, a Q-switchingPockels cell, and a polarizer. The Porro prismon the Q-switching arm is rotated by 45degrees in azimuth angle and provides a phaseshift resulting from double internal Fresnelreflections and an image reversal in the prism;

the phase shift eliminates the need for a waveplate. The degree of coupling of the output isadjusted by the rotation of azimuth angle ofthe prism on the laser rod arm, and the outputemanates through the polarizer. Two diodelasers are arranged on the side of the rod topump the rod.

Since the rod is pumped by the diodesonly from one side, excitation distribution inthe rod is extremely not uniform. However,each time the laser light travels back and forthin the resonator, the cross section of the beamrotates by twice the angle formed by the ridgelines of the two prisms; therefore, by setting aridge line angle equal to an irrational numbertimes 2π, any light ray passing through a cer-tain point in the rod cross section is directedso as not to pass through that point again.After the multiple round-trips, the nonuniformexcitation distribution can be averaged withrespect to the laser beam around the rod’s axis.With this setting, a top-hat beam with highrotational symmetry may be obtained.According to the conventional laser resonatortheory, it is assumed that the laser cannotoscillate unless the light ray in the resonatorreturns to the initial position and angle. How-ever, this resonator has indicated that the lasercan oscillate even if the laser light ray doesnot return to the initial position, provided thatthe overall amplitude distribution agrees withthe initial distribution after making a round-trip.

This resonator also displays excellent sta-bility against mechanical distortion such as the

ISHIZU Mitsuo

Received heterodyne beat signal oftransmitted cw laser beam

Fig.14

Configuration of diode-pumped Q-switched Nd:YAG laser for transmitterFig.15

104

same as cross-prism resonators. Furthermore,all parts are placed on the same plane, whichsimplifies the manufacture and assembly ofthe laser housing and parts holders. Heat gen-erated in the diodes can be removed along theshortest path through the base plate. This pla-nar construction ensures reliability in everysteps of laser manufacture, adjustment, andtesting.

For its satellite-borne laser, NASA hasemployed a zigzag-slab laser of planar con-struction according to the method developedfor MOLA[15] [16]. However, since the exci-tation distribution can be averaged only in theexcitation direction among the two orthogonaldirections in the rod cross section, the beamdiameter and beam divergence of the laser dif-fer between these two directions. The laser isthus susceptible to improvement. Further-more, since the laser is based on a standing-wave resonator, single-frequency operation isdifficult to achieve.

To achieve single-frequency operation ofthe laser, as required in the heterodyne detec-tion system, the spatial hole burning of excita-tion distribution in the rod must be eliminated.The laser being developed eliminates the spa-tial hole-burning by building up the twistedmode in the rod utilizing the phase shiftcaused by total reflections in the Porro prismof the rod side. In this configuration, single-frequency oscillation can be achieved with aminimum number of parts. Injection lockingis performed by injecting a part of the locallaser light through the polarizer from theopposite direction to the laser output.

This laser used a Nd:YAG crystal ofφ3 to4 diameter and 30 mm long, which waspumped by two LDs [model SDL-3231-A6(360 W), made by SDL Corp]. The hemi-cylindrical surface of the rod on the other sideof the pumping was contacted with a copperblock for cooling. The heat dissipated in therod and from the LD’s is extracted to the opti-cal bench through the base plate. The outputof the laser that used theφ3-mm rod was 5 mJat a pulse frequency of 200 Hz, and a pulsewidth of 8 ns was obtained. Fig.16 shows a

mode pattern of the output pulses at 4 m awayfrom the laser. Although a third-order angularmode of hexagonal symmetry appears slightlydue to the intersecting plane of the Porroprism and a damage spot on the Q-switchfacet, the nonuniform excitation distributionwas completely averaged. Moreover, in thenear-field image, a faint diffraction linecaused by the ridgelines of the Porro prism isseen, along with diffraction rings from the rodcircumference. From these results, it was con-cluded that a plane wave-front was obtainedacross the beam.

Fig.17 shows the relationship between thelaser output energy and the pumping energywhen using theφ4-mm rod. The pumpingduration was 200μs and pulse repetition fre-quency was 100 Hz. Although the efficiencycan be seen to have declined in the high-out-put region because LD temperature was notcontrolled, a differential quantum efficiency ofthe Q-switched oscillation of up to 24% wasobtained. The predicted input/output charac-teristics plotted in the figure were obtained byadjusting the parameters of the laser modelbased on the rate equation so that the outputsagreed at an optical pump energy of 94 mJ.The model agrees well with actual laser per-formance. From the model, the ratio ofabsorbed and pump energy of the rod was esti-mated to be 48%. Further study to increase

Journal of the Communications Research Laboratory Vol.49 No.2 2002

Beam pattern at 4 m from laserFig.16

105

this efficiency is necessary. Fig.18 shows the pulse waveform at the

same pump energy of 94 mJ. The waveformis shown from the end of the pumping atwhich Q-switch is turned ON. The waveformand the pulse width of 28 ns which are calcu-lated from the laser model agree well withthose of the actual laser, but the calculatedpulse rises up later than that of the actual pulseby 25 ns. The delay in the rise time wasattributed to a discrepancy between the initialsetting of number of photons in the model andthe actual laser.

The output-coupling coefficient of thislaser is small to compensate a relatively lowergain coefficient of the rod withφ4 mm diame-ter. This resulted in the high Q-value of thelaser resonator which stretches the laser pulsewidth. A small ripple in the actual waveformwas caused by mode-hopping, which is a char-acteristic of the waveform of single-frequencypulsed lasers. Through injection-locking, thismode-hopping will disappear and the entirepulse will be forced into a single frequency.Since the laser has displayed sufficient funda-mental performance as an transmitter, thealtimeter is scheduled to be tested using injec-tion-locking technique after assembly of thelaser.

7 Conclusions

We have built and tested an airborne laseraltimeter that uses a diode-pumped Nd:YAGlaser for the development of a satellite-bornelaser altimeter. Measurement resolution in the5-cm range (expected from main specifica-tions) was obtained from the results of theground test. This altimeter was deployed onan airplane to observe ice drifting off of theOkhotsk coast of Hokkaido. It was demon-strated for the first time that a strong correla-tion existed between the mean freeboard andconcentration of ice. For the deployment of asatellite-borne altimeter we are conductingbasic study of a heterodyne laser altimeter thatcan provide eye safety and allow the use of acompact receiving telescope. The develop-ment results of the transmitter-receiver andtransmitting laser of the heterodyne laseraltimeter have presented.

8 Acknowledgment

A part of this research was supported byEarth Remote Sensing Technology Researchin the Ocean Development and GeoscienceTechnology Research Program of the Ministryof Science and Technology (1990 though1997).

ISHIZU Mitsuo

Pumping input and laser outputFig.17

Waveform of Q-switched laser pulseFig.18

106 Journal of the Communications Research Laboratory Vol.49 No.2 2002

References1 NASA, Instrument panel report: LASA Lidar Atmospheric Sounder and Altimeter, Earth observing

system reportsⅡd, p4, 1987.

2 S. C. Cohen, J. J. Degnan, J. L. Bufton, J. B. Garvin and J.B.Abshire, "The Geoscience Laser

Altimetry/Ranging System", IEEE Trans. on Geosci. and Remote Sensing GE-25, pp.581-549, 1987.

3 Bufton J., "Laser Altimetry Measurements from Aircraft and Spacecraft", Proc. IEEE, 77, 463-477,

1989.

4 Abshire, J.B., X. Sun and R.S. Afzal, "Mars Orbiter Laser Altimeter: Receiver model and performance

analysis", Appl. Optics, 39, 2440-2460, 2000.

5 Smith, D.E., M.T. Zuber, H.V. Frey, J.B. Garvin, J.W. Head, D.O. Muhleman, G.H. Pettengill, R.J.

Phillips, S.C. Solomon, H.J. Zwally, W.B. Banerdt, T.C. Duxbury, M.P. Golombek, F.G. Lemoine,

G.A. Neumann, D.D. Rowlands, O. Aharonson, P.G. Ford, A.B. Ivanov, P.J. McGovern, J.B.Abshire,

R.S. Afzal, and X. Sun, "Mars Orbiter Laser Altimeter (MOLA): Experiment summary after the first

year of global mapping of Mars", J. Geophys. Res., 106, 23, 689-23, 722, 2001.

6 Garvin, J., J. Bufton, J. Blair, D. Harding, S. Luthcke, J. Frawley, and D.D. Rowlands, 1998, "Obser-

vations of the Earth's topography from the Shuttle Laser Altimeter (SLA): laser-pulse echo-recovery

measurement of terrestrial surfaces", Phys. Chem. Earth, 23 (9-10):1053-1068, 1998.

7 IEC825-2 and JIS C 6802.

8 Ishizu M. and T. Itabe, "Observation of sea ice in the Sea of Okhotsk by a laser altimeter", Proceed-

ings of SPIE, 3382, pp.146-151, 1998.

9 Wolfe, W. L. and G. J. Zissis, (Eds.), in the Infrared Handbook, Figure 3-124, Environmental

Research Institute of Michigan, Ann Arbor, MI, 1978.

10 McClatchyet al., Optical properties of the atmosphere, AFCRL-72-0497, AFCRL, Hanscom AFB,

MA,1970.

11 O'Brien H. H., Red and near-infrared spectral reflectance of snow, Report No. 332, U.S. Army Cold

Regions Research and Engineering Lab., Hanover, NH, Mar. 1975.

12 P. Wadhams, W. B. Ⅲ. Tucker, W. B. Krabill, R. N. Swift, J. C. Comiso and N. R. Davis, "Relation

between sea ice freeboard and draft in the Arctic Basin", J. Geophys. Res., 97(C12), 20325-20334,

1992.

13 M. Ishizu, K. Mizutani and T. Itabe, "Airborne freeboard measurements of sea ice and lake ice at the

sea of Okhotsk coast in 1993-95 by a laser altimeter", Int. J. of Remote Sensing, 20, pp.2461-2476,

1999.

14 Patent pending, No.400096 (2001), No.400097 (2001).

15 Afzal, R. S., "Mars observer laser altimeter: laser transmitter", Appl. Opt., 33, pp.3184-3188, 1994.

16 Errico A., C. Narrie, A. Cosentino, P. Laporta, P. Wazen and P. Maine, "Diode-pumped high-efficien-

cy high-brightness Q-switched Nd:YAG slab laser", Opt. Lett., 22, pp.1168-1170, 1997.

107ISHIZU Mitsuo

ISHIZU Mitsuo

Senior Researcher, Lidar Group,Applied Research and Standards Divi-sion

Laser Remote Sensing