A Portable Airborne Laser System
for Forest Inventory
Ross Nelson, Geoffrey Parker, and Milton Hom
were first applied to terrestrial concerns throughout the 1970sAbstract
(Link and Collins, 1981; Hoge et al., 1983; Krabill et al., 1984).A simple, lightweight, inexpensive, portable airborne laser
Over the past 30 years, the technology has moved from researchprofiling system has been assembled from off-the-shelf, com-
to commercial applications (Baltsavias, 1999). Applicationsmercially available components. The system, which costs
include the use of airborne laser data for creation of digital ter-approximately $30,000, is designed to fly aboard small
rain models, monitoring power lines, urban area mapping, andhelicopters and single- or twin-engine high-wing aircraft
land-cover surveys (e.g., http://www.airbornelasermapping.without airframe modification. The system acquires first-
com, last accessed 26 August 2002). Numerous commercialreturn range and amplitude measurements at data rates up
companies now offer turnkey airborne laser scanning systems orto 2000 hz (operator-controlled) and has an operational en-
for-hire airborne laser data collection.velope up to 300 m above terrain. The airborne laser profiling
Applications of laser technology to forest inventory prob-system includes the laser transmitter/receiver, differential GPS
lems, however, remain largely in the research arena, for the fol-receiver, a CCD video camera and recorder, and a laptop com-
lowing reasons:puter which interleaves and records the GPS and laser range/
amplitude data. The portable airborne laser system (PALS) was
? Laser data post-processing to derive inventory estimates
requires the attention of specialists and, frequently, the applica-designed to acquire forest height measurements along linear
tion of specialized, home-grown software;flight transects in order to conduct regional or subcontinental
? Although forest/laser research has been ongoing for at least twoforest inventories worldwide. This economical laser system
decades, it is only in the last few years that researchers havenow puts airborne laser mensuration within reach of opera-
demonstrated beyond a reasonable doubt that laser data can betional foresters and researchers interested in making rapid
used to create reliable maps of tree height, canopy density, bio-forest structure and/or timber surveys in remote areas. PALS mass, and volume; and
has been used to acquire over 5000 km of flight transect data
? The flight hardware, to this point, has been expensive.
over the state of Delaware.
Certainly, some of the more sophisticated commercial and
research laser scanners maintain price tags in the million dollarIntroduction
plus range. Research systems such as SLICER (Scanning LidarAirborne lasers may be used to acquire ranging data to measure
Imager of Canopies by Echo Recovery; Blair et al., 1994; Har-tree heights. The tree height data may be used directly to meas-
ding et al., 1994), LVIS (Laser Vegetation Imaging Sensor; Blairure such forest biophysical characteristics as average canopy
et al., 1999), AOL (Airborne Oceanographic Lidar; Krabill et al.,height, height variability or canopy roughness, and canopy
1984), and EAARL (Experimental Advanced Airborne Researchclosure (Arp et al., 1982; Nelson et al., 1984; Aldred and
Lidar) are experimental, one-of-a-kind systems that are de-Bonner, 1985; Schreier et al., 1985; Ritchie et al., 1992; Ritchie
signed to push technological envelopes rather than to be cost-et al., 1993; Nilsson, 1996; N?sset, 1997a; Blair et al., 1999;
effective. Commercial airborne lasers currently sold areMeans et al., 1999; Means et al., 2000; Popescu et al., 2002).
sophisticated, turn-key scanning systems used primarily to cre-These height and density measurements, in turn, can serve as
ate accurate airborne digital elevation models (DEMs). Thesethe independent variables in predictive models to estimate for-
laser scanning systems record precise range and positioningest basal area, merchantable volume, biomass, and carbon
data so that georeferenced DEMs can be produced. Some com-(Maclean and Krabill, 1986; Nelson et al., 1988a; Nelson et al.,
mercially available, airborne laser scanning systems host1988b; Nelson et al., 1997; N?sset, 1997a; N?sset, 1997b; Lef-
multi-return receivers. Data from these multiple-return, lasersky et al., 1999a; Lefsky et al., 1999b; Popescu et al., 2000).
scanning systems can be used to make DEMs of the top of theA wide variety of laser systems now exist, and numerous
forest canopy, the ground, and significant subcanopy layers.researchers are investigating uses of these systems with respect
Commensurate with this level of sophistication are price tags ofto forestry applications. Originally designed for bathymetry
hundreds of thousands of dollars, which effectively puts these(Hickman and Hogg, 1969; Hoge et al., 1980), airborne lasers
data out of the reach of operational foresters.
The authors, being interested in the uses of laser data for
large-area forest inventory and forest canopy characterization,
set about to design a simple, inexpensive, portable airborneR. Nelson is with the Biospheric Sciences Branch, Code 923,
laser system. Their objective was to design, build, test, andNASA/Goddard Space Flight Center, Greenbelt, MD 20771
(ross@ltpmail.gsfc.nasa.gov).
G. Parker is with the Smithsonian Environmental Research
Center, 647 Contees Wharf Road, P.O. Box 28, Edgewater, MD
Photogrammetric Engineering & Remote Sensing21037-0028 (parker@serc.si.edu).
Vol. 69, No. 3, March 2003, pp. 267?273.
M. Hom is with Science Systems and Applications, Inc., Code
0099-1112/03/6903?267$3.00/0923, Biospheric Sciences Branch, NASA/Goddard Space
 2003 American Society for PhotogrammetryFlight Center, Greenbelt, MD 20771 (mhom@pop900.
gsfc.nasa.gov). and Remote Sensing
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING March 2003 267
operate an airborne laser profiling system, not married/dedi-
cated to any particular airframe, that could be used to structur-
ally characterize and inventory forests. It is their hope that this
type of system (1) will make airborne laser hardware economi-
cally more accessible, and (2) will allow scientists to conduct
laser investigations in isolated locales far from home, e.g., the
circumpolar boreal forests, South America, the Congo, and
Asia. This report describes the result of their efforts to build a
small, robust airborne laser profiling system.
System Description
With rapid technological advances, electronic miniaturization,
orders-of-magnitude increases in computing speed and data
storage capability, and software advances, systems with price
ranges in the tens of thousands of dollars are appearing. The
laser system herein presented was designed with the following
priorities paramount: portability, simplicity, low cost, ease of
operation, and applicability to forest/vegetation applications.
To this end, the Portable Airborne Laser System (PALS) was
designed using only commercially available components,
including
? a near-infrared laser transmitter/receiver to measure first-return
ranges and amplitudes from laser to target;
? a differential Global Positioning System (dGPS) receiver to moni-
tor aircraft location;
? a charge-coupled device (CCD) video camera with GPS video
titling and video recording to maintain a time/location synchro-
nized video record of the flightlines;
? a laptop computer to record the digital dGPS and laser data; and
? a commercial software package to control, monitor, and record
the dGPS and laser data streams.
Each of these components is described in some detail
below. Name brand hardware and software packages and the
Plate 1 Components of the Portable Airborne Laser System.approximate cost at the time of purchase (1999) are noted so
that the reader can reproduce the system exactly. Certainly,
other components can be used to construct a PALS-like instru-
ment, and company names are mentioned only as a starting
point and are not meant to imply exclusivity or highest quality.
System components are shown in Plate 1, and a summary of
the PALS components are reported in Table 1. serial port handling the laser data must handle at least that baud
rate in order to capture the data. The firing rate of the laser can-
Laser not be adjusted; it is always running at 2000 hz. The high-speed
A class IIIB, non-eyesafe, near-infrared (0.905 m), pumped laser data stream can only be subsampled or averaged on the
diode laser transmitter/receiver is used to collect range meas- computer side or averaged, in standard mode, onboard the
urements from the flight platform to the target. The laser is eye- laser.
safe when operated from a moving platform. The Riegl LD90- The transmitter has a 2-mr divergence and 10-cm optics so
3800-VHS laser, which comprises the heart of the PALS system, that, at a nominal flight altitude of 150 m, the system is illumi-
accounts for approximately half of its cost. The transmitter nating a 0.3-m spot on the ground. Ground speed and the laser
(i.e., pulse generator) emits a 20-ns (full width, half max.), 500- sampling rate determine the horizontal (i.e., along-track) post
nJ, near-infrared pulse at 2000 hz. The laser can be pro- spacing between sequential laser shots.
grammed by the analyst to measure range to first target or range
to last target. Company specifications for this laser report rang-
Differential GPSing accuracy on the order of 2.5 cm (http://www.riegl.co.at, last
The PALS system employs a handheld GPS satellite receiver andaccessed 26 August 2002; see laser altimeters - LD90-3 series).
beacon receiver for on-the-fly differential correction. A Gar-Unfortunately, the laser employed in this study cannot be tog-
min GPS III satellite receiver with a remote antenna is used togled such that first-, then last-return ranges are measured
acquire signals from the GPS satellite constellation, and a Gar-sequentially, though such lasers are available today. The laser
min GBR-21 beacon receiver is used to acquire land-based sig-can be configured to provide amplitude (i.e., strength-of-
nals from U.S. Coast Guard dGPS beacons scattered across thereturn) measurements, and can be programmed to run freely
U.S. (http://www.navcen.uscg.gov, last accessed 26 August(laser-initiated firing) or to be triggered digitally, i.e., com-
2002). The GPS III updates differentially corrected positionputer-initiated firing. Finally, the laser can be set up to average
information - latitude, latitude, longitude, elevation abovepulses (standard mode) or to collect individual pulses at 2000
mean sea level (MSL), speed, heading, Greenwich Mean Timehz (high speed mode).
(GMT), and signal quality metrics - once every two seconds. InAs currently configured, PALS is a first-return, free-run-
the event that a beacon signal is not available, the system col-ning, high speed, 2000-hz transmitter. The receiver detects the
lects non-differential GPS data. The serial stream is sent simul-return pulse and generates a 4-byte serial stream where the first
taneously to a video titler (discussed below) to be incorporatedthree bytes record the range and the fourth byte reports the
into the video stream and also to a second serial port on thestrength of the return from the laser shot on a 0 to 255 scale. At
2000 hz, the laser outputs a serial stream at 64,000 baud, so the computer to be interleaved with the laser ranging data.
268 March 2003 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
of 180 km/hr (50 m/sec), the image remains sharp, and individ-
ual tree crowns are easily discerned.
The camera picture is routed to a video titler which accepts
both the S-video signal and the ASCII string from the GPS III sat-
ellite receiver. The GPS ASCII string is the same as that written to
the dGPS/laser file, so that the video and computer data records
are synchronized. The Horita GPT50 GPS video titler integrates
the two and labels the video stream with the current latitude,
longitude, GMT, various measures of GPS signal quality, inter-
nal (local) time, and date.
The annotated video stream is then routed to an 8-mm
video cassette recorder (Sony AV500). The 8-mm tapes record
approximately 2 hours of flight data. The VCR also records the
cockpit conversation via a jack from the internal helicopter
communications system to the audio jack on the VCR.
Computer
Two serial data streams, laser and GPS, are interleaved and
recorded on a laptop personal computer with a 12-GB hard
drive, 384 MB of RAM, and two serial ports; one internal and one
on a PCMCIA card. In the aircraft, the computer runs on battery
power alone, so a total of four 2.5-hour batteries were pur-
chased, enough to last for an entire day?s mission (batteries were
partially recharged at every fueling stop). A typical 10-hour
day?s mission collected on the order of 50 to 75 MB of data.
With an effective storage capacity on the computer of 10 GB, thePlate 2 PALS mounted beneath a Bell 206 JetRanger.
computer could store 130 to 200 days worth of data. Typically,
however, the data were backed up nightly onto 100-MB zip
drives.
Video System
Data Collection ProgramA video history synchronized with the laser data is acquired
with a CCD camera mounted next to and boresighted with the LABView, a software package designed to control and collect
data from a myriad of scientific instruments, was used to re-downward-looking laser. The Pulnix TMC-7 color camera
refreshes every other line in the 768 column by 494 line pixel cord the laser and GPS data. A LABView program, called a VI
(or Virtual Instrument), was written (1) to ingest the 2000-hzarray at 60 hz, effecting a complete refresh every thirtieth of a
second. A 16-mm, manually focused and shuttered lens fronts laser data stream and the 0.5-hz GPS data from the serial ports;
(2) to translate the 4-byte laser observations into characterthe camera. The focus is set to infinity, and on all but the darkest
flights, i.e., dusk or dawn acquisitions, the f-stop was set at f16. strings of range and amplitude; (3) to strip the appropriate GPS
information from the GPS III serial stream; (4) to subset theBright afternoon sun with an f11 setting would saturate the
camera over bright targets, e.g., concrete and dirt roads. At 150 2000-hz laser data stream, i.e., process every pulse, every other
pulse, every third pulse, . . ., every 50th pulse; (5) to displaym AGL (above ground level), the 16-mm lens has a field-of-view
of 60.5 m along-track by 45.4 m across-track. At a flight speed range, amplitude, and GPS data, real time, on the computer
TABLE 1. PORTABLE AIRBORNE LASER SYSTEM COMPONENTS. COSTS ARE IN 1999 US DOLLARS (INC.  INCLUDED IN PRICE )
Subsystem Component Description Cost(USD)
laser range-finder Riegl LD90-3800-VHS laser 2000-hz, near-infrared transmitter/receiver 16,150
video system Pulnix TMC-7 video camera 900
14-mm lens manual focus/shutter 140
clear filter protective lens cover 20
Horita GPT50 GPS video titler 590
Sony AV500 8-mm cassette recorder 870
Sony NP-F950 9-hour battery for AV500 150
dGPS system Garmin GPS III satellite receiver 400
Garmin GA26 remote satellite antenna 40
Garmin GBR21 beacon receiver 400
Garmin GBR21 beacon antenna 165
Garmin GPS45 power/data cable 50
computer system Dell Inspiron 7000 12-GB Hard drive, 128-MB RAM 6,000
PCMCIA card two additional serial ports inc.
three batteries 2.5 hour each inc.
LABView software package 2,500
power system Artesyn BXB100 18- to 36-vdc to 12-vdc power converter 150
Total Cost of Hardware (excluding aircraft mounting costs): 28,525
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING March 2003 269
Figure 1. A 2.5-km section of flightline in Delaware. The ground line is a spline
fit to local minima. The aircraft elevation is offset (i.e., reduced) by 120 m to
permit plotting on the graph.
screen; and (6) to record a GPS-only file (for later use in GIS pro- km of flight data over the state of Delaware. The flight profile
grams) and a separate laser file with GPS records interleaved. followed in Delaware is reported in Table 2.
The program, i.e., the VI, has a number of operator-defined During a data collection mission, the operator controls the
controls that may be altered by the PALS operator during data airborne profiling laser by turning on the laser and initiating the
acquisition. The operator can control and change the rate at LABView program enroute to the flightline starting point.
which the 2000-hz laser data stream is subsampled. The opera- The operator is provided with running traces of the first-return
tor can also adjust the graphics to size and position the trace of range and amplitude as well GPS position, speed and altitude
the laser ranges, i.e., the laser profile, on the screen. All of these information, updated every 2 seconds. The range from aircraft
settings can be changed on the fly while the VI is running. to the first target is subtracted from a constant so that a realistic,
not inverted, profile is presented to the operator, i.e., so that
Power trees ?grow up? from the ground instead of ?into? the ground.
Much of the airborne laser profiling system runs off aircraft An illustration of the type of data collected by PALS is pre-
power. The Bell JetRanger?s 28 vdc power supply was tapped sented in Figure 1.
and converted to 12 vdc for use by the laser, video titler, GPS sys- A number of forest structural measurements can be
tem, VCR, and CCD camera. Only the laptop runs on its own extracted from discrete segments of the profiling data with lit-
internal batteries. PALS, minus the laptop, draws approximately tle effort. Height measurements include average canopy height,
4 amps at 12 vdc. The aircraft power, converted to 12 vdc, is height variance, quadratic mean height, maximum height(s),
distributed to the laser, GPS receiver, and video systems through and various decile or quartile heights. The ratio between tree
the power hub noted in Plate 1.
PALS was designed specifically for forestry applications.
The system measures relative, i.e., local, heights, and PALS
records locations which can be used to approximately position TABLE 2. FLIGHT PROFILE AND INSTRUMENT SET-UP FOR THE DELAWARE
OVERFLIGHTSthe flightlines on the ground or in a GIS to within 10 meters.
The system does not measure absolute elevations, i.e., heights
General: 5000 km of flight data acquired June?August, 2000, 5 days.above an invariant datum such as MSL, nor does it record the 56 systematic, N-S flightlines 1 km apart; longest: 163 km
location data needed to position the laser trace on the ground shortest: 
1 km
with sub-meter precision. Sub-meter X, Y, Z locations would Cumulative flight time, including transit: 49 hours over 5 days
require the inclusion of a centimeter-level dGPS system (cost in Flight Costs:  $30,000 USD
Nominal flight altitude AGL: 150 mthe $10k to $15k range) and a tiltmeter or INS system capable of
Nominal flight speed: 50 m/sec (180 km/hr,measuring aircraft attitude changes on the order of tenths of a
97.2 knots)degree.
Laser subsampling interval: 10:1
Effective laser firing rate: 200 hzData Collection and Performance
Effective post spacing: 0.25 m
PALS was fitted to a cargo hook frame beneath a Bell 206 Jet- Laser spot size (2 mr divergence): 0.3 m
Ranger (Plate 2) and was used to collect approximately 5000
270 March 2003 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
hits and ground shots yields canopy density. Canopy rough-
ness can also be quantified by calculating height variance or
rumple, where rumple is the distance traced over the canopy
divided by the horizontal distance along the flight segment.
Many of these profiling variables have been shown to be corre-
lated with forest volume and biomass (Nelson et al., 1988a; Nel-
son et al., 1988b; Nelson et al., 1997; also the Delaware inven-
tory results, not yet published). Coefficients of determination,
however, are typically in the 0.5 to 0.6 range due to the fact that
height is not a good predictor of volume or biomass; tree diame-
ter is the predictive driver.
In order to assess the stability of the PALS, the system was
bench-tested for two one-hour periods at the Goddard Space
Flight Center after the U.S. Department of Defense had stopped
dithering the GPS signal. The test was run to quantitatively
assess (1) the locational stability of the GPS system (both differ-
entially corrected and uncorrected), (2) the stability of the free-
running laser?s firing rate, and (3) the power stability of the laser
as described by the strength of the laser return from an un-
changing target. Table 3 reports the results from the two one-
hour tests.
The results presented in Table 3 quantify characteristics
which limit the utility of this laser profiling system. The sys-
tem acquires highly accurate ranging data from the aircraft to
first target, but the locational inaccuracies of the aircraft and
the profile track are on the order of 10 meters in X and Y and 10
Figure 2. Relationship between ground-measured buildingto 20 meters in Z. PALS data should be used to measure local
heights and airborne laser measurements of the sameheight differences, e.g., tree heights. It cannot be used to gather
buildings.anything more than gross topographic information because the
aircraft position and laser pulse positions may be off by 10 m or
more. These errors could be reduced below 1 m by incorporat-
ing an inertial navigation system or tiltmeters and a dGPS sys-
tem which works with local base stations to differentially
using a handheld laser rangefinder, the Jenoptik LEDHA-GEO.correct the aircraft location. But the cost of the system would go
The linear equation relating the laser to ground heights has aup appreciably, and the need to establish local base stations
slope close to but significantly different from 1.0 at the 95 per-would hamper this system?s use in remote, inhospitable areas.
cent level of confidence (p 
 0.0001). The mean differenceFigure 2 presents results which compare building heights
between laser and ground measurements was 17 cm  59 cm,measured in the field to heights measured by the laser. Build-
insignificant at the 95 percent level of confidence (two-sidedings were measured rather than trees because it is easier to iden-
paired t-test, p 0.0423). Laser-ground differences ranged fromtify and measure the top of a building (better line-of-sight,
2.1 to 2.0 meters; 90 percent of the differences fell betweenobvious surfaces) and because buildings - collections of straight
0.6 and 1.0 meters. It is the opinion of the primary authorlines, flat surfaces, and right angles - are much more forgiving
that much of this variability reflects a lack of precision in thewith respect to errors in flightline location. Man-made struc-
ground measurements, not laser inaccuracies.tures ranging in height from 1.2 to 63.1 meters were measured
Post-Flight Processing
TABLE 3. BENCH TEST RESULTS FOR PALS
PALS is a first-return, airborne laser profiling system. In order to
stan. dev. range measure tree heights along these first-return transects, a
ground-finding program was developed to define a ground
1. GPS Location Stability: (meters) (meters) trace beneath the forest canopy. Once this ground line wasDifferential: Latitude: 0.93 5.56
defined by identifying and interpolating between suspectedLongitude: 0.57 7.16
ground hits, a canopy height measure could be defined forElevation-MSL: 1.79 11.6
each pulse.Non-Differential: Latitude: 1.30 9.27
Longitude: 1.43 11.46 The frequency with which ground is found beneath a forest
Elevation-MSL: 2.40 16.5 canopy is a function of the density of the forest, the laser spot
2. Range Stability to a 70-m size at target, and the sensitivity of the laser (Aldred and
target (brick wall): Bonner, 1985). The sensitivity of the PALS laser receiver cannot
first hour: mean 70.39 m be adjusted in the field, and it was common to range off of
sd: 0.06 m power lines which undoubtedly intercepted less than 10 per-range: 0.4 m (70.1 m to 70.5 m)
cent of the 0.3-m laser footprint. Although ground returns wereResults duplicated in 2nd hour-long test.
infrequent in the typically dense, young secondary forests ofNo degradation over time (5 minute intervals over hour).
Delaware, enough occurred within the stands and at field/stand3. Repetition rate stability over time:
5 minute blocks of data, e.g., 0?5 min, 15?20 min, 30?35 minutes, edges, grade crossings, and waterways that a coherent ground
. . . over a two hour period. trace could be fashioned.
overall mean (2 hr): 2086.34 hz The ground-finding program identifies local minima
mean, 1st hr: 2087.00 within a user-defined window and then fits a spline to the sus-
range, 5 min blocks, 1st hr: 2085.0 to 2087.5 hz pected ground hits. The window size used in Delaware was 1 or
mean, 2nd hr: 2085.67 hz 2 seconds - 1 second in the hilly terrain of northern Delawarerange, 5 min blocks, 2nd hr: 2084.2 to 2086.7 hz
and 2 seconds on the flats. The analyst reviews a flightline and
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING March 2003 271
may adjust one or more of the minima in order to produce what helicopter?s radio, the squelch adjustment reduced the
operating radius of the radio. Tests were run in-flight to try tohe/she believes is an adequate representation of the ground.
The program also interpolates, for each pulse, the GPS latitude/ identify the offending electronics, and it appears that it came
from two sources. The GPS remote satellite antenna was one sig-longitude, GMT, aircraft elevation above MSL, and the aircraft
ground speed and heading. These position/time data plus the nificant source of radio frequency noise. We recommend that a
GPS avionics satellite antenna be used and located away fromrange-to-canopy, range-to-ground, and amplitude are recorded
for each pulse. The file output by the ground-finding program the aircraft radio. A second source of noise was the 18- to 36-
vdc to 12-vdc converter. It is recommended that the converterserves as the base file in all subsequent processing involving GIS
and forest inventory analyses. be housed in a shell constructed from materials which block
rf noise.Subsequent PALS-like systems will want to incorporate a
laser which, at a minimum, sequentially toggles between first The system has limitations. PALS can effectively operate up
to 300 m above terrain. Above 300 m, the laser return becomesand last returns. The identification of a reliable ground line was
straightforward in Delaware where topography is minimal and weak enough that significant data dropout occurs, especially
over forests. (Below 300 m, significant dropout is limited tothe forests are highly dissected. Establishing a reliable ground
line in mountainous or jagged terrain in areas of contiguous, areas of standing water, which absorbs the near-infrared laser
pulse.) This 300-m limit makes the system ill-suited to over-high-LAI forests would be problematic using a first-return laser.
When the PALS components were purchased in 1999, the se- flights of mountainous areas in fixed-wing aircraft. If use is con-
templated in areas with significant topography, then neces-quential toggling was not available; it is today. Likewise, rela-
tively inexpensive laser scanners are also available, and one of sarily the system will have to be mounted aboard a helicopter
so that the operational envelope can be maintained. The effectsthese might be substituted for the first-return laser transmitter/
receiver. of aircraft speed and elevation changes necessitated by undu-
lating topography can be removed from the laser ranging data
in the data post-processing phase.Summary
A simple, portable airborne laser profiling system has been The simplicity, the compactness, and the meager power
requirements of the system preclude the need for dedicatedassembled from off-the-shelf, commercially available compo-
nents. The system is designed to be small, lightweight, simple, support aircraft and staff, thereby greatly reducing research or
operational budgets. These cost savings will permit foresters torelatively inexpensive, easy-to-install, and easily operated by
one person. The system is designed to be transported to remote employ such systems in locales where reconnaissance-level
forest inventory data are most needed, e.g., in understudiedsites and installed aboard local, for-hire aircraft. One person
transports, installs, and operates the system; support staff regions in the Amazon, the Congo, Southeast Asia, and the cir-
cumpolar boreal forests of Canada and Russia.(other than a pilot) is unnecessary.
The system was designed with economy, portability, and
component availability in mind. PALS is undoubtedly one of
Referencesthe simplest laser profiling systems around, and it provides a
relatively simple data stream of first-return measurements. It Aldred, A.H., and G.M. Bonner, 1985. Application of Airborne Lasers
to Forest Surveys, Information Report PI-X-51, Petawawa Nationallacks the dGPS and inertial information needed (1) to accurately
Forestry Institute, Canadian Forest Service, Agriculture Canada,locate the laser trace on the ground to better than 10 meters,
Chalk River, Ontario, Canada, 62 p.and (2) to accurately map topography. As such, it should be con-
Arp, H., J-C. Griesach, and J.P. Burns, 1982. Mapping in tropical forests:sidered a tool which can be used to accurately measure relative
A new approach using the laser APR, Photogrammetric Engi-heights, e.g., vegetation canopy heights or building heights
neering & Remote Sensing, 48(1):91?100.locally (sub-meter). It is not designed to accurately measure
Baltsavias, E.P., 1999. Airborne laser scanning: Existing systems andtopography relative to a fixed datum, e.g., elevation above mean
firms and other resources, ISPRS Journal of Photgrammetry andsea level. In this respect, PALS gathers approximate measures,
Remote Sensing, 54:164?198.reporting heights with an accuracy above a fixed datum of 10 to
Blair, J.B., D.B. Coyle, J.L. Bufton, and D.J. Harding, 1994. Optimization20 meters along regional profiles tens or hundreds of kilome-
of an airborne laser altimeter for remote sensing of vegetation andters long. Given such characteristics, PALS can be used to meas-
tree canopies, Proceedings, IGARSS?94, 08?12 August, Pasadena,ure forest heights, but should not be used to establish or
California, 2:939?941.validate DTMs or DEMs.
Blair, J.B., D.L. Rabine, and M.A. Hofton, 1999. The Laser VegetationImprovements to this design are already on the shelf and
Imaging Sensor: A medium-altitude, digitisation-only, airborneshould be considered for incorporation into new PALS-like laser altimeter for mapping vegetation and topography, ISPRS Jour-
instruments. Riegl, and perhaps others, currently sell lasers nal of Photogrammetry and Remote Sensing, 54:115?122.
which sequentially toggle between measurements of first and
Harding, D.L., J.B. Blair, J.G. Garvin, and W.T. Lawrence, 1994. Laserlast returns. Use of this laser would significantly facilitate altimeter waveform measurement of vegetation canopy structure,
ground-finding under dense canopies growing on rugged ter- Proceedings, IGARSS?94, 08?12 August, Pasadena, California,
rain. Certainly, more expensive dGPS systems are available 2:1250?1253.
which differentially correct on the fly without the need for Hickman, G.D., and J.E. Hogg, 1969. Application of airborne pulsed
base stations or U.S. Coast Guard beacons. Unfortunately, these laser for near shore bathymetric measurements, Remote Sensing
units currently cannot leave the country, so they?re ruled out if of Environment, 1:47?58.
the laser might be used for work outside the U.S.A. A third Hoge, F.E., R.N. Swift, and E.B. Frederick, 1980. Water depth measure-
improvement would be the incorporation of tiltmeter data into ment using an airborne pulsed neon laser system, Applied
the laser data stream. Two-axis electronic tiltmeters accurate to Optics, 19(6):871?883.
tenths of a degree can be purchased for a few hundred dollars, Hoge, F.E., R.N. Swift, and J.K. Yungel, 1983. Feasibility of airborne
and incorporation of pitch and roll information would further detection of laser-induced fluorescence emissions from green ter-
refine the ground location of the laser trace. restrial plants, Applied Optics, 22(19):2991?3000.
One problem was noted during the Delaware overflights. A Krabill, W.B., J.G. Collins, L.E. Link, R.N. Swift, and M.L. Butler, 1984.
few of the components reported in this configuration are Airborne laser topographic mapping results, Photogrammetric
Engineering & Remote Sensing, 50(6):685?694.unshielded, and operation of the airborne laser system did
interfere with aircraft-tower communications. Although the Lefsky, M.A., D. Harding, W.B. Cohen, G. Parker, and H.H. Shugart,
1999a. Surface lidar remote sensing of basal area and biomass ininterference could be controlled by reducing the squelch on the
272 March 2003 PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
deciduous forests of eastern Maryland, USA, Remote Sensing of Nelson, R., R. Swift, and W. Krabill, 1988b. Using airborne lasers to
Environment, 67:83?98. estimate forest canopy and stand characteristics, Journal of For-
estry, 86(10):31?38.Lefsky, M.A., W.B. Cohen, S.A. Acker, G.G. Parker, T.A. Spies, and D.
Harding, 1999b. Lidar remote sensing of the canopy structure and Nelson, R., R. Oderwald, and T.G. Gregoire, 1997. Separating the
biophysical properties of douglas-fir western hemlock forests, ground and airborne laser sampling phases to estimate tropical
Remote Sensing of Environment, 70:339?361. forest basal area, volume, and biomass, Remote Sensing of Environ-
Link, L.E., and J.G. Collins, 1981. Airborne laser systems use in terrain ment, 60:311?326.
mapping, Proceedings, 15th International Symposium on Remote
Nilsson, M., 1996. Estimation of tree heights and stand volume usingSensing of Environment, 11?15 May, Ann Arbor, Michigan
an airborne lidar system, Remote Sensing of Environment, 56:1?7.(ERIM) 1:95?110.
Popescu, S., R.H. Wynne, and R.F. Nelson, 2000. Estimating forestMaclean, G.A., and W.B. Krabill, 1986. Gross-merchantable timber vol-
vegetation biomass using airborne lidar measurements, Proceed-ume estimation using an airborne LIDAR system, Canadian Jour-
ings, 2nd International Conference - Geospatial Information innal of Remote Sensing, 12(1):7?18.
Agriculture and Forestry, 10-12 January, Lake Buena Vista, FloridaMeans, J.E., S.A. Acker, D.J. Harding, J.B. Blair, M.A. Lefsky, W.B.
(ERIM), 2:346?353.Cohen, M.E. Harmon, and W.A. McKee, 1999. Use of large-foot-
print scanning airborne lidar to estimate forest stand characteris- ???, 2002. Estimating plot-level tree heights with lidar: Local filter-
tics in the western Cascades of Oregon, Remote Sensing of ing with a canopy-height based variable window size, Computers
Environment, 67:298?308. and Electronics in Agriculture, accepted for publication.
Means, J.E., S.A. Acker, B.J. Fitt, M. Renslow, L. Emerson, and C. Ritchie, J.C., J.H. Everitt, D.E. Escobar, T.J. Jackson, and M.R. Davis,Hendrix, 2000. Predicting forest stand characteristics with air- 1992. Airborne laser measurements of rangeland canopy coverborne scanning lidar, Photogrammetric Engineering & Remote and distribution, Journal of Range Management, 45(2):189?193.Sensing, 66(11):1367?1371.
Ritchie, J.C., D.L. Evans, D. Jacobs, J.H. Everitt, and M.A. Weltz, 1993.N?sset, E., 1997a. Determination of mean tree height of forest stands
Measuring canopy structure with an airborne laser altimeter,using airborne laser scanner data, ISPRS Journal of Photogramme-
Transactions of the American Society of Agricultural Engineers,try and Remot Sensing, 52(2):49?56.
36(4):1235?1238.???, 1997b. Estimating timber volume of forest stands using airborne
laser scanner data, Remote Sensing of Environment, 61:246?253. Schreier, H., J. Lougheed, C. Tucker, and D. Leckie, 1985. Automated
measurements of terrain reflection and height variations usingNelson, R., 1997. Modeling forest canopy heights: The effects of canopy
an airborne infrared laser system, International Journal of Remoteshape, Remote Sensing of Environment, 60:327?334.
Sensing, 6(1):101?113.Nelson, R., W. Krabill, and J. Tonelli, 1988a. Estimating forest biomass
and volume using airborne laser data, Remote Sensing of Environ-
ment, 24:247?267. (Received 21 November 2001; revised and accepted 14 June 2002)
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING March 2003 273