effect of target moisture on laser scanner intensity

2128 IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 48, NO. 4, APRIL 2010

Effect of Target Moisture on Laser Scanner IntensitySanna Kaasalainen, Henri Niittymäki, Anssi Krooks, Katharina Koch,

Harri Kaartinen, Ants Vain, and Hannu Hyyppä

Abstract—We have studied the effect of moisture on thebackscatter reflectance of mineral aggregate (such as sand andgravel) targets that are used as reflectance standards in airbornelaser scanning intensity calibration. Target moisture has turnedout to be a crucial factor when using external reference targets(either commercial or naturally available at the measurement site).The most common targets used thus far are different types ofgravel and sand, which emphasizes the role of moisture because ofthe long time required for drying. We carried out laboratory andfield experiments to find out how moisture affects the performanceof these targets as reference standards. We found that even smallamount of moisture has a crucial effect on target reflectanceproperties: The backscattered reflectance decreases strongly up to10% gravimetric water content, after which the reflectance levelmostly remains constant. The moisture effect has to be furtherstudied and taken into account in the retrieval and calibration oflaser scanner intensity data acquired in moist conditions. We alsodemonstrate a practical method of deriving the surface backscat-tered reflectance at different moisture contents with a digitalcamera.

Index Terms—Laser measurements, laser radar, laser radiationeffects, remote sensing.

I. INTRODUCTION

INTENSITY information from airborne laser scanners hasbeen actively studied in recent years in search for correction

and calibration methods [1]–[6]. The intensity is recorded bycurrent instruments either in the form of echo amplitude or fullwaveform of the backscattered signal. Recently, there have alsobeen efforts for calibration of full-waveform laser scanner data[6], [7]. Since the finding of Hug and Wehr [8], in which the

Manuscript received November 26, 2008; revised August 24, 2009. Firstpublished December 22, 2009; current version published March 24, 2010. Thiswork was supported by the Academy of Finland (Projects “Improving theapplicability of intensity information in laser scanning” and “New techniques inactive remote sensing: Hyperspectral laser in environmental change detection”).

S. Kaasalainen and H. Kaartinen are with the Department of Remote Sensingand Photogrammetry, Finnsih Geodetic Institute, 02431 Masala, Finland(e-mail: [email protected]; [email protected]).

H. Niittymäki is with the Department of Physics, University of Helsinki,00014 Espoo, Finland (e-mail: [email protected]).

A. Krooks is with Helsinki University of Technology, 02150 Espoo, Finland,and also with the Department of Remote Sensing and Photogrammetry, FinnsihGeodetic Institute, 02431 Masala, Finland (e-mail: [email protected]).

K. Koch was with the Department of Remote Sensing and Photogram-metry, Finnsih Geodetic Institute, 02431 Masala, Finland. She is now withthe Technical University of Munich, 80333 Munich, Germany (e-mail:[email protected]).

A. Vain is with the Estonian University of Life Sciences, 51014 Tartu,Estonia, and also with the Department of Remote Sensing and Photogrammetry,Finnsih Geodetic Institute, 02431 Masala, Finland (e-mail: [email protected]).

H. Hyyppä is with the Research Institute of Measuring and Modellingfor the Built Environment, Department of Surveying, Helsinki University ofTechnology, 02150 Espoo, Finland (e-mail: [email protected]).

Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TGRS.2009.2036841

reflectance value was found to be the most reliable discrim-inator in the automatic identification of objects, the intensityhas been used in various studies for object classification (see[4] for an extensive review and references). As the environ-mental applications of both airborne laser scanning (ALS) andterrestrial laser scanning (TLS) are increasing (e.g., in forestryand glaciology [9]–[12]), the use of intensity and waveformdata is becoming more common in the analysis, showing alarge potential in the development of interpretation tools [13]–[15]. Accurate calibration techniques are needed for moreeffective use of the intensity/waveform information in theseapplications. As the signal intensity also affects the distancemeasurement of a laser scanner (e.g., [11]), effective correctionmethods would improve the distance accuracy. This wouldbe particularly important in the increasing implementations oflaser scanners in ground-based mobile platforms [16]–[18]. Theuse of laser scanners is also increasing in data fusion andcomparison efforts, such as combining Light Detection AndRanging (LIDAR) data with, e.g., hyperspectral imaging [19],synthetic aperture radar[20], or surface elevation measurementsfrom spaceborne instruments [10], which increase the potentialof using the radiometrically calibrated intensity data.

In practice, the laser scanner detectors measure the flux ofphotons entering the receiver from a given direction and solidangle (i.e., the scattered radiance). The radiance is related tothe power of the received signal, which is proportional to thebidirectional reflectance distribution function (BRDF) of thetarget, which can be written in terms of laser power as [21]

BRDF =∂2Ps

∂Pi∂Ω1

cos Θs(1)

where Ps and Pi are the scattered and incident powers, re-spectively, Ω is the scattering solid angle, and Θs is the anglebetween the scattering direction and the surface normal. Infact, most commercial reflectance spectrometers measure the(directional) hemispherical reflectance (i.e., albedo), which isthe total fraction of the incident collimated power on a unitsurface area scattered into upper hemisphere by a unit areaof surface [22] and often referred to as “reflectance.” Laserscanners measure only the fraction of the (hemispherical) re-flectance that occurs in the direction of illumination (0◦ anglebetween light source and detector), which can be approximatedby the BRDF at Θs = Θi (Θi being the angle between theincident laser beam and the surface normal), which we here callthe backscattered reflectance.

Since the bidirectional reflectance rather describes the direc-tional scattering properties of a unit area of surface, the targetsize (i.e., the receiving/scattering area) must also be taken into

0196-2892/$26.00 © 2009 IEEE

KAASALAINEN et al.: EFFECT OF TARGET MOISTURE ON LASER SCANNER INTENSITY 2129

account. The power entering the laser scanner detector alsodepends on the sensor parameters and the distance to the target.The measurement geometry, scattering properties of the target,and the sensor parameters are all summarized by the radarequation, which defines the received power Pr as [7]

Pr =PtD

2r

4πR4β2t

σ (2)

where Pt is the transmitted power, Dt is the receiver aperture,R is the range, and βt is the transmitter beamwidth. σ is thetarget backscatter cross section, defined as

σ =4π

ΩρAs (3)

where Ω is the scattering solid angle, As is the receiving area,and ρ describes the reflectance of the target (see [4] and [6]for more detailed reviews of the physical concepts of laserscanner intensity correction). While the use of backscatteringcross section is common in radar remote sensing to describethe target scattering properties, the calibration of received laserpower into values that are proportional to target reflectancewould help the comparison of ALS intensity data with thoseobtained from, e.g., aerial imagery, reflectance spectroscopy,or field reference measurements carried out to validate theradiometric calibration. In these cases, the results are oftenpresented in terms of reflectance.

The signal received by a laser scanner detector is also af-fected by several instrumental and atmospheric factors. There-fore, corrections must be applied to convert the received powerPr (which is often called simply “intensity”) into a value that isproportional to the target reflectance [4] or, at first point, to thebackscattered reflectance (BRDF at Θi = 0) of the surface. Theradar equation [see (2)] includes the sensor and target parame-ters, as well as the measurement geometry (the scattering solidangle), but all the sensor parameters are not known (or reportedby the manufacturer) at the time of the measurement, e.g., thetransmitted power Pt may be subject to continuous variation(such as automatic gain control to reduce the signal for brighttargets). The atmospheric attenuation must also be corrected torelate Pt to the scattered power Ps and, hence, to the targetreflectance. The recent calibration approaches for laser scannerintensity data are based on either investigating the power inci-dent on the target (related to the laser power of the scanner)or using external targets of known reflectance measured in thesame conditions (reflectance standards). Few applications ofreflectance standards in ALS are thus far available [5], [23],and results from further validation campaigns (laboratory andin situ) are needed to make these techniques fully operative.One of the challenges is related to the backscatter geometry(where the incident and reflected light paths coincide; see [5]for more references), which strongly increases the scatteredflux and causes uncertainty in reflectances measured in thebackward direction [24]. Since most of the remote-sensing landtargets are not isotropic scatterers, further studies are neededto find out how well the backscattered reflectance describesthe hemispherical reflectance properties of the targets. Thebackscattered reflectance could however be used in data in-

terpretation algorithms, e.g., target classification and changedetection.

Another factor that may significantly complicate theALS-based reflectance measurement and calibration is the pres-ence of water (moisture) in the targets, particularly in borealregions, where weather conditions (such as rainfall or dew)often disturb the ALS campaigns. As an important climate andhydrologic parameter, soil moisture is a widely studied subject,and measurement techniques range from in situ gravimetricand moisture probes to microwave remote sensing [25], [26].While it is known that soil moisture can be retrieved from radarbackscatter [25], [27], there is little information available on theeffects of moisture on the backscattered signal of laser scanners.LIDAR digital terrain models (digital elevation model) havebeen used in hydrologic applications, such as analyzing the soilwater content [28].

We have developed a practical approach for radiometriccalibration of airborne and terrestrial laser scanner intensitydata, based on commercially available sand and gravel beingused as reflectance standards [5], [29], for which the backscat-ter reflectance is calibrated in the laboratory. To establisha reference-based calibration method, repeated measurementcampaigns were carried out with both airborne and terrestriallaser scanners, as well as laboratory experiments. These cam-paigns included the testing and search for practical and stabletargets that produce statistically meaningful results (see [5] formore details). Validation campaigns are underway, as well asthe first tests and demonstrations of the calibration techniquesin environmental applications.

In this paper, we investigate the performance of ALS radio-metric calibration standards when small amounts of moistureare present in order to study the accuracy of reflectance mea-surement they provide. In the boreal climate, the ALS appli-cations often suffer from moisture due to weather conditions(e.g., flight campaign occurs shortly after rain), and informationon the usability of land targets as reflectance standards inwet conditions would therefore be valuable concerning theincreasing number of ALS activities in the northern latitudes.It will also be important to find out whether it is possible toestimate the backscatter reflectance of the ALS land targets inwet conditions and how much this affects the interpretation ofdata (e.g., classification of targets).

II. MEASUREMENTS AND DATA

A. Laboratory Experiments

To get an overall image about the effect of moisture onbackscattered reflectance, we carried out a laboratory exper-iment for sand and gravel samples used in ALS intensity-calibration campaigns [5]. The moisture was measured bymeans of gravimetric analysis by adding water into the sam-ple of certain weight [about 0.5 kg, except 0.25 kg for lightexpanded clay aggregate (LECA)] in 10-mL increments andmeasuring the backscatter reflectance after each addition. Ateach step, the samples were thoroughly stirred for even distri-bution of moisture and weighed to determine the gravimetricwater content of the sample. Water was added up to the point ofsaturation, i.e., where the samples were thoroughly soaked, and


Fig. 1. Outdoor (field) sample of sandblasting sand of 0.1–0.6 mm.

the backscatter measurement indicated a strong reflection fromwater rather than the moistened gravel itself. Seven sampleswere measured: sandblasting sand with two different grain sizesof 0.1–0.6 and 0.5–1.2 mm, coarse gravel used for sanding theroads (sanding gravel), crushed redbrick, crushed LECA, whitegabbro gravel, and a beach-sand sample from Kivenlahti, Espoo(Finland). The beach-sand sample has been a ground sample ofrepeated previous ALS flight campaigns [5] and represents anexample of typical ground targets that could be used in ALSintensity calibration.

The backscatter-reflectance measurement was carried outwith a laboratory laser instrument consisting of a 10-mW1064-nm Nd:YAG (neodymium-yttrium aluminum garnet)laser and a 16-bit monochrome digital camera (Sbig ST-7CCD). To observe the geometry of a laser scanner (i.e., atbackscatter), a beam splitter was used: The laser beam was firstreflected into the sample by the beam splitter, and the retrore-flected signal was observed through the beam splitter with thecharge-coupled device (CCD) detector placed above the beamsplitter and sample. Detailed descriptions of the instrument arein [5] and [30]. Spectralon (Labsphere Inc.) reference plate with99% reflectance was used as a standard. Each data point is anaverage of five 5-s CCD exposures.

B. Field Data

To investigate the role of surface drying on laser backscatterreflectance, a field sample of sandblasting sand of 0.1–0.6 mmwas also prepared (see Fig. 1) by laying out a 6-cm layer ontop of a filter fabric. The sample size was about 1 m × 1 m.Water was added up to over 20% volumetric moisture, whichwas measured with a ThetaKit TK2-BASIC soil moisture-measuring device (Delta-T Devices). The device has steelspikes that measure the moisture from a 6-cm layer. The samplewas left outdoor (in sunny conditions, 24 ◦C temperature) for4 h, during which the surface was dried in the sun. A samplewas taken from the surface at intervals from 10 to 30 min, andthe backscattered reflectance was measured with the laboratorylaser instrument described in Section II-A. The volumetricmoisture content was measured simultaneously.

C. Near-Infrared Digital Camera

We also made an effort to measure the surface moistureusing an Exotec MC-60A wood and building moisture meter,

Fig. 2. Setup for the digital camera experiment. The camera is facing down-ward to the Spectralon reference panel (the white 99% strip was used as areference in the calibration).

based on measuring the dielectric constant. Most of thesemeters are, however, not designed for sand-type samples (whichresulted in, for example, incorrect moisture values for sandand gravel), whereas the soil moisture probes (such as theThetaKit used in the field experiment) rather measure thetotal moisture instead of just the target surface. Therefore,we made an experiment with a digital camera system con-sisting of Fuji IS PRO with an 850-nm infrared (IR) filterand Nikon SB800 flash, for which the output power variationwas ±2%. Similar approach was used in [31] for the determi-nation of soil color indexes using a digital camera. The useof flash means that the measurement technique is active, andthe noise caused by external illumination (such as sunlight)is minimized in two ways: 1) using a reflectance standard(such as that in Fig. 2) photographed in similar conditions and2) minimizing the exposure time and camera aperture (we used1/250 s and F/16 aperture in this study).

The camera was mounted on a tripod with the lens facingdownward to the sample, which was placed about 70 cm belowthe camera objective (Fig. 2). The 14-bit raw format imageswere exported into linear 16-bit Tagged Image File Formatimages, and the green channel intensity values were recorded(using ImageJ software). The 99% Spectralon panel was againused as a reflectance standard, and it was photographed andsampled similarly to the gravel samples. Four gravel sampleswere measured (sandblasting sand of 0.1–0.6-mm grain size,sanding gravel, crushed redbrick, and crushed LECA), andwater was added in 10-mL increments, similarly to the laserand CCD experiment described earlier. To investigate the roleof surface moisture, some dry gravel was added on the surfaceof the wet sample to observe the difference in backscatteredreflectance compared to the samples with evenly distributedwater content.


Fig. 3. Palokangas forest road (top) at 8:33 A.M. after inserting water and(bottom) at 9:30 A.M. The first Leica ALS50-II flight occurred at 10:15 A.M.

D. Airborne Laser Scanner

Two case studies with dry and wet targets were carried outwith both airborne and terrestrial laser scanners to demonstratethe effect of wetness on laser scanner intensity. The ALSexperiment took place on July 17, 2008, during the LeicaALS50-II (1064 nm) campaign in Palokangas forest area (eastof Finland near Ilomantsi) in a sand-covered forest road. Waterwas introduced to a 7-m stretch of the road, while the other partsremained dry for comparison (see Fig. 3). The moisture wasmeasured with the ThetaKit TK2 moisture probe, after whichthe first ALS flight occurred within 2 h. Some 195 and 167 laserhits from the dry and wet roads, respectively, were analyzedfrom two flight lines, which occurred at about 10:15 A.M. and13:45 P.M. The flight altitudes were 520 and 550 m for thoseflight lines. More details of the ALS radiometric calibration arein [5].

E. Terrestrial Laser Scanner (FARO)

The TLS case study was carried out using the FAROLS880HE80 (785 nm) scanner. The scanner used a phasemodulation technique for the distance measurement with anaccuracy of 3–5 mm and 360◦ × 320◦ field of view. Threegravel targets (and a four-step Spectralon reflectance standardof 12%, 25%, 50%, and 99% reflectances) were scanned. Thegravel targets included coarse gravel used for sanding the roads,crushed redbrick, and sandblasting sand with a grain size of0.1–0.6 mm (all these samples were also investigated in thelaser laboratory experiment described in Section II-A). Both dryand thoroughly soaked samples were scanned for each graveltarget. The close-up images of the samples are shown in Fig. 4.A FARO intensity image showing the samples and the standardis shown in Fig. 5.

Instrumental corrections to the FARO intensity data werecarried out for the effects of the amplifier and reducer toget a linear brightness scale, based on instrument calibration

Fig. 4. Close-up images of the samples measured with the FARO scanner.(Top) Sandblasting sand of 0.1–0.6 mm. (Middle) Crushed redbrick. (Bottom)Sanding gravel.

Fig. 5. FARO intensity image of the (top row) dry and (bottom row) soakedsamples: (Left) Sanding gravel, (middle) crushed redbrick, and (right) sand-blasting sand. The Spectralon reference plate is placed behind the samples.

experiments for both distance and reflectance scale (see [30] formore details). The average error in the backscattered reflectancein our previous experiments has been on the order of 1%–2%for FARO [5]. The (raw) intensity data were calibrated usingthe 99% Spectralon reference plate.

III. RESULTS

Table I provides a sketch of the results from different mea-surements, which are discussed in this section. In the ideal case,the ground reference measurements for ALS intensity shouldbe carried out at the same wavelength as in the scanner. Inpractice, however, it may occur that the wavelengths of theALS and reference sensor(s) are different. For example, theperformance of a digital camera is most ideal below 1000 nm.


TABLE ICHART OF RESULTS FROM ALL DIFFERENT EXPERIMENTS. THE ALS

RESULTS ARE DISCUSSED IN SECTION III-D. NOTE THAT THE

LABORATORY AND FIELD SAMPLES WERE MEASURED

WITH THE SAME INSTRUMENT

Our first studies with the digital camera indicate a good agree-ment with laboratory laser (1064-nm) measurements [32]. Aspectroscopic study of the targets presented in this paper hasshown that the wavelength differences for most of these targetsare small and do not play a strong role [5]. Another reason forusing several different instruments (including some differencesin wavelength) is to know whether the measurement of themoisture effect, and the intensity effects in the first place, is in-dependent on the instrument (as it should be to be meaningful).

Another question is the footprint versus surface inhomo-geneity, which is significantly different for laboratory and ALSexperiments. As the laboratory laser beam diameter is about5 mm, we rotated the sample during the exposures to get datafrom a larger area (than just the laser spot). However, the near-infrared (NIR) digital-camera-based approach is more practicalfor collecting ALS ground reference since the image field ofview is larger than a small laboratory sample, and images can betaken and averaged from a large area in the target to reduce theeffect of surface inhomogeneity (such as that shown in Fig. 3 forthe wet surface area); see [32] for more details. Nevertheless,the laboratory laser experiment provides information on thesurface scattering properties in the exact backscatter (and thebackscattered reflectance of the surface material), which iscrucial in the calibration of ALS data.

A. Laboratory Experiments

The gravimetric moisture-content versus backscatter-reflectance graphs measured in laboratory are shown inFigs. 6–8. All samples (except crushed redbrick, for whichdeviation in data played a strong role) showed somewhatsimilar trend in backscatter reflectance: The decrease inbackscattered reflectance was most prominent after the firstaddition of water (10 mL, which corresponded to about 2%water content, except for LECA, where the water contentwas almost 4% because of the light weight of the samplecompared with the others). After the following additions, thebackscattered reflectance still decreased, but at certain point(at about 10%–15% water content), there was no significantchange until beyond the level of saturation, where liquid waterstarted to corrupt the measured intensities (those points havebeen excluded from Figs. 6–8). The moisture content wherethe saturation happened was different from target to target,depending on the grain size and the porosity of the grains.The deviation in the crushed-redbrick data may be due to the

Fig. 6. Laboratory experiments: Gravimetric water content (in percentages ofweight) versus backscatter reflectance at 1064 nm for sandblasting sand with(top) 0.1–0.6-mm and (bottom) 0.5–1.2-mm grain sizes. Each data point is anaverage of five CCD exposures. The error bars are the standard deviations.

fact that the grains are large and apparently more stronglyabsorbing than, e.g., those of the sanding gravel and sand.

B. Field Measurement

The result for the 0.1–0.6-mm sandblasting-sand field exper-iment is shown in Fig. 9. Comparison with the laboratory data(Fig. 6) shows that the backscatter-reflectance values for bothsamples at moisture content above 20% are at the same level,while the increase in backscatter reflectance toward smallermoisture content is stronger for the outdoor sample. This maybe caused by the fact that its surface dried faster than the entirebulk, causing the surface reflectance to be higher than that forthe laboratory sample with even moisture distribution. Duringthe 4 h in open-air (sunny) conditions, the sample dried up to3% volumetric moisture content, and the overall backscatter-reflectance trend is similar to the laboratory measurement. Itmust be noted, however, that there is some inaccuracy in theThetaKit meter results, but it still demonstrates the fact that theoverall moisture is not the primary factor in surface-reflectanceproperties.

C. Digital Camera

The backscatter reflectances measured with the digital-camera system are shown in Figs. 10 and 11, with compar-ison with the laser measurements, up to the point where thewater content no longer affected the laser results (as shownin Figs. 6–8). The error for the camera measurements wasestimated from the standard deviations of five separate camera


Fig. 7. Laboratory experiments: Gravimetric water content versus backscatterreflectance at 1064 nm for (top) sanding gravel, (middle) crushed redbrick, and(bottom) crushed LECA, similar to that in Fig. 6.

exposures to be about 2.5% for gravel and sand-type samples.The decrease in the backscatter reflectance is evident in thedigital camera results, similar to the laser backscatter mea-surement, even though there are differences in the backscatter-reflectance levels, particularly for the dry samples (i.e., near 0%water content). The difference may be due to the physical dif-ferences in scattering properties caused by the variation in themeasurement geometry between laboratory and digital camerasetups, but the relative decrease in the backscattered reflectanceas the moisture increases is well measurable, making the digitalcamera a feasible method of surface moisture versus reflectanceevaluation.

We also investigated the role of surface wetness versusgravimetric water content by adding a surface of dry materialon top of the wet sample. This was done to simulate thesituation where the sample surface had dried (e.g., in the sun)while the layers under the surface still contain moisture. Forall samples, the surface backscatter reflectance increased close

Fig. 8. Laboratory experiments: Gravimetric water content versus backscatterreflectance at 1064 nm for (top) white gabbro and (bottom) Kivenlahti beachsand, similar to that in Fig. 6.

Fig. 9. Field experiment: Backscatter reflectance versus water percentage ofthe 0.1–0.6-mm sandblasting-sand field sample. The time interval between thefirst (3% water percentage) and last (22%) data points was 4 h.

to the original (0% water) level right after the addition. Thosedata points are also shown in Figs. 10 and 11. The remarkabledifference in the backscatter reflectance between the evenlydistributed water content and the dry-surfaced sample indicatesthat the overall water content is not the primary factor in surfacereflectance, and therefore, the surface moisture/reflectance mustbe measured separately during ALS field campaigns in suchweather conditions where, e.g., traces of recent rain might stillaffect the surface-scattering properties.

D. Case Study With Airborne Laser Scanner

The wetness was clearly visible in both Leica ALS50-II flightlines. There was a 27% decrease in backscatter reflectance


Fig. 10. Digital camera: Moisture content (gravimetric water content) versusbackscattered reflectance from digital camera images (camera) compared withthe laboratory laser backscatter measurement (laser; also shown in Figs. 6 and7), for (top) sandblasting sand of 0.1–0.6-mm grain size and (bottom) sandinggravel. The “Surface” points, plotted in yellow, are the surface backscatter-reflectance values when dry sand/gravel was added on top of the sample afterthe last addition of water.

Fig. 11. Digital camera: Gravimetric moisture content versus backscatteredreflectance from digital camera images (camera) compared with the laboratorylaser backscatter measurement (laser; also shown in Fig. 7), for (top) crushedredbrick and (bottom) crushed LECA. Similar to that in Fig. 10, the yellow“Surface” points represent the surface backscatter reflectance when a drymaterial was added on top of the sample after the last addition of water.

Fig. 12. TLS case study for dry and wet samples with FARO terrestrial laserscanner: Sanding gravel, crushed redbrick, and sandblasting sand (0.1–0.6 mm).

between wet and dry stretches of road in the first (10:15) flightline. In the second (13:45) flight line, the decrease was only 7%,which means that the surface had dried significantly betweenthese two flight lines. The original volumetric water percentagewas about 4%. Comparison with laboratory data for similar typeof sand samples, e.g., sandblasting sand (0.1–0.6 mm) in Fig. 6,for which the drop in backscatter reflectance was on the order of47% from 0% to 4% (gravimetric) moisture, indicates that thewet stretch of road had dried from the original 4% volumetricmoisture even before the first flight line, which occurred about2 h after the addition of water. It must be noted, however, thatthese results are preliminary and call for more extensive andquantitative analysis of moisture–reflectance relations in caseswhere airborne laser scanner data taken at wet conditions haveto be analyzed.

E. Case Study With Terrestrial Laser Scanner

Fig. 12 shows the backscatter reflectance of sanding gravel,crushed redbrick, and sandblasting sand (0.1–0.6-mm grainsize) extracted from the FARO TLS intensity data. The intensity(raw) data were corrected for the nonlinear detector effectsand calibrated into backscattered reflectance using the 99%Spectralon standard (see Section II-E and Fig. 5, where theSpectralon panel is shown). Strong decrease in backscatteredreflectance is observed for all samples. The effect is also clearlyvisible in the intensity image (Fig. 5), in spite of slightlydifferent incidence angles for wet and dry samples, which mighthave marginally smoothed out the effect (cf. [33]). Althoughthe wavelength difference and deviation in data typical tolaser scanner geometry (see [5] and [34] for more details)produce backscatter-reflectance values that are different fromthe laboratory measurement, the relative decrease in backscatterreflectance (compared in Fig. 13) is similar for each samplein both laboratory and FARO results. The decrease in thebackscattered reflectance is about 70% for the crushed redbrickand about 45% for the sandblasting sand in both measurements,whereas for the sanding gravel, there is more deviation. Thismay indicate that, when a certain level of water content (hereabout 10%–12% gravimetric moisture) is reached, the decreasein backscatter reflectance remains the same, i.e., the reflectance


Fig. 13. TLS: Relative differences in backscatter reflectance for wet and drysamples measured with FARO (785 nm) and with the 1064-nm laboratory laser(i.e., the ratio of wet and dry sample backscatter reflectance in percentages).The wet sample backscatter reflectance for laboratory samples represents thevalue of about 12% gravimetric water content.

behavior of a wet sample stabilizes (up to the point where liquidwater starts to dominate in the bulk sample).

IV. CONCLUSION

We have tested some practical methods to evaluate the de-crease in backscatter reflectance related to the surface moistureof typical ALS land targets, which could also be used asreferences in the ALS radiometric calibration (cf. [23] and[32]). Since most moisture measurement devices are designedto measure the water content of the entire sample (or at least abulk of a sample) rather than just the surface, the gravimetricwater content turned out to be the most practical measurableunit of target moisture to get some reference for the laboratorymeasurements. However, since the laser scanners only measurethe surface of the target, the gravimetric water content of theentire bulk may not describe the surface wetness properties,particularly when the surface has dried up faster. Therefore,the change in surface reflectance might be the best measureof surface moisture, for which a digital-camera-based approachturned out to be a promising method.

The laboratory experiments show that small water contents(up to 10%) are most crucial to the reflectance measurement,and call for accurate field methods of surface moisture de-termination of ALS land targets. At a certain level of watercontent, no more decrease in backscatter reflectance was ob-served, i.e., the reflectance behavior of a wet sample stabilizes,which occurs when the saturation is reached. However, this maynot make a difference for ALS campaigns since other factorsmay restrict or even prevent the scanner operation in very wetconditions.

We have also presented a digital-camera-based approach tomeasure the backscatter reflectance for typical ALS referencetargets (see also [32]). We have shown that the effect of surfacemoisture is visible in the NIR images of these targets. Thisinformation can be used to determine the decrease in targetbackscatter reflectance, e.g., with IR camera-based surfacebackscatter reflectance measured in situ (simultaneously withthe flight) and subsequently after the sample has dried, eitherin the field or in the laboratory. Further studies from future

ALS campaigns are needed to validate and test the performanceof this approach in both dry and wet conditions and also toinvestigate how well the backscatter reflectance measured fromthe images actually represents the surface reflectance of thetargets.

Since ALS flights are often carried out in moist (or evenalmost wet) conditions, information on the effect of moistureon the intensity data is important to evaluate the usability ofsuch data in, e.g., target classification and comparison. Theperformance of wet surfaces as reference targets and also themeasurement of wet land targets in the first place, as well asthe methods of wetness determination, must be further evalu-ated in future ALS flight campaigns.

ACKNOWLEDGMENT

The authors would like to thank E. Ahokas for the helpwith the Palokangas data and all others who participated in thecampaign, particularly P. Lyytikäinen-Saarenmaa who helpedin searching for the ground targets, arrangements, etc. Theauthors would also like to thank W. Wagner and M. Kaasalainenfor lots of comments and discussion and, particularly, J. Hyyppäfor the comments on this paper.

REFERENCES

[1] F. Coren and P. Sterzai, “Radiometric correction in laser scanning,” Int. J.Remote Sens., vol. 27, no. 15/16, pp. 3097–3104, Aug. 2006.

[2] E. Ahokas, S. Kaasalainen, J. Hyyppä, and J. Suomalainen, “Calibrationof the Optech ALTM 3100 laser scanner intensity data using bright-ness targets,” in Proc. ISPRS Comm. I Symp., Int. Arch. Photogramm.,Remote Sens. Spatial Inf. Sci., Marne-la-Vallee, France, Jul. 3–6, 2006,p. 36(A1).

[3] C. Hopkinson, “The influence of flying altitude, beam divergence, andpulse repetition frequency on laser pulse return intensity and canopyfrequency distribution,” Can. J. Remote Sens., vol. 33, no. 4, pp. 312–324,Aug. 2007.

[4] B. Höfle and N. Pfeifer, “Correction of laser scanner intensity data: Data-and model-driven approaches,” ISPRS J. Photogramm. Remote Sens.,vol. 62, no. 6, pp. 415–433, Dec. 2007.

[5] S. Kaasalainen, H. Hyyppä, A. Kukko, P. Litkey, E. Ahokas, J. Hyyppä,H. Lehner, A. Jaakkola, J. Suomalainen, A. Akujärvi, M. Kaasalainen,and U. Pyysalo, “Radiometric calibration of LIDAR intensity with com-mercially available reference targets,” IEEE Trans. Geosci. Remote Sens.,vol. 47, no. 2, pp. 588–598, Feb. 2009.

[6] W. Wagner, J. Hyyppä, A. Ullrich, H. Lehner, C. Briese, andS. Kaasalainen, “Radiometric calibration of full-waveform small-footprintairborne laser scanners,” in Int. Arch. Photogramm., Remote Sens. SpatialInf. Sci., Beijing, China, 2008, vol. XXXVII, pp. 163–168.

[7] W. Wagner, A. Ullrich, V. Ducic, T. Melzer, and N. Studnicka, “Gaussiandecomposition and calibration of a novel small-footprint full-waveformdigitising airborne laser scanner,” ISPRS J. Photogramm. Remote Sens.,vol. 60, no. 2, pp. 100–112, Apr. 2006.

[8] C. Hug and A. Wehr, “Detecting and identifying topographic objects inlaser altimeter data,” in Int. Arch. Photogramm., Remote Sens. Spatial Inf.Sci., Stuttgart, Germany, 1997, vol. 32, pp. 19–26.

[9] F. M. Danson, D. Hetherington, F. Morsdorf, B. Koetz, andB. Allgower, “Forest canopy gap fraction from terrestrial laserscanning,” IEEE Geosci. Remote Sens. Lett., vol. 4, no. 1, pp. 157–160,Jan. 2007.

[10] N. T. Kurtz, T. Markus, D. J. Cavalieri, W. Krabill, J. G. Sonntag, andJ. Miller, “Comparison of ICESat data with airborne laser altimeter mea-surements over arctic sea ice,” IEEE Trans. Geosci. Remote Sens., vol. 46,no. 7, pp. 1913–1924, Jul. 2008.

[11] A. Prokop, “Assessing the applicability of terrestrial laser scanning forspatial snow depth measurements,” Cold Reg. Sci. Technol., vol. 54, no. 3,pp. 155–163, Nov. 2008.

[12] C. Wang and N. F. Glenn, “Integrating LIDAR intensity and elevation datafor terrain characterization in forest area,” IEEE Geosci. Remote Sens.Lett., vol. 6, no. 3, pp. 463–466, Jul. 2009.


[13] D. N. M. Donoghue, P. J. Watt, N. J. Cox, and J. Wilson, “Remotesensing of species mixtures in conifer plantations using LIDAR heightand intensity data,” Remote Sens. Environ., vol. 110, no. 4, pp. 509–522,Oct. 2007.

[14] I. S. Korpela, “Mapping of understory lichens with airborne discrete-return LIDAR data,” Remote Sens. Environ., vol. 112, no. 10, pp. 3891–3897, Oct. 2008.

[15] J. Boudreau, R. F. Nelson, H. A. Margolis, A. Beaudoin, L. Guindon, andD. S. Kimes, “Regional aboveground forest biomass using airborne andspaceborne LIDAR in Québec,” Remote Sens. Environ., vol. 112, no. 10,pp. 3876–3890, Oct. 2008.

[16] D. M. Barber and J. P. Mills, “Vehicle based waveform laser scanning ina coastal environment,” in Proc. 5th Int. Symp. MMT , 2007.

[17] A. Kukko, C.-O. Andrei, V.-M. Salminen, H. Kaartinen, Y. Chen,P. Rönnholm, H. Hyyppä, J. Hyyppä, R. Chen, H. Haggrén,I. Kosonen, and K. Capek, “Road environment mapping system of theFinnish Geodetic Institute—FGI ROAMER,” in Proc. ISPRS WorkshopLaser Scanning SilviLaser. Int. Arch. Photogramm., Remote Sens. SpatialInf. Sci., Espoo, Finland, Sep. 12–14, 2007, vol. XXXVI, pp. 241–247.

[18] A. Jaakkola, J. Hyyppä, H. Hyyppä, and A. Kukko, “Retrieval algorithmsfor road surface modelling using laser-based mobile mapping,” Sensors,vol. 8, no. 9, pp. 5238–5249, 2008.

[19] M. Dalponte, L. Bruzzone, and D. Gianelle, “Fusion of hyperspectraland LIDAR remote sensing data for classification of complex forestareas,” IEEE Trans. Geosci. Remote Sens., vol. 46, no. 5, pp. 1416–1427,May 2008.

[20] D. C. Mason, M. S. Horritt, J. T. Dall’Amico, T. R. Scott, and P. D. Bates,“Improving river flood extent delineation from synthetic aperture radarusing airborne laser altimetry,” IEEE Trans. Geosci. Remote Sens., vol. 45,no. 12, pp. 3932–3943, Dec. 2007.

[21] J. C. Leader, “Analysis and prediction of laser scattering from rough-surface materials,” J. Opt. Soc. Amer., vol. 69, no. 4, pp. 610–628,Apr. 1979.

[22] B. Hapke, Theory of Reflectance and Emittance Spectroscopy.Cambridge, U.K.: Cambridge Univ. Press, 1993.

[23] C. Briese, B. Höfle, H. Lehner, W. Wagner, M. Pfennigbauer, A. Ullrich,and C. Doppler, “Calibration of full-waveform airborne laser scanningdata for object classification,” in Proc. SPIE—Laser Radar Technologyand Applications XIII, M. D. Turner and G. W. Kamerman, Eds., Orlando,FL, 2008, vol. 6950.

[24] Y. L. Kim, Y. Liu, V. M. Turzhitsky, H. K. Roy, R. K. Wali, andV. Backman, “Coherent backscattering spectroscopy,” Opt. Lett., vol. 29,no. 16, pp. 1906–1908, Aug. 2004.

[25] A. T. Joseph, R. van der Velde, P. E. O’Neill, and R. H. Lang, “Soilmoisture retrieval during a corn growth cycle using l-band (1.6 GHz)radar observations,” IEEE Trans. Geosci. Remote Sens., vol. 46, no. 8,pp. 2365–2374, Aug. 2008.

[26] J. E. Lawrence and G. M. Hornberger, “Soil moisture variability acrossclimate zones,” Geophys. Res. Lett., vol. 34, p. L20 402, Apr. 2007.DOI:10.1029/2007/GL031382.

[27] G. Macelloni, S. Paloscia, P. Pampaloni, E. Santi, and M. Tedesco, “Mi-crowave radiometric measurements of soil moisture in Italy,” Hydrol.Earth Syst. Sci., vol. 7, no. 6, pp. 937–948, 2003.

[28] S. P. Wechsler, “Uncertainties associated with digital elevation models forhydrologic applications: A review,” Hydrol. Earth Syst. Sci., vol. 11, no. 4,pp. 1481–1500, Aug. 2007.

[29] S. Kaasalainen, J. Hyyppä, P. Litkey, H. Hyyppä, E. Ahokas, A. Kukko,and H. Kaartinen, “Radiometric calibration of ALS intensity,” in Proc.ISPRS Workshop Laser Scanning SilviLaser. Int. Arch. Photogramm.,Remote Sens. Spatial Inf. Sci., Espoo, Finland, Sep. 12–14, 2007,vol. XXXVI, pp. 201–205.

[30] S. Kaasalainen, A. Kukko, T. Lindroos, P. Litkey, H. Kaartinen, J. Hyyppä,and E. Ahokas, “Brightness measurements and calibration with airborneand terrestrial laser scanners,” IEEE Trans. Geosci. Remote Sens., vol. 46,no. 2, pp. 528–534, Feb. 2008.

[31] N. Levin, E. Ben-Dor, and A. Singer, “A digital camera as a tool tomeasure colour indices and related properties of sandy soils in semi-arid environments,” Int. J. Remote Sens., vol. 26, no. 24, pp. 5475–5492,Dec. 2005.

[32] A. Vain, S. Kaasalainen, U. Pyysalo, A. Krooks, and P. Litkey, “Use ofnaturally available reference targets to calibrate airborne laser scanningintensity data,” Sensors, vol. 9, no. 4, pp. 2780–2796, Apr. 2009.

[33] A. Kukko, S. Kaasalainen, and P. Litkey, “Effect of incidence angleon laser scanner intensity and surface data,” Appl. Opt., vol. 47, no. 7,pp. 986–992, Mar. 2008.

[34] M. Kaasalainen and S. Kaasalainen, “Aperture size effects on backscatterintensity measurements in earth and space remote sensing,” J. Opt. Soc.Amer. A, Opt. Image Sci., vol. 25, no. 5, pp. 1142–1146, May 2008.

Sanna Kaasalainen received the M.Sc. degree in physics and the Ph.D. degreein astronomy from the University of Helsinki, Espoo, Finland, in 1998 and2003, respectively.

She is currently an Academy of Finland Research Fellow with the De-partment of Remote Sensing and Photogrammetry, Finnish Geodetic Institute,Masala, Finland. Her current research interests are the development of laser-based remote sensing methods, laser reflectance measurements, laser physics,and their application in environmental remote sensing, particularly snow/forestmonitoring, as well as laboratory reference experiments for Earth and spaceremote sensing.

Henri Niittymäki received the M.Sc. degree from the University of Helsinki,Espoo, Finland, in 2007, where he is currently working toward the Ph.D. degree.

He was a Research Scientist in the Department of Remote Sensing andPhotogrammetry, Finnish Geodetic Institute, Masala, Finland, in 2007–2008.

Anssi Krooks is currently working toward the M.Sc. degree at the HelsinkiUniversity of Technology, Espoo, Finland.

He is currently a Research Assistant with the Department of Remote Sensingand Photogrammetry, Finnish Geodetic Institute, Masala, Finland.

Katharina Koch is currently working toward the M.Sc. degree in electricalengineering at the Technical University of Munich, Munich, Germany.

She was a Research Assistant with the Department of Remote Sensing andPhotogrammetry, Finnish Geodetic Institute, Masala, Finland, in 2008.

Harri Kaartinen received the M.Sc. degree in geodesy from the HelsinkiUniversity of Technology, Espoo, Finland, in 1993.

He is currently a Senior Research Scientist with the Department of RemoteSensing and Photogrammetry, Finnish Geodetic Institute, Masala, Finland. Hisresearch interests are laser-based mapping methods, use of terrestrial laserscanners in traffic and construction applications, reference measurements, andmobile mapping.

Ants Vain received the M.Sc. degree from the Estonian University of LifeSciences, Tartu, Estonia, in 2008, where he is currently working toward thePh.D. degree.

He is currently a Research Scientist with the Department of Remote Sensingand Photogrammetry, Finnish Geodetic Institute, Masala, Finland.

Hannu Hyyppä received the M.Sc., Lic.Tech., and D.Tech. degrees from theHelsinki University of Technology, Espoo, Finland, in 1986, 1989, and 2000,respectively.

He is currently an Academy of Finland Research Fellow with the ResearchInstitute of Measuring and Modelling for the Built Environment, Departmentof Surveying, Helsinki University of Technology. His specialization rangesfrom laser scanning, photogrammetry, geoinformatics, and remote-sensingapplications to computer-aided civil engineering.

effect of target moisture on laser scanner intensity

Documents