towards autonomous control of quadrotor unmanned aerial vehicles in gps denied urban areas via laser range finer

c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ier .com/ locate /compag

Comparison of two 2D laser scanners for sensing objectdistances, shapes, and surface patterns

Kyeong-Hwan Lee, Reza Ehsani ∗

University of Florida, IFAS, Agricultural and Biological Engineering Department, Citrus Research and Education Center,700 Experiment Station Road, Lake Alfred, FL 33850, USA

a r t i c l e i n f o

Article history:

Received 25 April 2007

Received in revised form

10 July 2007

Accepted 17 August 2007

Keywords:

Laser scanner

Measurement drift

Measurement accuracy

a b s t r a c t

Laser scanners are increasingly used in automation and robotic applications as a sensing

device for navigation and safety. They have agricultural applications in measuring plant

growth rate, tree volume, tree count, 3D imaging, and pattern recognition. Laser scanners

are commercially available, but there is very little published information on characteristics

and performance of these laser scanners. This study compared two laser scanners, the Sick

LMS200 and the Hokuyo URG-04LX, for measurement drift over time, the effect of material

and color on measurement accuracy, and the ability to map different surface patterns.

Measurement drift over time was studied by determining the distance between the laser

scanner sensor and a stationary object at different fixed distances and angles. Distance

measurements over time fluctuated with a peak-to-peak value of 10–20 mm. The settling

time, which is the time required for the averaged distance data to reach a stable level,

increased when measurement distance increased but for a given distance, the settling time

remained constant for different angles. At the measurement angle of 90◦, the settling times

for the LMS200 and the URG-04LX for 50% of the maximum scanner measurement distances

were 53 min and 70 min, respectively. Therefore, to obtain accurate distance measurements,

the laser scanners should be warmed up for the duration of the settling time before recording

measurement data.

The measured distance for soft material objects, such as a styrofoam plate and a sheet

of dry sponge, was longer than the actual distance. For shiny objects, such as orange tree

leaves, transparency film, and a stainless steel plate, the measurement distance was shorter

than actual distance. At the measurement angle of 90◦, the difference between the longest

and shortest measured distance (90% of the maximum scanner measurement distance) was

21.3 mm for the LMS200 and 29.7 mm for the URG-04LX. At the measurement angle of 45◦,

this difference increased to 73.2 mm for the LMS200; the URG-04LX was not able to detect

any objects at 45◦.

The surface shapes of a cylindrical pipe, a folded cardboard plate with a square-shaped

valley, and a folded cardboard plate with a V-shaped valley were well-depicted by the laser

scanner. For the object with a V-shaped valley with a true depth of 6.1 cm, the averaged

depths measured by the LMS200 and URG-04LX were 6.8 cm and 3.6 cm, respectively. The

larger discrepancy in the URG-04LX depth measurement may be caused by the relatively

lower angular resolution of the URG scanner, compared to that of the LMS scanner.

© 2007 Elsevier B.V. All rights reserved.

∗ Corresponding author. Tel.: +1 863 956 1151x1228; fax: +1 863 956 4631.E-mail address: [email protected] (R. Ehsani).

0168-1699/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.compag.2007.08.007

mailto:[email protected]

dx.doi.org/10.1016/j.compag.2007.08.007

a g r

1

Ftfpmscpim

stliarehsacf

rdbflabtobsrm

aeJsisatocnspscbsttt

Their specifications, provided from the manufacturers, areshown in Table 1. The LMS200 has a longer measurement dis-tance, larger size, and is heavier compared to the URG-04LX.

c o m p u t e r s a n d e l e c t r o n i c s i n

. Introduction

or agricultural and industrial applications, the distance to aarget object is a valuable measurement because it can be usedor determining a variety of other measurements. For exam-le, objects in an open space can be detected and counted byeasuring the distances to the objects. Even the positions and

hapes of the objects can be obtained. Distance measurementan be useful information in generating the surface topogra-hy of a target object, like for example, fruit trees. The 3D

mage of an object can be also reconstructed using distanceeasurements, obtained by moving a sensor in a 2D plane.With advances in sensing technology, various types of sen-

ors for distance measurement have been developed. Amonghem, sensors which use lasers have shown dominance. Aaser is a light source device which converts external energynto electromagnetic radiation. The word LASER came fromn acronym for light amplification by stimulated emission ofadiation, but it is now accepted as a single word. What differ-ntiates a laser from other light sources is that a laser beamas a single wavelength, a same phase, and high energy den-ity. Thus, a laser beam can travel to quite a long distance instraight line, maintaining a narrow beam. Because of this

haracteristic, lasers are commonly used as a sensing sourceor distance measurement.

A laser scanner, which is also called a laser radar or a laserange finder, is a non-contact optical device that measures theistance to an object in a scanning field using a pulsed laseream. The scanner’s measurement is based on the time-of-ight (TOF) principle. A laser source inside the scanner emitspulsed laser beam. If this beam hits an object, part of the

eam is reflected back to the scanner and hits a detector insidehe scanner. The time between transmission and receptionf the pulsed signal is directly proportional to the distanceetween the scanner and the object. The laser pulse is divertedequentially with a specific angular interval using an internalotating mirror. Thus, a fan-shaped two-dimensional scan is

ade of the surrounding area.Laser scanners are becoming a common sensing device to

id the steering device to avoid obstacles, and in mappingnvironments for use in robotics and agricultural applications.imenez et al. (1999) built a laser scanner-based measurementystem to recognize fruits in field tree conditions, consider-ng it as a sensing device for a fruit-harvesting robot. Thecanner provided the distance to an object and the attenu-tion of laser signal which occurred in the round-trip travelo the object. The information obtained was merged to rec-gnize the fruit and find the final fruit position. Hebert (2000)ompared the characteristics of several range sensing tech-ologies used in robotics. The measurement range of a lasercanner using the TOF principle was relatively long, com-ared to other technologies. The scanner provided relativelytable, accurate measurements under hostile environmentalonditions such as fog, dust, or smoke. Monta et al. (2004)uilt a three-dimensional sensing system, composed of a laser
canner and a scanner table moving vertically, for an agricul-ural robot. The sensing system could detect objects such asree trunks, branches, and leaves in a vineyard, and calculatehe diameter of the tree trunk and the distance between the
i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262 251

tree trunks. Kise et al. (2005) presented an obstacle detectionand identification algorithm of a laser scanner-based sens-ing system for autonomous agricultural vehicles. The sensingsystem was capable of detecting a moving object within asemicircle of an 8 m radius and reconstructing a 2D silhou-ette of the obstacle progressively in real time. Subramanian etal. (2006) developed machine vision and laser scanner-basedguidance systems to navigate a tractor through the alleywayof a citrus grove and compared the performance of thesesystems. They reported that the laser scanner-based guid-ance was the better guidance system for straight and curvedpaths.

Measuring the surface topography of soil and plants orknowing the shape of an object is important for many pre-cision agriculture applications. For this, laser scanners haveshown great potential. Darboux and Huang (2003) developeda laser scanning system composed of two diode lasers anda digital camera to measure soil surface microtopography.Gonzalez et al. (2007) demonstrated the capability of a 3Dlaser scanning system, which consisted of a laser transmitterand two cameras, in describing the evolution of an under-water sediment bed in real time. Ehsani and Lang (2002)developed a laser scanner-based plant volume measurementsystem. The system was able to measure plant volume andheight, indicating the possibility to measure the biomassand leaf area index of the plant. Wei and Salyani (2004,2005) showed the potential of a laser scanner for simulta-neous measurement of tree canopy height, width, volume,and foliage density. While applications of laser scanners inautomation, robotics, and agriculture are increased signifi-cantly in the recent years, very little published informationis available on characteristics and performance of these laserscanners.

2. Objectives

The overall goal of this study was to analyze and compare thecharacteristics of two commercially available laser scanners.The specific objectives were: (i) to test distance-measurementdrifts over time at different measurement distances andangles, (ii) to examine the effect of different materials andcolors of target objects on distance measurements, and (iii)to measure accuracy in mapping the surface patterns of theobjects of different shapes.

3. Materials and methods

Two laser scanners, LMS200 (Sick Ag, Germany) and URG-04LX(Hokuyo Automatic Co., Japan), were used for the tests (Fig. 1a).

It also has the ability to change angular resolution. To controlthe scanners and download the measurements to a computer,a computer program was written using LabVIEW (NationalInstruments Co., Austin, TX).

252 c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262

rs (a
Fig. 1 – Laser scanne
3.1. Laser scanners

3.1.1. LMS200The light source of the LMS200 is a pulsed infrared laser of905 nm, not visible to the human eyes. It operates in either mmmode or cm mode. The maximum measurement distances are8 m in the mm mode and 80 m in the cm mode. The LMS200has two scanning range options: (i) from 40◦ to 140◦ with angu-lar resolutions of 0.25◦, 0.5◦, and 1◦, and (ii) from 0◦ to 180◦

with angular resolutions of 0.5◦ and 1◦. The times for scan-ning one cycle are 53.28 ms, 26.64 ms, and 13.32 ms at 0.25◦,0.5◦, and 1◦ angular resolutions, respectively, in both scan-ning ranges. The scanner requires about 13.32 ms for one cyclerotation of an internal mirror with a 1◦ step. To achieve 0.25◦

and 0.5◦ angular resolutions, the 1◦ step is shifted to 0.25◦ and0.5◦ at the start of the mirror wheel rotation, respectively, and
four and two mirror rotations are required. For this reason, ascan with an angular resolution of 0.5◦ takes as twice as longas a scan with an angular resolution of 1◦; and a scan withan angular resolution of 0.25◦ takes four times as long. The
Table 1 – Specifications of the LMS200 and URG-04LX

LMS200

Maximum measurement distance (m) 8 (mm mode), 80 (cm modScanning angle (◦) 180 and 100 (selectable)Angular resolution (◦) 0.25, 0.5, and 1 (selectableScanning time (ms/cycle) 53, 26, and 13 at 0.25◦, 0.5◦

resolution, respectivelyMeasurement resolution (mm) 10Measurement error (mm) ±20 (mm mode), ±40 (cm

Data interface and transfer rate RS232 and RS422 (9.6 kbit/38.4 kbit/s, and 500 kbit/s)

Supply voltage (VDC) 24 ± 15%Current consumption (mA) 830Weight (kg) 4.5External dimensions (mm) 185 (W) × 156 (L) × 210 (H)

) and steel stand (b).

laser scanner can communicate with a computer via a serialport at a baud rate of 9.6 kbit/s, 19.2 kbit/s, and 38.4 kbit/s. Forhigh-speed data transmission, the manufacturer offers specialhigh-speed data interface cards. With the cards, the baud ratecan increase up to 500 kbit/s. Our test results indicated thatsome of the collected data were occasionally lost when com-munication took place at 500 kbit/s via a PCMCIA high-speedRS422 interface card (CSM GmbH, Germany), while no datawas lost at the slower speed of 38.4 kbit/s. Thus, for more reli-able communication between the scanner and the computer,a 38.4 kbit/s data transfer rate was chosen for the experiment.The LMS200 ran in the mm mode and scanned target objectsin the range of 40–140◦ with an angular resolution of 0.25◦.

3.1.2. URG-04LXThe URG-04LX uses a semiconductor laser beam of 785 nm to
measure distance. It has a fixed scanning range of 60–300◦ witha 0.36◦ angular resolution. Its maximum measurement dis-tance is 4 m when an object is white paper. The scanner andexternal devices can interface with each other via a RS232 port
URG-04LX

e) 4240

) 0.36, and 1◦ angular 100

1mode) ±10: up to 1 m distance; 1% of distance:

1–4 m distances, 19.2 kbit/s, RS232 (19.2 kbit/s, 57.6 kbit/s, 115.2 kbit/s,

500 kbit/s, and 750 kbit/s), USB 2.0(12 Mbit/s)5 ± 5%5000.1650 (W) × 50 (L) × 70 (H)

a g r

wrt

3

3Eln(wbsiFa

1(aUstnpstopnan

do9ml

Fm

c o m p u t e r s a n d e l e c t r o n i c s i n

ith a baud rate of 19.2–750 kbit/s and a USB port with a baudate of 12 Mbit/s. In this study, the scanner communicated withhe computer via a USB port.

.2. Experimental methods

.2.1. Drift testxperiments were designed to test measurement drifts of theaser scanners over a given period of time. The laser scan-ers were placed on a stand at a height of 96 cm. A steel stand

Fig. 1b) was made to place a target object at a certain distanceith a constant height. The stand was designed with the capa-ility of attaching a target object to the backside of the openingquare on the stand. A laser beam passes through the open-ng (25 cm × 25 cm) and hits the surface of the target object.or the tests, a sheet of white paper (30 cm × 30 cm) was useds the target object.

The stands were placed at angles (�) of 45◦, 90◦, and35◦, facing the scanners (Fig. 2a). The intended distancesD) between the object and the scanners were 0.8 m, 4.0 m,nd 7.2 m for the LMS200, and 0.4 m, 2.0 m, and 3.6 m for theRG-04LX, which are 10%, 50%, and 90% of the maximum mea-urement distances of the scanners, respectively. It was hardo measure the true distance between the object and the scan-ers because the laser beam detector, which is the referenceoint on the scanner for distance measurement, is inside theealed case of the scanners. Therefore, to keep intended dis-ances constant for each test, a new reference point was maden the outside of the scanner housing. A point at which aower plug and a communication plug meet was set as theew reference point on the LMS200 (Fig. 1a). On the URG-04LX,marker indicating the front of the scanner was used as theew reference point (Fig. 1a).

At 50% of the maximum scanner measurement distances,istance data were collected at 45◦, 90◦, and 135◦ simultane-
usly every second for 4 h. This test was repeated at 10% and0% of the maximum measurement distances. These experi-ents were conducted in a room illuminated with fluorescent
amps. The light intensity of the room was between 650 lx

ig. 2 – Schematic of the experimental setup (top view) (a) for driaterials and colors.

i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262 253

and 700 lx, and the temperature was between 12 ◦C and 20 ◦C.Before the tests, the power of the scanners was shut off forabout 10 h.

3.2.2. Test on objects of different materials and colorsFor examining the effect of different materials and col-ors of objects on distance measurement, 11 target objects(30 cm × 30 cm) were used. The materials included five sheetsof colored paper with a 0.5 mm thickness (white, blue, yellow,red, and black), a 20-mm thick laminated wood plate, a 2-mmthick stainless steel plate, a 26-mm thick styrofoam plate, a 25-mm thick sheet of sponge, a 4-mil thick sheet of transparencyfilm, and a sheet made of orange tree leaves that were affixedon a transparency film using double-sided tape without anyfree space.

The three intended distances from the scanner to theobject used in the experiment for drift measurement werealso used in this experiment. The target objects were placed atangles (�) of 45◦ and 90◦ with same distance (D) to the scanner(Fig. 2b). The objects were turned towards the front so that theeffect of an incidence angle from the laser beam to the surfaceof the object on distance measurement could also be studied.The distance data to the 11 objects were collected first at 90◦

for 90% of the maximum scanner measurement distances. Theexperimental order of the objects was randomly determined,and then the test was conducted again at 45◦ with the samedistance in the new random experimental order of the objects.These tests were repeated at 10% and 50% of the maximumscanner measurement distances.

In order to analyze the data, a multiple comparison analysiswas conducted using the “multicompare” function in MAT-LAB’s statistics toolbox (The MathWorks Inc., Natick, MA, Ver5.1). The multicompare function follows Tukey’s procedure,which is based on the Studentized range distribution.

3.2.3. Determining the laser beam spot sizeImmediately after the test on the objects of different materialsand colors at a certain angle and distance, another experimentwas conducted to investigate the size of the laser beam spot

ft measurement and (b) for testing objects of different


l pipe, (b) folded cardboard plate with square-shaped valleys,ces in cm.

Fig. 3 – Objects for generating surface patterns: (a) cylindricaand (c) folded cardboard plate with V-shaped valleys. Distan

at the same angle and distance. The square stand openingwas blocked using two sets of two sheets of white paper. Bymoving one set of two sheets of paper horizontally from thecenter to the left-hand side and right-hand side, respectively,and the other set of two sheets of papers vertically from thecenter to the top and bottom, respectively, a small rectangularopen area was made at the center of the opening. When thelaser beam, generated at a single angle, passed through therectangular opening, the papers were fixed. The rectangularopening was considered as the approximate size of the laserbeam spot at the angle and distance.

3.2.4. Generating surface patterns of objects of differentshapesExperiments for investigating the capability of the scanners ingenerating the surface patterns of objects of different shapeswere designed. For the tests, three objects shown in Fig. 3 wereprepared: (i) a cylindrical pipe, (ii) a folded cardboard platewith square-shaped valleys, and (iii) a folded cardboard platewith V-shaped valleys. The size of the cardboard plates was91 cm (W) × 15 cm (H). These objects were put on a table withadjustable height. The table was fixed at a height at which thelaser beam hit the mid-height of the objects. The objects werealso positioned where the laser beam, generated at 90◦, hitthe center of the objects. The intended distance between thescanners and the objects was 100 cm.

To build the surface patterns of the objects, the paralleldistance (P) between the scanner and the object was calcu-lated from the measured distance (M) and the measurementangle (�) using the definition of the sine trigonometric function(Fig. 4):

P = M sin � (1)

The diameter of the cylindrical pipe and the width (D) ofthe hill in the square-shaped object were obtained based onthe distances (R1 and R2) and measurement angles (�1 and �2),
which were measurements at the right-most and left-mostedges of the objects, respectively (Fig. 5):
D =√

R21 + R2

2 − 2R1R2 cos(�2 − �1) (2)

Fig. 4 – Geometry for obtaining the parallel distancebetween the laser scanner and an object.

4. Results and discussion

4.1. Measurement drift

Distance measurements by the LMS200 over time at anintended distance of 4.0 m and three different angles (45◦, 90◦,and 135◦) are shown in Fig. 6. The distance data fluctuatedwith a peak-to-peak value of about 20 mm. The period of thefluctuation at about 2 min of run time was in the range of0.3–0.5 min. The period increased with run time, and was inthe range of 15–20 min when the run time reached 200 min.To examine the trend of measurement drift, the distance datawere averaged every 20 min. This time interval was selectedbecause the longest period of the fluctuation was close to thetime interval. The averaged distance data decreased with runtime until about 53 min, and then stayed at a constant level. In

◦ ◦
this stable region, the averaged distance data at 45 , 90 , and135◦ differed a little from each other. This might be caused bythe difference of the distance between the scanner and theobject at each angle.

c o m p u t e r s a n d e l e c t r o n i c s i n a g r

Fig. 5 – Geometry for obtaining the width of an object.

Fig. 6 – Distance measurements by the LMS200 at an intended d

i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262 255

The differences between the averaged distance data at2 min and that at 53 min were 7.8 mm, 7.7 mm, and 5.7 mmat 45◦, 90◦, and 135◦, respectively. These were considered asthe measurement errors caused only by insufficient warm-uptime of the laser scanner. The effect of measurement angle onmeasurement drift was trivial. The pattern of the averaged dis-tance data over time at an intended distance of 4.0 m (Fig. 6)was also observed in the data measured at 0.8 m and 7.2 m.The settling time, which is the time required for the averageddistance data to reach a stable level, was different dependingon measurement distances.

Fig. 7 presents distance measurements by the LMS200 atan angle of 135◦ for three intended distances (0.8 m, 4.0 m, and7.2 m). Like the data measured at a constant distance for differ-ent angles (Fig. 6), the distance measurements shown in Fig. 7fluctuated. Again, the averaged distance data decreased withrun time at the beginning of the scanner’s operation, and thenbegan to stabilize at a settling time. The averaged distancesin the stable regions deviated from the intended distances.This may have been caused by disagreement between the
location of the new reference point on the scanners and thelocation of the laser beam detector inside the sealed case. Inaddition, these could have been inaccurate distance measure-ment using a tape measurement when the scanner and the
istance of 4.0 m for the angles: (a) 45◦, (b) 90◦, and (c) 135◦.


for th
Fig. 7 – Distance measurements by the LMS200 at 135◦
object were setup before the test. The settling times at 0.8 m,4.0 m, and 7.2 m distances were 32 min, 53 min, and 137 min,respectively, showing the settling time increases when themeasurement distance increases. The differences betweenthe averaged distance at 2 min of run time and that at thesettling times were 2.4 mm, 5.7 mm, and 14.1 mm at 0.8 m,4.0 m, and 7.2 m, respectively. This shows a measurementerror caused by insufficient warm-up time of the laser scannerincreases when measurement distance increases.

From this experiment, it was shown that in order to mea-sure the distance to an object accurately, the LMS200 needsto warm up for some time before measurement. The requiredwarm-up time differs depending on measurement distanceand the data measured should be averaged over a specific timeinterval.

Distance measurements by the URG-04LX at a single angleof 135◦ for three intended distances of 0.4 m, 2.0 m, and 3.6 m
are shown in Fig. 8. Like the distance measurements by theLMS200 (Fig. 7), the distance data shown in Fig. 8 also fluc-tuated with a peak-to-peak value of about 10–15 mm. Theamplitude of the fluctuation tended to be larger when the
e intended distances: (a) 7.2 m, (b) 4.0 m, and (c) 0.8 m.

measurement distance increased, but the period of the fluc-tuation was not recognizable. The distance data, averagedevery 20 min, presented a different pattern with that of theaveraged distance data measured by the LMS200 (Fig. 7). Theaveraged distances increased in the beginning of the scan-ner’s operation at the intended distances of 0.4 m and 2.0 m,which had settling times of 50 min and 70 min, respectively.At an intended distance of 3.6 m, the averaged distance beganto decrease, reached a bottom limit at 70 min, increaseduntil 111 min, and then stabilized. The averaged distancesat the settling times were quite different with the intendeddistances. This might also have been caused by inaccuratedistance setup between the scanner and the object beforethe test. The settling time was larger when the measurementdistance increased. The differences between the averaged dis-tance at 2 min of run time and that at the settling times were9.3 mm, 12.3 mm, and 4.1 mm at intended distances of 0.4 m,
2.0 m, and 3.6 m, respectively. Thus, like the LMS200, the URG-04LX should also be warmed up for the settling time before thetest to minimize measurement error, and the distance datashould be averaged. In the test with the URG-04LX at a single

c o m p u t e r s a n d e l e c t r o n i c s i n a g r i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262 257

for

dw

4o

TwpsT1dsslwd

oi

Fig. 8 – Distance measurements by the URG-04LX at 135◦

istance for three different angles, the distance measurementas not affected by the angles.

.2. Effect of different materials and colors of objectsn distance measurements

he previous experimental results demonstrated that somearm-up settling time is required for the laser scanners torovide stable distance measurements (Figs. 6–8). The highestettling time for the LMS at a distance of 7.2 m was 137 min.he longest period of the fluctuating distance data was about5–20 min in the stable region. Thus, to avoid error in theistance measurement by insufficient warm-up time of thecanners and a short data-sampling period, the tests weretarted after running the scanners for 3 h without data col-ection. The scanners recorded 1000 readings on each object,

hich roughly corresponded to one period of the fluctuating
istance data in the stable region.
Tables 2 and 3 show the mean and standard deviationf distance measurements to each object by the LMS200 at

ntended distances of 0.8 m and 7.2 m, respectively, and the

the intended distances: (a) 3.6 m, (b) 2.0 m, and (c) 0.4 m.

results of multiple comparison analysis at the distances. Themean distances of the objects were sorted in an ascendingorder of alphabet indexes in the multiple comparison col-umn. When the objects have the same index, there is nosignificant difference among them in a 95% confidence level.In general, the shortest distance measurements were foundwith shiny objects such as orange tree leaves, transparencyfilm, and a stainless steel plate. The longest distance mea-surements were found with objects made of soft materialssuch as styrofoam and dry sponge. In particular, the distancemeasurement to the transparent film was very sensitive tomeasurement angles. At 45◦, the transparent film could not bedetected by the LMS200. Some portion of the laser beam mightpenetrate the film, and a large portion of the beam bouncedoff the film might be deviated from the route to the scanner.Thus, the amount of the laser beam returning to the scannermay have been insufficient for the scanner to detect the tar-
get object. When a laser beam is directed towards an objectof high reflectivity such as a stainless steel plate, most of thebeam is bounced off the object immediately after hitting itand comes back to the scanner. However, when the object has


Table 2 – Mean, standard deviation, and multiple comparison analysis of distance measurements to the different objectsby the LMS200 at an intended distance of 0.8 m

Material 45◦ 90◦

Mean (mm) S.D. (mm) Multiple comparison* Mean (mm) S.D. (mm) Multiple comparison*

Tree leaves 788.4 3.8 a 788.2 3.3 dBlue paper 789.4 3.8 b 787.5 4.7 cdBlack paper 790.2 1.7 c 786.3 5.0 cYellow paper 790.4 1.6 c 785.3 2.0 bRed paper 791.8 5.0 d 787.2 4.8 cdWood plate 792.1 2.5 d 788.2 3.5 dWhite paper 793.8 1.7 e 788.3 1.9 dSteel plate 799.9 3.9 f 772.1 1.6 aStyrofoam 802.3 1.6 g 796.7 1.9 fSponge 811.1 2.5 h 802.5 2.0 gTransparency – – – 794.4 9.8 e

analyes wa

∗ “a–h” represent the alphabet indexes of the multiple comparisondifference. The confidence level of the multiple comparison analys

low reflectivity such as sponge, a large portion of the laserbeam is absorbed on the surface of the object for some time,and then returns to the scanner. Thus, the flight time of thelaser beam to an object of low reflectivity is longer than one ofhigh reflectivity, and the distance measurement to the objectof low reflectivity is larger. At an intended distance of 0.8 m(Table 2), the differences between the longest distance mea-sured on the soft objects and the shortest distance measuredon the shiny objects were 22.7 mm and 30.4 mm at 45◦ and90◦, respectively; at an intended distance of 7.2 m (Table 3),the differences at 45◦ and 90◦ were 73.2 mm and 21.6 mm,respectively.

The effect of color on distance measurement was moresensitive at 45◦ than at 90◦. At an intended distance of 0.8 m,the differences between the longest and shortest distanceson the colored papers at 45◦ and 90◦ were 4.4 mm and 3.0 mm,respectively; at an intended distance of 7.2 m, the differences
at 45◦ and 90◦ were 5.0 mm and 2.0 mm, respectively. Theresults showed that the effect of the measurement distanceson distance measurement of the different colored objects wasminor.
Table 3 – Mean, standard deviation, and multiple comparison aby the LMS200 at an intended distance of 7.2 m

Material 45◦

Mean (mm) S.D. (mm) Multiple comparis

Tree leaves 7168.6 2.7 aSteel plate 7169.3 4.8 aWood plate 7173.1 4.5 bBlue paper 7175.1 4.0 cBlack paper 7175.2 4.9 cYellow paper 7176.3 3.8 dRed paper 7178.6 4.5 eWhite paper 7180.2 2.9 fStyrofoam 7185.2 6.7 gSponge 7241.8 5.3 hTransparency – – –

∗ “a–h” represent the alphabet indexes of the multiple comparison analydifference. The confidence level of the multiple comparison analyses wa

ses. When the objects have the same index, there is no significants 95%.

The mean and standard deviation of distance measure-ments to the different objects by the URG-04LX at intendeddistances of 0.4 m and 3.6 m and the results of multiple com-parison analysis are shown in Tables 4 and 5. As observed inTables 2 and 3, the shortest distance measurements were alsofound with shiny objects and the longest distance measure-ments were also found with soft objects. The URG-04LX couldnot detect transparency film at 45◦ for an intended distanceof 0.4 m. The capability of the URG-04LX was not enough todetect any objects at 45◦ for an intended distance of 3.6 m.At an intended distance of 0.4 m (Table 4), the differencesbetween the longest distance observed at the soft objects andthe shortest distance at the shiny objects at 45◦ and 90◦ were55.4 mm and 21.6 mm, respectively. At an intended distance of3.6 m (Table 5), the differences at 90◦ were 29.7 mm. The mea-surement angles at an intended distance of 0.4 m might notinfluence on the results of the multiple comparison analysis
for the different colored papers. However, at a measurementangle of 90◦, the result of multiple comparison analysis for thecolored papers at an intended distance of 0.4 m might be moresensitive than that at an intended distance of 3.6 m.
nalysis of distance measurements to the different objects

90◦

on* Mean (mm) S.D. (mm) Multiple comparison*

7188.0 5.0 cd7178.2 4.8 b7192.0 2.1 f7189.4 1.5 e7187.9 2.4 cd7187.4 2.4 c7188.4 2.6 cde7188.8 2.4 de7197.7 3.3 g7197.4 2.3 g7176.1 3.0 a



Table 4 – Mean, standard deviation, and multiple comparison analysis of distance measurements to the different objectsby the URG-04LX at an intended distance of 0.4 m

Material 45◦ 90◦

Mean (mm) S.D. (mm) Multiple comparison* Mean (mm) S.D. (mm) Multiple comparison*

Steel plate 360.2 4.1 a 442.6 2.5 iBlack paper 397.9 2.2 b 424.0 1.8 dTree leaves 398.8 2.2 c 424.8 1.7 eWood plate 399.0 2.2 c 421.9 1.8 bBlue paper 399.4 2.2 c 424.3 2.0 dYellow paper 400.5 2.2 d 427.2 1.8 fRed paper 402.7 2.2 e 422.9 1.8 cWhite paper 405.7 2.4 f 425.0 1.6 eSponge 412.6 2.0 g 428.7 1.7 gStyrofoam 415.6 2.0 h 437.5 1.8 hTransparency – – – 415.9 2.8 a

∗ “a–h” represent the alphabet indexes of the multiple comparison analydifference. The confidence level of the multiple comparison analyses wa

Table 5 – Mean, standard deviation, and multiplecomparison analysis of distance measurements to thedifferent objects by the URG-04LX at an intendeddistance of 3.6 m

Material 90◦

Mean (mm) S.D. (mm) Multiplecomparison*

Steel plate 3643.3 3.3 aBlue paper 3647.6 2.9 bRed paper 3648.3 2.9 cYellow paper 3648.6 3.0 cdBlack paper 3649.1 2.9 cdeWhite paper 3649.5 2.9 cdeTree leaves 3649.9 3.4 eWood plate 3652.3 2.9 fStyrofoam 3658.1 2.8 gTransparency 3666.6 2.8 hSponge 3673.0 2.9 i

∗ “a–i” represent the alphabet indexes of the multiple compari-son analyses. When the objects have the same index, there isno significant difference. The confidence level of the multiplecomparison analyses was 95%.

Table 6 – Sizes of the laser beam spot, approximated with a recdistances

Distance (m) Rectangle

45◦

0.8 for LMS200, 0.4for URG-04LX

Width (cm) 2.0Height (cm) 1.8






4.3. Determining the laser beam spot size

Table 6 shows the sizes of the laser beam spot, approximatedwith a rectangle, for the LMS200 and URG-04LX at differentangles and distances. In the spot for the LMS200, the rect-angle corresponded to a single laser beam spot. However,in the spot for the URG-04LX, it corresponded to three laserbeam spots generated at consecutive three angles. The laserbeam could not be separated to a single beam with physicalmethods.

The widths of the rectangles at 45◦ were greater thanthose at 90◦. The ratios of the widths at 45◦ to the widthsat 90◦ were in the range of 1.63–1.73 for the LMS200 and1.25–1.29 for the URG-04LX, but the heights of the rectanglesat both angles were almost the same. Both the width andthe height of the LMS beam spot increased when measure-ment distance increased. In the URG beam spot, the widthwas greater with increased distance, but the change in heightwas minor. The linear regression analysis between the area(Y) of the laser beam spot and measurement distance (X) wasconducted (Eqs. (3)–(6)). The beam spot area increased linearlywith distance. The linear models for the LMS beam at 45◦

and 90◦ are:

Y = 6.70X − 3.52 (R2 = 0.98) (3)

Y = 3.86X − 1.63 (R2 = 0.99) (4)

tangle, for the LMS200 and URG-04LX at three intended

LMS200 URG-04LX

90◦ 45◦ 90◦

1.2 1.5 1.21.8 1.0 0.9

3.2 4.0 3.13.9 0.8 0.9

4.8 – 4.85.6 – 1.2

i n a g
260 c o m p u t e r s a n d e l e c t r o n i c s
The linear models for the URG beam at 45◦ and 90◦ are:

Y = 1.46X + 0.29 (R2 = 0.98) (5)

Y = 1.06X + 1.08 (R2 = 1.00) (6)

4.4. Surface pattern of objects

Fig. 9 shows the surface pattern of a cylindrical pipe gen-erated by the laser scanners. The pattern provided a sketchof the surface shape of the pipe, but the resolution was notenough to depict the pipe surface in detail. In the surface pat-tern generated by the LMS200, the radii of the pipe obtainedby calculating the difference between the highest distanceat the edge and the shortest distance in the middle, andobtained using the cosine law (Eq. (2)) were 3.5 cm and 3.7 cm,respectively. The radii were 1.4 cm and 2.8 cm, respectively,in the surface pattern generated by the URG-04LX. The radiiobtained in the pattern by the LMS were close to the trueradius of 3.5 cm, while the radii obtained in the pattern by theURG were underestimated. The URG scanner has the lowerangular resolution of 0.36◦, compared to the LMS angular res-olution of 0.25◦. Thus, the URG scanner might miss the right-and left-most edges of the pipe, and this may explain thereason why the measured radii are smaller than the trueradius.

The surface pattern of an object with flat hills and square-shaped valleys generated by the laser scanners are shown inFig. 10. The pattern clearly reconstructed the hills and valleys.The numbers of hills and valleys in the pattern agreed withthose in the actual object. In the pattern by the LMS200, the
widths of the hills and valleys in the middle, left-most, andright-most were obtained using the cosine law. The widthsof the hills were 6.9 cm, 6.1 cm, and 6.1 cm, respectively; andthe widths of the valleys were 6.1 cm, 4.0 cm, and 4.4 cm,
Fig. 9 – Surface pattern of a cylindrical pipe gen

r i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262

respectively. The depths of the valleys were 7.2 cm, 7.5 cm,and 7.8 cm, respectively. In the pattern by the URG-04LX, thewidths of the hills were 6.8 cm, 5.0 cm, and 5.6 cm, respec-tively; and the widths of the valleys were 6.0 cm, 4.6 cm, and4.6 cm, respectively. The depths of the valleys were 7.2 cm,6.6 cm, and 7.0 cm, respectively.

The measured widths of the hills in the middle were closeto the true width of 7.0 cm. The widths of the hills at the leftand right sides were smaller than the true width. Since thelaser beam generated at 90◦ hit the middle of the object, andthe scanners were in parallel with the object, the resolutionof the surface pattern at the left and right sides became lowerthan that in the middle. This might cause the smaller widthsat the left and right sides. The measured widths of the valleyswere much smaller than the true width of 7.0 cm. When thelaser beam was projected on the area of the valley, some ofthe beam was blocked by the hill before it reached the valley.Therefore, the measured widths of the valleys became smaller.The measured depths of the valleys were close to the truedepth of 7.0 cm.

Fig. 11 shows the surface pattern of an object with V-shapedvalleys generated by the laser scanners. The pattern recon-structed the surface shape of the object well. The numbersof hills and valleys in the pattern agreed with those in theactual object. The averaged depths of the valleys in the pat-terns by the LMS and URG were 6.8 cm and 3.6 cm, respectively.The depth measured by the URG was much smaller than thetrue depth of 6.1 cm. Since the URG scanner has a relativelylarger angular interval (0.36◦), compared to the angular inter-val (0.25◦) of the LMS, it might miss the crests of the hills andthe bottom limits of the valleys. This might cause to under-
estimate the depth of the valley. This also can be confirmedfrom the fact that the crests of the hills and the bottom limitsof the valleys in the pattern by the LMS are sharper than thosein the pattern of the URG.
erated by the LMS200 and the URG-04LX.


Fig. 10 – Surface pattern of an object with flat hills and square-shaped valleys generated by the LMS200 and the URG-04LX.

d va

5

Tntf

Fig. 11 – Surface pattern of an object with V-shape

. Conclusions

he characteristics of two commercially available laser scan-ers, LMS200 and URG-04LX, were analyzed and comparedhrough several tests. The following conclusions can be drawnrom these tests:

lleys generated by the LMS200 and the URG-04LX.

• Distance measurements by the laser scanners over run timefluctuated with a peak-to-peak value of 10–20 mm. The dis-tance data measured by the LMS200 showed a decreasing
pattern until a settling time, whereas that by the URG-04LXshowed an increasing pattern.
• The warm-up settling time was greater at a longer mea-surement distance, but was not affected by measurement

i n a g

r

262 c o m p u t e r s a n d e l e c t r o n i c s

angle. At a measurement angle of 135◦ for 10%, 50%, and 90%of the maximum measurement distances of the scanners,the setting times of the LMS200 were 32 min, 53 min, and137 min, respectively; those of the URG-04LX were 50 min,70 min, and 111 min, respectively.

• From distance measurements to objects of different materi-als and colors, the longest measurements were found withsoft objects such as styrofoam and sponge; the shortest oneswere found with shiny objects such as orange tree leaves,transparency film, and a stainless steel plate. The effect ofcolor on distance measurement was more sensitive at 45◦

than at 90◦; but the effect of measurement distances ondistance measurement of the different colored objects wasminor.

• At 90% of the maximum scanner measurement distances,the differences between the longest measurement with thesoft objects and shortest measurements with the shinyobjects were 73.2 mm and 21.3 mm at 45◦ and 90◦ for theLMS200, respectively, and 29.7 mm at 90◦ for the URG-04LX.The capability of the URG-04LX was not enough to detectany objects at 45◦. The transparency film could not bedetected by either laser scanner at 45◦ for 10% and 50% ofthe maximum scanner measurement distances.

• The size of the laser beam spot was approximated witha rectangle. Both the width and height of the LMS beamspot increased when measurement distance increased.Regarding the URG beam spot, the width was greater withincreased distance, but the change of the height was minor.The beam spot areas of both the scanners increased linearlywith distance.

• The surface patterns of different shapes of objects mappedand reconstructed by the laser scanners depicted the sur-
face of the target objects well. From the surface patternof an object with V-shaped valleys, the averaged depths ofthe valleys in the patterns generated by the LMS and URGwere 6.8 cm and 3.6 cm, respectively. The depth measured
r i c u l t u r e 6 0 ( 2 0 0 8 ) 250–262

by the URG was much smaller than the true depth of 6.1 cm.This may have been caused by the relatively larger angularinterval of the URG scanner.

e f e r e n c e s

Darboux, F., Huang, C., 2003. An instantaneous-profile laserscanner to measure soil surface microtopography. Soil. Sci.Soc. Am. J. 67, 92–99.

Ehsani, M.R., Lang, L., 2002. A sensor for rapid estimation of plantbiomass. In: Proceedings of the 6th International Conferenceon Precision Agriculture, Bloomington, MN.

Gonzalez, E.P., Diaz-Pache, F.S.-T., Mosquera, L.P., Agudo, J.P., 2007.Bidimensional measurement of an underwater sedimentsurface using a 3D-scanner. Opt. Laser Technol. 39,481–489.

Hebert, M., 2000. Active and passive range sensing for robotics.In: Proceedings of the 2000 IEEE International Conference onRobotics & Automation, San Francisco, CA,pp. 102–110.

Jimenez, A.R., Jain, A.K., Ceres, R., Pons, J.L., 1999. Automatic fruitrecognition: a survey and new results using range/attenuationimages. Pattern Recogn. 32 (10), 1719–1736.

Kise, M., Zhang, Q., Noguchi, N., 2005. An obstacle identificationalgorithm for a laser range finder-based obstacle detector.Trans. ASAE 48 (3), 1269–1278.

Monta, M., Namba, K., Kondo, N., 2004. Three-dimensionalsensing system using laser scanner. ASABE Paper No. 041158.ASABE, St. Joseph, MI.

Subramanian, V., Burks, T.F., Arroyo, A.A., 2006. Development ofmachine vision and laser radar based autonomous vehicleguidance systems for citrus grove navigation. Comput.Electron. Agric. 53, 130–143.

Wei, J., Salyani, M., 2004. Development of a laser scanner for
measuring tree canopy characteristics: phase 1. Prototypedevelopment. Trans. ASAE 47 (6), 2101–2107.
Wei, J., Salyani, M., 2005. Development of a laser scanner formeasuring tree canopy characteristics: phase 2. Foliagedensity measurement. Trans. ASAE 48 (4), 1595–1601.

Documents

towards autonomous control of quadrotor unmanned aerial vehicles in gps denied urban areas via laser range finer