Study of bat flight behavior by combining thermalimage analysis with a LiDAR forest

reconstruction

Xiaoyuan Yang, Crystal Schaaf, Alan Strahler, Thomas Kunz, Nathan Fuller, Margrit Betke,Zheng Wu, Zhuosen Wang, Diane Theriault, Darius Culvenor, David Jupp,

Glenn Newnham, and Jenny Lovell

Abstract. The nature of forest structure plays an important role in the study of foraging behaviors of bats. In this study, we

demonstrate a new combined methodology that uses both thermal imaging technology and a ground-based LiDAR

system to record and reconstruct Eptesicus fuscus (big brown bats) flight trajectories in three-dimensional (3-D) space.

The combination of the two 3-D datasets provided a fine-scale reconstruction of the flight characteristics adjacent to and

within the forests. A 3-D forest reconstruction, assembled from nine Echidna Validation Instrument LiDAR scans over

the 1 ha site area, provided the essential environmental variables for the study of bat foraging behaviors, such as the

canopy height, terrain, location of the obstacles, and canopy openness at a bat roosting and maternity site in Petersham,

Massachusetts. Flight trajectories of 24 bats were recorded over the 25 m � 37.5 m region within the LiDAR forest

reconstruction area. The trajectories were reconstructed using imaging data from multiple FLIR ThermoVision SC8000

cameras and were co-registered to the 3-D forest reconstruction. Twenty-four of these flight trajectories were categorized

into four different behavior groups according to velocity and altitude analysis of the flight trajectories. Initial results

showed that although all bats were guided by echolocation and avoided hitting a tree that was in all of their flight paths,

different bats chose different flight routes. This study is an initial demonstration of the power of coupling thermal image

analysis and LiDAR forest reconstructions. Our goal was to break ground for future ecological studies, where more

extensive flight trajectories of bats can be coupled with the canopy reconstructions to better establish responses of bats to

different habitat characteristics and clutter, which includes both static (trees) and dynamic (other bats) obstacles.

Resume. La nature de la structure forestiere joue un role important dans l’etude des comportements d’approvisionnement

en nourriture des chauves-souris. Dans cette etude, on demontre une nouvelle methodologie combinee qui utilise a la fois

la technologie des images thermiques et un systeme LIDAR au sol pour enregistrer et reconstruire les trajectoires de vol

des Eptesicus fuscus (grandes chauves-souris brunes) dans l’espace tridimensionnel. La combinaison des deux ensembles

de donnees 3D a fourni une reconstruction a echelle fine des caracteristiques de vol adjacents et a l’interieur des forets.

Une reconstruction forestiere en 3D, assemblee a partir de neuf balayages LIDAR du capteur « Echidna Validation

Instrument » au-dessus du site d’etude de un hectare, a fourni les variables environnementales essentielles pour l’etude des

comportements d‘approvisionnement en nourriture des chauves-souris comme la hauteur du couvert, le terrain, la

localisation des obstacles et l’ouverture du couvert sur un site de nidification et de maternite de chauves-souris a

Petersham, au Massachusetts. Les trajectoires de vol de 24 chauves-souris ont ete enregistrees au-dessus de la region de

25 m � 37,5 m a l’interieur de la zone de reconstruction LIDAR de la foret. Les trajectoires ont ete reconstruites a l’aide

des donnees images de cameras FLIR ThermoVision SC8000 multiples et superposees sur la reconstruction

tridimensionnelle. Vingt-quatre de ces trajectoires de vol ont ete classifiees en quatre groupes differents de comportements

Received 9 January 2012. Accepted 21 July 2013. Published on the Web at http://pubs.casi.ca/journal/cjrs on 2 October 2013.

Xiaoyuan Yang1, Crystal Schaaf, and Zhuosen Wang. School for the Environment, College of Science and Mathematics, University ofMassachusetts Boston, Boston, MA, USA, 02125.

Alan Strahler. Department of Earth and Environment, Boston University, Boston, MA, USA, 02215.

Thomas Kunz and Nathan Fuller. Department of Biology, Boston University, Boston, MA, USA, 02215.

Margrit Betke, Zheng Wu, and Diane Theriault. Department of Computer Science, Boston University, Boston, MA, USA, 02215.

Darius Culvenor. Environmental Sensing Systems, Melbourne, Victoria 3000, Australia.

David Jupp. CSIRO Marine and Atmospheric Research, Canberra, ACT 2601, Australia.

Glenn Newnham. CSIRO Land and Water, Clayton South, Victoria 3169, Australia.

Jenny Lovell. CSIRO Marine and Atmospheric Research, Hobart, Tasmania 7000, Australia.

1Corresponding author (e-mail: Xiaoyuan.Yang@umb.edu).

Can. J. Remote Sensing, Vol. 39, No. S1, pp. S112�S125, 2013

S112 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

selon une analyse de la vitesse et de l’altitude des trajectoires de vol. Les resultats initiaux ont montre que, bien que toutes

les chauves-souris etaient guidees par echolocation et evitaient de frapper un arbre qui etait dans toutes leurs trajectoires

de vol, differentes chauves-souris choisissaient differentes trajectoires de vol. Cette etude constitue une demonstration

initiale du potentiel du couplage de l’analyse d’images thermiques et des reconstructions forestieres LIDAR. Notre

objectif etait de montrer la voie aux etudes ecologiques futures, ou des trajectoires de vol a plus grande echelle de chauves-

souris peuvent etre couplees avec les reconstructions du couvert pour mieux etablir les reponses des chauves-souris face

aux differentes caracteristiques des habitats et au fouillis d’echos, incluant a la fois les obstacles statiques (arbres) et

dynamiques (autres chauves-souris).

[Traduit par la Redaction.]

Introduction

The nature of forest structure (e.g., canopy cover,

distribution of overstory and understory, foliage density,

forest gaps, stand-scale variability, forest fragmentation,

availability of standing deadwood, and forestry practices)

are all important contributors to the maintenance of

successful bat populations (Lacki et al., 2007; Kunz et al.,

2008). Bat species have different foraging strategies: some

capture insects on the wing (aerial), whereas others collect

prey from surfaces (gleaning) (Fenton, 1990; Schnitzler and

Kalko, 1998). Previous research suggests that the vertical

stratification of a forest may also play an important role in

the foraging behavior of bats (Bonaccorso, 1979; Bernard,

2001; Kalko and Handley, 2001; Weinbeer and Kalko, 2004).

Forest attributes such as clumped leaf area with stand height

and patterns of disturbance and gap formation have been

shown to play an important role in determining optimum

bat habitat use (Parker and Brown, 2000). Jung et al. (2012)

found a positive association between European bat activities

and forest structure parameters, as well as associations

between similar adaptations in wing morphology and

foraging strategy with similar forest structural parameters.

Studies by Krusic et al. (1996) and Zimmerman and Glanz

(2000) showed that a mix of forest types, including clear-cuts

and regenerated areas and over-mature hardwoods, in

combination with water and trails, provide the optimum

summer habitat for the bat species of New England. For big

brown bats (Eptesicus fuscus) in New England, which were

targeted specifically in this study, the most common roosting

places are in hollow trees, beneath loose tree bark, in the

crevices of rocks, or in man-made structures such as attics,

barns, old buildings, eaves, and window shutters. These

roosting places vary from obstacle-rich space in a forest

interior to the less cluttered space near forest edges and in

forest gaps (Kurta and Baker, 1990). It remains unclear,

however, to what extent that this species uses open spaces

away from clutter, such as the space above the canopy and

the gaps within a canopy.

To detect, quantify, and understand bat activities inside a

forest is a challenging task. It requires appropriate integra-

tion of diverse tools, technologies, and multiple databases

(Kunz et al., 2008). Scientists have been exploring the way

bats fly at the edge of forests or above forest canopies using

radar (Kunz, 2004; Cleveland et al., 2006; Cryan et al.,

2008), radio-telemetry technologies (Hodgkison et al., 2003;

Thies and Kalko, 2004; Thies et al., 2006), ultrasonic sensors

(Holderied et al., 2005; Gillam et al., 2009; Holderied and

Jones, 2009; Parsons and Szewczak, 2009), and high-

resolution infrared cameras (Simmons, 2005; Betke et al.,

2007; Fleck et al., 2007; Betke et al., 2008; Hristov

et al., 2008; Wu et al., 2009a; Hristov et al., 2010; Theriault

et al., 2010). Without a way to describe the structure of the

forest quantitatively, using these methodologies alone can

only provide a limited assessment of foraging behavior.

Recent advances in remote sensing of the three-

dimensional (3-D) structure of forests have been made

with under-canopy Light Detection and Ranging (LiDAR)

systems (Strahler et al., 2008; Jupp et al., 2009b; Yang et al.,

2013), airborne sensors (Blair et al., 1999; Anderson

et al., 2006; 2008), and multi-angular passive data (Kimes

et al., 2006; Schull et al., 2007). The under-canopy LiDAR,

the Echidna Validation Instrument (EVI), is a near-infrared,

pulsed-laser, mirror scanning LiDAR with full-waveform

digitizing. The full-waveform recording enables character-

ization of the canopy structure of fully forested regions and

captures details about partially obscured objects that are not

possible with discrete return systems. Culvenor et al. (2005)

and Jupp et al. (2009b) described the theory and mechanical

details for the retrieval of Leaf Area Index (LAI) and the

foliage profile from EVI and demonstrated its application.

Strahler et al. (2007; 2008; 2009), Zhao et al. (2011), and

Yao et al. (2011) further validated retrieval of EVI-derived

forest structural parameter. Assembled from multiple EVI

scans, digital reconstructions can be built in 3-D space for

forests with varying density and species composition (Yang

et al., 2013). Three-dimensional forest reconstruction per-

mits direct characterization of the forest structure that

provided essential environmental variables for the detection

and analysis of bat foraging behaviors. The integration of

ground-based LiDAR data and high-resolution thermal

infrared cameras provided the opportunity to explore and

characterize the responses of bats to forest structure. This

integration may eventually help describe ecosystems on a

coarser scale and establish optimum landscape models for

bat assemblages in both temperate and tropical ecosystems

(Meyer et al., 2009; Bass et al., 2010).

In this study, the emergence patterns of a small colony of

big brown bats roosting in a barn located adjacent to the

edge of a deciduous forest (a 25 m � 37.5 m region of

interest) in Petersham, Mass., was investigated using

coupled thermal infrared imaging technology and forest

reconstruction using LiDAR. Our study focused on the big

brown bats, but similar studies may be very illuminating, for

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S113

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

example in tropical forests where different bat species use

different parts of a forest (Heithaus et al., 1975; Bonaccorso,

1979; Bernard, 2001; Kalko and Handley, 2001).

Methods

FLIR ThermalVision SC8000

Thermal infrared cameras detect the characteristic infra-

red radiation (IR) of objects in the form of photons. The

higher the temperature of an object of interest, the greater

the intensity of emitted radiation, and thus the brighter theresulting thermal image (Kastberger and Stachl, 2003).

Therefore, thermal infrared imaging has the ability to record

the behavior of free-ranging organisms in the dark and has

been proven to be a valuable paradigm for investigating bat

flight behavior (Betke et al., 2007; 2008; Hristov et al., 2008;

Wu et al., 2009a; 2009b; Hristov et al., 2010). In this study,

two FLIR ThermoVision SC8000 cameras were deployed

and positioned to reconstruct 3-D images of foraging bats asthey emerged from their daytime roost. The SC8000 camera

is a high-resolution and high-speed thermal imaging instru-

ment. It has a spatial resolution of 1024 � 1024 pixels and a

frame rate of Rf � 131.5 frames per second. For a given

viewing distance, the field of view (FOV) of the camera

determines the dimensions of the total surface area detected

by the instrument, which depends on the camera lens and

focal plane dimensions and is expressed in two angles(FLIR, 2005). In our experiments, we used a 25 mm lens

with each camera, which produced an approximate 408 �408 FOV. The camera relies on a sealed cryogenic chamber

that lowers the operating temperature of the detector array

to a temperature that is much lower than ambient (typically

70�80 K). Because most objects of interest are at higher

temperature, such cooled imaging devices have high thermal

sensitivity. Thermal images of objects from the FLIRThermoVision SC8000 result in a stereotypical intensity

pattern that is generally highest in the centre of mass of the

body and cooler at the periphery (Figure 1). This usually

results in a distinct intensity pattern that is ideally suited for

analysis by computer vision algorithms that rely on stan-

dardized regions of an image for detection and recognition

(Betke et al., 2007).

Echidna Validation Instrument

The EVI is a ground-based, full-waveform digitizing,

scanning LiDAR. The EVI is equipped with a diode-

pumped solid-state 1064 nm Nd:YAG laser that emits a

beam of 14.9 ns pulses at a frequency of 2 kHz. The beamdiameter of EVI is 29 mm with a manually adjustable

divergence of 2�15 mrad. In this application, all the EVI

scans were collected at 5 mrad beam divergence. A rotating

458 angled mirror directs the laser beam of EVI to cover

zenith angles of 91308 in the vertical plane while the

instrument base rotates 1808 in the horizontal plane

simultaneously (Figure 2). Unlike most of the terrestrial

LiDAR systems that record a single range for each laser

shot, the return sensor system digitizes and records the

intensity of the pulse reflected from targets along the entire

transmission path continually throughout a scan. The

waveform is recorded using a sampling rate of 2 GHz,

which equates to one sample every 7.5 cm of range from the

instrument. A complete EVI scan takes about 20 minutes.

The EVI data is recorded at 4 mrad angular resolution and

then resampled to 5 mrad resolution. The construction

theories and mechanical details of EVI are summarized in

Culvenor et al. (2005). The EVI can record full-waveform

signals from the entire field of view in the entire upper

hemisphere, as well as part of the lower hemisphere, from a

Figure 2. Multiangular scanning mode of the Echidna Validation

Instrument.

Figure 1. Composite of a sequence of a high-speed thermal

infrared images, subsampled at every 10th frame, that show the

flight path of one bat that emerged from the barn on the left and

flew into the forest at the lower right of the image frame.

Vol. 39, No. S1, Suppl. 1 2013

S114 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

distance of up to 100 m or more in medium density forests.

The EVI scan geometry is shown in Figure 2.

Plot layout

Big brown bats (Eptesicus fuscus), belonging to a roosting

colony in Petersham, Mass., were recorded as they emerged

from a crevice near the peak of the barn and flew into and

above the adjacent forest after dusk on 2 June 2010. We

deployed two FLIR ThermoVision SC8000 cameras with a

3 m baseline and pointed them towards the barn with an

overlapping FOV to record thermal video of bats as they left

the barn and commuted through the forest. Three fixed

calibration targets, one with a hot pack and two with cold

packs taped to the end of poles, were positioned so that they

would appear in the FOV of both cameras (A and B in

Figure 3). Owing to the temperature difference of the

calibration targets with the surrounding environment, the

hot target appeared as a patch of bright pixels and the cold

targets as dark pixels in the image data collected with the

thermal camera.

Nine EVI scans were acquired in a 25 m grid pattern over

the forest area near the barn (Figure 3). The site area

consists of a barn on the left, a dense broadleaf forest near

the barn, and several adjacent areas of an open field and

nearby paths without vegetation. Following the methodol-

ogy described in Yang et al. (2013), EVI scans can be

combined by adjusting translation and rotation matrices to

reconstruct a 100 m � 100 m 3-D structure of the forest on

the site. Because the flight activities of bats captured by

thermal cameras were limited to areas close to the barn and

did not extend far into the forest, a 25 m � 37.5 m region of

interest (x � [�50, �25] and y � [�12.5, 25]), was cropped

from the original hectare. The center of the barn is located at

(� 45, � 5) over the (x, y) plane.

The locations of the thermal cameras, as well as the

locations of thermal calibration targets were surveyed using

a Nikon AX-2S instrument and were then co-registered into

the 3-D coordinate system of the forest reconstruction.

Reconstruction of bat flight trajectories

Calibration of thermal camera

The thermal cameras were calibrated both in space and

time to reconstruct the flight trajectories of bats. To

synchronize datasets collected by multiple cameras in time,

a function generator was connected to the cameras, which

triggered them to start at the same time. Spatial calibration

was used to acquire the relative orientation between the two

cameras. That is, the calibration matrix P registered

the image coordinates (u, v) for the two cameras into one

(x, y, z) world coordinate system (Figure 4). In the field, a

calibration video was recorded prior to acquiring the actual

video emergences of bats. A mobile calibration device was

fabricated from PVC and aluminum tubing (Figure 5A).

A field crew member rotated and waved the calibration

device over the area detected by both cameras. N pairs of

image coordinates (u, v) of targets and were recorded

and extracted from the calibration video (Figure 5B). From

the N pairs of (u, v), a ‘‘wand-based self-calibration

Figure 3. Plot layout over the Petersham site.

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S115

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

strategy’’ (Hartley and Zisserman, 2004) for the spatial

calibration of thermal camera was applied. The calibration

matrices PA and PB are produced when the standard

deviation of the N pairs of estimated distances betweencalibration targets and , Da,b, normalized by the known

length of the wand is below 0.1.

Direct linear transformation (DLT)

EcoTracker software (Betke et al., 2010) was used torecord the bat flight trajectory datasets. By clicking on the

position of each bat in the first frame, this software can

automatically track and record the frame ID, as well as the

image coordinates (u, v) corresponding to the two-dimen-

sional position of the bat. The jth bat flight trajectory F j,

which is composed by m pairs of corresponding points that

are visible for both cameras, can be described by the

following sequence of point pairs:

Fjm �

uj;A; vj;A

� �1; uj;A; vj;A

� �2� � � uj;A; vj;A

� �m

uj;B; vj;B

� �1; uj;B; vj;B

� �2� � � uj;B; vj;B

� �m

8<: (1)

where the subscripts A and B are used to denote the

information collected by camera A and camera B.

The DLT method (Abdel-Aziz and Karara, 1971) is a

traditional reconstruction procedure that has been widely

used in biomechanics to obtain 3-D world coordinatesfrom film and video records. For each bat trajectory, the

image coordinates (u?, v?) were expressed as 3-D homo-

genous coordinates Q � (u, v, w) and the target 3-D world

coordinates (x?, y?, z?) were expressed as 4-D homogeneous

coordinates O � (x, y, z, w) where uw; v

w

� �¼ u0; v0ð Þ and

xw; y

w; z

w

� �¼ x0; y0; z0ð Þ. Using the calibration matrix PA and

PB obtained from the calibration process described pre-

viously, assuming the rows of each calibration matrix isdenoted as ri,i � 1, 2, 3, we can express the transformation

of camera coordinates and world coordinates as:

QA ¼ PAO)uA

vA

wA

0@

1A ¼ PA

x

y

z

w

0BB@

1CCA (2)

QB ¼ PBO)uB

vB

wB

0@

1A ¼ PB

x

y

z

w

0BB@

1CCA (3)

Two cameras will give four linear constraints as follows:

u0A rPA

3 �O� �

� rPA

1 �O ¼ 0 (4a)

v0A rPA

3 �O� �

� rPA

2 �O ¼ 0 (4b)

u0B rPB

3 �O� �

� rPB

1 �O ¼ 0 (4c)

v0B rPB

3 �O� �

� rPB

2 �O ¼ 0 (4d)

Figure 4. Reference frame for image coordinate system (u, v) and

world coordinate system (x, y, z).

Figure 5. Thermal calibration devices. (A) Diagram of mobile

calibration device; (B) thermal calibration targets in camera B’s

FOV as an example. (Fixed calibration devices: and , cold;

, hot. Mobile calibration targets: , cold and , hot.)

Vol. 39, No. S1, Suppl. 1 2013

S116 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

Here O is unknown and can be solved based on Equations

(4a) to (4d). Therefore, the flight trajectories of bats Fjm in

Equation (1) can be represented by a 3-D point group after

the reconstruction of the 3-D world coordinate system fromthe two thermal cameras using the DLT method

Fjm � x01; y01; z01ð Þj; x02; y02; z02ð Þj; � � � x0m; y0m; z0mð Þ

j; (5)

The EcoTracker (Betke et al., 2010) software uses the

DLT method to compute the 3-D trajectory in Equation (5)

based on the 2-D trajectories as in Equation (1) and the

calibration matrices.

Three-dimensional forest reconstruction

In any EVI scan, returned pulses within each laser shot

can be located unambiguously in space relative to the

instrument by three geometric dimensions: zenith angle

(u), azimuth angle (8), and the effective range (r) (Figure 2).

The intensity peaks within each EVI waveform are detectedand their intensity aj are directly displayed as 3-D point

clouds by converting the original polar coordinates of u, 8,

and r to Cartesian coordinates of x, y, and z. The full width

at half maximum (FWHM) that describes the pulse shape,

the time sequence along the laser shot (1st, 2nd, 3rd, 4th . . .),and the polar coordinates of the peak value are also

recorded as auxiliary information. Thus, the 3-D point

clouds rearrange the laser signals in a more convenient wayfor visualization while leaving the original information

unchanged. Following the methodology described in Yang

et al. (2013), a 3-D forest reconstruction was built from nine

EVI scans utilizing surveyed scan geolocations. Based on the

recorded pulse shape information, the merged point cloud of

forest reconstruction was further classified into ground,

foliage, and trunk. A Digital Elevation Model (DEM) and a

Canopy Height Model (CHM) were interpolated andmapped based on these categories of points to characterize

the terrain and basic forest structure over the site.

Registration of forest reconstruction and bat flight trajectory

After the 3-D thermal camera system is reconstructed

using the DLT method, the coordinates of the thermal

calibration targets , , and can be extracted as

x01ð Þ; y01ð Þ; z01ð Þ

� �; x02ð Þ; y02ð Þ; z02ð Þ

� �; and x03ð Þ; y03ð Þ; z03ð Þ

� �. The

thermal calibration targets, surveyed in the field, were co-

registered into the 3-D forest reconstruction with coordi-

nates x001ð Þ; y001ð Þ; z001ð Þ

� �; x002ð Þ; y002ð Þ; z002ð Þ

� �; and x003ð Þ; y003ð Þ; z003ð Þ

� �.

Rotation (R?) and the translation matrix (T?) that can

transfer the 3-D world coordinate system of the thermal

infrared camera into the 3-D forest reconstruction were

calculated using the coordinates of the fixed calibration

targets. Thus, the flight trajectories of bats, Fjm, were

registered to the 3-D forest reconstruction coordinate system

Fj0

m ¼ FjmR0 þ T 0 (6)

Fjm � x001; y001; z001ð Þj; x002; y002; z002ð Þj; � � � x00m; y00m; z00mð Þj; (7)

Properties of bat flight

For the jth bat, the magnitude of its instantaneous

velocity during flight Vj at ith location, can be calculated

from the registered 3-D coordinates of its 3-D flight

trajectory Fjm, the associated frame ID I, and the frame

rate Rf

Vji ¼

Di�1;i

Ii � Ii�1ð Þ=Rf

(8)

where Di�1, i is the distance between the ith location and the

(i�1)th location:

Di�1;i ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffix00i � x00i�1ð Þ2þ y00i � y00i�1ð Þ2þ z00i � z00i�1ð Þ2

q(9)

The flying altitude Hji can be calculated from the

registered 3-D coordinates of bat flight trajectory Fj0

m and

the DEM datasets generated from 3-D forest reconstruction

Hji ¼ z00i � Gi (10)

where Gi is the ground height at the ith location.

Results

Registration of 3-D flight trajectories of bats and 3-D forest

reconstruction

The registration of flight trajectories of bats and the 3-D

forest reconstruction provided a fine-scale reconstruction of

each flight path within and along the forest edge and

canopy. During the 15 minute recording periods, 26 bats

(j � 1�26) were recorded and are numbered in the order of

appearance. Among the 26 bats, 24 of them emerged from a

crevice at the peak of the barn and flew into the forest,

whereas bat 1 and bat 13 flew into the FOV of the camerasfrom behind the cameras. Therefore, bat 1 and bat 13 were

not included in the following flight velocity and altitude

analysis of the trajectories immediately following emergence.

The 24 flight trajectories of the bats emerging from the

crevice at the peak of the barn are shown in Figure 6, each

with a unique color. The barn can be seen on the upper left

corner of Figure 6A and front region of Figure 6B. The

registration of flight trajectories of the bats and the 3-Dforest reconstruction provides a fine-scale reconstruction of

each flight path within and along the forest edge and canopy.

DEM and CHM for the Petersham bat site

A DEM and a CHM helped in defining the 3-D forest

space where the bats foraged. They were derived from

the 3-D forest reconstruction following the methodology

described in Yang et al. (2013). This height information was

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S117

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

used to characterize flight characteristics such as altitude

and flight velocity of the emerging bats and their correlation

with primary forest structure.

This site is relatively flat over most of the area; however, it

contains some obvious terrain variation over our selected

region of interest where the bats were observed (Figure 7A).

Considering the location of the barn, there is a stable

downward slope on the ground along the path that bats

emerged and flew into the forest. The canopy over the

hectare is less than 15 min height and is composed of

deciduous trees (from field observations) (Figure 7B). The

canopy over the selected region of interest where the bat

trajectories were detected is comparatively sparser and

shorter than the rest of the forested area, with canopy

heights of around 5�8 m.

Flight velocity and altitude

To analyze flight velocity and altitude second by second, we

labeled the 5 seconds after emergence as T1, T2, T3, T4, and T5.Twenty-four bat flight velocity curves, (excluding bat 1 and bat

13 of the total 26 trajectories) were imaged emerging from the

crevice at the peak of the barn (Figure 8A). After the bats

emerged, they accelerated to approximately 6 m/s to 8 m/s

Figure 6. 3-D trajectories of emerging bats reconstructed from videos recorded with two

thermal infrared cameras and a 3-D forest reconstruction using EVI. (A) Top view (the front

and two sides of the barn can be seen as three green lines in the left centre of the LiDAR

image); and (B) side view (a profile of the front and one side of the barn can be seen in the

LiDAR image in the lower centre of the image).

Vol. 39, No. S1, Suppl. 1 2013

S118 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

within T1. They then reduced their velocity for a brief period

before entering into a more or less random velocity variation.

The maximum velocity recorded was approximately 12 m/s

and the minimum velocity was approximately 2.5 m/s, exclud-

ing the acceleration period T1.

As shown in Figure 8B, bats emerged from the barn,

which is about 8 m above ground level (AGL) and descended

1�2 m during T1 of the emergence. Then, during T1 and T2,

some bats started to climb to 8 m while others continued to

descend. The maximum flight altitude of these bats after

emergence was also about 8 m AGL and the minimum flight

altitude was about 2 m AGL.

The average velocity after T1 of emergence for the whole

bat group is 5.9 m/s and the average flight altitude for the

whole bat group is 6 m, which corresponds to the average

height of the forest canopy (6.5 m). T1 of the emergence was

excluded from the calculation to emphasize the difference

between individual bats (Figure 9).

Categorized bat flight patterns

The 24 flight trajectories can be grouped into four

categories with different flight patterns based on the 3-D

bat trajectory-forest registration shown in Figure 6: (A)

trajectories that curved to left of the flight direction (bats 4,

12, 22, 16); (B) high trajectories that curved to the right of

the flight direction (bats 7, 8. 9, 15, 16, 17, 20, 23, 24, 25);

(C) low trajectories that curved to the right of the flightdirection (bats 2, 5, 6, 10, 14, 19, 21); and (D) mixed

trajectories (bats 3, 11, and 18).

The flight altitude showed a systematic behavior between

different categories (Figure 10). Category A and category C

Figure 7. (A) Digital Elevation Model and (B) Canopy Height Model over the Petersham bat

site (units in metres). (Orange box is the selected region of interest.)

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S119

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

trajectories dropped in altitude after T1 following emer-

gence. Category C trajectories increased in altitude during

T2, and then started to drop in altitude. Category D

trajectories were more unstable and displayed mixed beha-

viors. The velocity showed a less systematic behavior

between different categories: the velocity of category A

and category D trajectories became more variable after the

T1, whereas the variation of category B and category C

trajectories contained similar trends within the category.

Bats in category B slowed down during T2 and then speeded

up during T3. The category C bats generally slowed down

during T2 and T3. Both the flight altitude and the velocity

became unpredictable after T3.

The behaviorof the bats during the first five seconds after their

appearance can be generalized based on the above information

as: T1, sped up and dropped height; T2, slowed and adjusted

flight direction; T3, slowed further down or sped up while

continuing to drop in height; and T4 and T5, spread to forest.

Flight pattern to avoid obstacles

The registration of bat flight trajectories and 3-D forest

reconstruction enabled us to examine the initial bat flight

trajectories in fine detail. Videos available at ftp://crsos.bu.

edu/geog-grad/xiaoyuan/bat_movies/ show the bat trajectories

Figure 8. Properties of flight paths of 24 big brown bats that emerged from a small

crevice located at the peak of the barn. (A) Flight velocity vs. time; (B) flight

altitude vs. time.

Vol. 39, No. S1, Suppl. 1 2013

S120 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

associated with the forest reconstruction for each category.2

A short 3 m tree with coordinates [�37, 2] blocked the direct

path of all trajectories during T2 and T3. The bats

presumably used echolocation to avoid hitting any obstacles

by flying around it following different routes. Combining the

information from previous sections, it would appear that the

bats in category A and category D chose more variable

routes; however, bats in category B flew along one side of the

tree with higher trajectories and lower velocities, and bats in

category C flew along the opposite side of the tree with

lower trajectories and higher velocities.

Discussion

Improving the identification and reconstruction of the bat

flight trajectories

In this study, we present a demonstration of the potential

power of combining two novel technologies, thermal cam-

eras and terrestrial LiDAR, to study bat flight behavior.

Owing to the initial experimental design, the observation

and analysis of bat flight behavior for this particular field

acquisition was limited. Better experimental design should

be adopted in the future to extend both the temporal and

spatial scale of the thermal observations of bat flight

trajectories. For example, despite installation and mainte-

nance difficulties, a tower-mounted set of thermal cameras,

which would enable repeat data collection at the same site

over several time sequences, might be more suitable to

record flight activity of bats within a particular forest.

Supporting data sources such as an acoustic tracking device

(Jung et al., 2012) and information from small GPS units

that can be installed in individual bats, would also help

increase the observation scales and accuracy. More ad-

vanced methodology such as improved computer vision

algorithms also needs to be incorporated in the future to

identify and reconstruct the flight trajectory of bats. The

EcoTracker software used in the paper was recently updated,

which enables it to automatically detect and track dispersing

and foraging colonies with a large number of bats and over a

coarser spatial scale.

Characterizing bat foraging behavior with respect to the forest

structure

By registering the flight trajectories of bats and the forest

reconstruction to the same 3-D space, a fine-scale recon-

struction of the flight patterns along forest edge was built.

However, little ecological hypothesis and prediction can be

drawn from this initial demonstration because of the limited

data collection. LiDAR data played an important role in

defining the 3-D forest space where bats forage. To analyze

the relationship between bat activities and forest structure, a

larger more expanded 3-D forest reconstruction model is

needed. The 1 ha reconstruction area assembled from nine

scans of terrestrial LiDAR EVI can be enlarged by combin-

ing multiple LiDAR sensors if a larger spatial scale analysis

is needed. Recent studies linking airborne LiDAR with

terrestrial LiDAR provide the possibility of characterizing

Figure 9. Averaged flight velocity and flight altitude for 24 individual big brown bats

as they emerged from a crevice at the peak of the barn at the Petersham site.

2Categorized flight trajectories around a tree. Video 1A, category A,trajectories curved to left. Video 1B, category B, higher trajectoriescurved to the right. Video 1C, category C, lower trajectories curvedto the right. Video 1D, category D, mixed trajectories.

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S121

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

Figure 10. Four distinct group behaviors were found for big brown bats as they emerged from the barn roost, based a

categorization of their flight velocity and altitude. (A) Category A, trajectories curved to left; (B) category B, higher

trajectories curved to the right; (C) category C, lower trajectories curved to the right; and (D) category D, mixed trajectories.

Vol. 39, No. S1, Suppl. 1 2013

S122 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

3-D forest structure over several square kilometres. For exam-

ple, Zhao et al. (2012) showed a general agreement between

foliage profiles retrieved from EVI 3-D reconstructions and

airborne Laser Vegetation Imaging Sensor foliage profiles.

Moreover, a simplified yet representative 3-D digital forest

model that can accurately characterize the distribution and

size of 3-D gaps is needed for the future. Researchers are

beginning to build similar models in the field of architecture

through surface reconstruction algorithms, e.g., Bolle and

Vemuri (1991); however, these applications are limited in

forest scenarios because of their more complex 3-D struc-

tures. Several commercial LiDAR scanners such as Riegl,

FARO, and Leica have been successfully used to scan forest

structures and natural landscapes with first returns (Bienert

et al., 2006; 2007; Fleck et al., 2007; Omasa et al., 2008;

Calders et al., 2011; Forbriger et al., 2012; Vaccari et al.,

2013). However, the application of these instruments is

limited over larger spatial areas (e.g., 1 ha) or for providing

complex foliage clumping and understory characteristics of a

densely vegetated multistory canopy. The full waveform

property of EVI, on the other hand, helps to identify and

characterize multiple scattering events along the entire path

of a laser pulse return; therefore, the EVI scans contain 3-D

gap information that can be used to derive important

structural parameters such as LAI, foliage profile, and

clumping (Strahler et al., 2008; Jupp et al., 2009a; Zhao

et al., 2011; 2012). Three-dimensional gap information is

essential for characterizing bat flight trajectories within the

canopy that allows more detailed study over bat foraging

behaviors. The comparison between field collected tree

densities and EVI retrieved tree densities in a New England

forest (Yao et al., 2011) also showed a high R2 value around

0.9. Thus, this technique of merging point clouds from

multiple LiDAR scans to reconstruct both individual trees

and entire stand segments provides more complete forest

structural information. In addition, 3-D forest reconstruc-

tion provides a more efficient way to acquire many types of

forest structural information with less effort and at a

potentially lower cost than standard inventory and measure-

ment methods (Yang et al., 2013). Therefore, a closer

examination of the structural information in 3-D forest

reconstructions, such as horizontal and vertical gaps, can

provide an improved representation of actual forest structure

for further bat foraging behavior studies.

Conclusion

In this paper, we explored an integrated methodology to

study how New England big brown bats (Eptesicus fuscus)

responded to forest clutter by combining two 3-D datasets: the

flight trajectories of bats recorded and reconstructed using

two FLIR ThermoVision SC8000 cameras and forest recon-

structions generated by full-waveform terrestrial LiDAR EVI.

The combination of the two 3-D datasets provided a fine-scale

reconstruction of the flight characteristics adjacent to and

within the forests. In future research, with improved

recording protocol with thermal cameras, additional and

longer bat flight trajectories can be obtained and studied

using this technology. Forest structure variables such as

height, canopy openness, gap length, understory, and

overstory foliage extracted from the digital forest model

represent future possibilities to study the larger scale

correlations between bat foraging behavior and forest

structure in species specific projects.

Acknowledgements

This research was supported by NASA grant NNG-

0GG192G, NSF grant DBI-0923389 and NSF grant IIS

0910908. The support is gratefully acknowledged. The

authors acknowledge Mr. James Baird of Petersham,

Mass., for making the site available for study and scanning

and thank the anonymous reviewers for critical comments.

ReferencesAbdel-Aziz, Y.I., and Karara, H.M. 1971. Direct linear transformation

from comparator coordinates into object space coordinates in close-

range photogrammetry. In American Society of Photogrammetry, Falls

Church, VA, USA. pp. 18.

Anderson, J., Martin, M.E., Smith, M.L., Dubayah, R.O., Hofton, M.A.,

Hyde, P., Peterson, B.E., Blair, J.B., and Knox, R.G. 2006. The use of

waveform LiDAR to measure northern temperate mixed conifer and

deciduous forest structure in New Hampshire. Remote Sensing of

Environment, Vol. 105, No. 3, pp. 248�261. doi: 10.1016/j.rse.2006.

07.001.

Anderson, J.E., Plourde, L.C., Martin, M.E., Braswell, B.H., Smith, M.L.,

Dubayah, R.O., Hofton, M.A., and Blair, J.B. 2008. Integrating wave-

form LiDAR with hyperspectral imagery for inventory of a northern

temperate forest. Remote Sensing of Environment, Vol. 112, No. 4, pp.

1856�1870. doi: 10.1016/j.rse.2007.09.009.

Bass, M.S., Finer, M., Jenkins, C.N., Kreft, H., Cisneros-Heredia, D.F.,

McCracken, S.F., Pitman, N.C.A., English, P.H., Swing, K., and Villa,

G. 2010. Global conservation significance of Ecuador’s Yasunı National

Park. PloS one, Vol. 5, No. 1, pp. e8767. doi: 10.1371/journal.pone.

0008767.

Bernard, E. 2001. Vertical stratification of bat communities in primary

forests of Central Amazon, Brazil. Journal of Tropical Ecology, Vol. 17,

No. 01, pp. 115�126. doi: 10.1017/S0266467401001079.

Betke, M., Hirsh, D.E., Bagchi, A., Hristov, N.I., Makris, N.C., and Kunz,

T.H. 2007. Tracking large variable numbers of objects in clutter. In IEEE

Computer Society Conference on Computer Vision and Pattern Recogni-

tion (CVPR), June, Minneapolis, MN, USA. pp. 8.

Betke, M., Hirsh, D.E., Makris, N.C., McCracken, G.F., Procopio, M.,

Hristov, N.I., Tang, S., Bagchi, A., Reichard, J.D., and Horn, J.W. 2008.

Thermal imaging reveals significantly smaller Brazilian free-tailed bat

colonies than previously estimated. Journal of Mammalogy, Vol. 89,

No. 1, pp. 18�24. doi: 10.1644/07-MAMM-A-011.1.

Betke, M., Wu, Z., Immermann, E., Theriault, D.H., and Permerlani, L.

2010. EcoTracker: Software to annotate thermal videos of bats by visual

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S123

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

inspection, detect and track bats automatically, and reconstruct their

three-dimensional flight paths.

Bienert, A., Scheller, S., Keane, E., Mohan, F., and Nugent, C. 2007. Tree

detection and diameter estimations by analysis of forest terrestrial

laserscanner point clouds. In ISPRS WG III/3 III/4 V/3 VIII/11

Workshop on ‘‘Laser Scanning 2007 and SilviLaser 2007’’, 12�14 Sep,

2007, Espoo, Finland. pp. 50�55.

Bienert, A., Scheller, S., Keane, E., Mullooly, G., and Mohan, F. 2006.

Application of terrestrial laser scanners for the determination of forest

inventory parameters. In ISPRS Commission V Symposium on ‘‘Image

Engineering and Vision Metrology’’, 25�27 Sep, 2006, Dresden,

Germany. pp. Part 5.

Blair, J.B., Rabine, D.L., and Hofton, M.A. 1999. The laser vegetation

imaging sensor: a medium-altitude, digitisation-only, airborne laser

altimeter for mapping vegetation and topography. ISPRS Journal of

Photogrammetry and Remote Sensing, Vol. 54, No. 2, pp. 115�122. doi:

10.1016/S0924-2716(99)00002-7.

Bolle, R.M., and Vemuri, B.C. 1991. On three-dimensional surface

reconstruction methods. IEEE Transactions on Pattern Analysis and

Machine Intelligence, Vol. 13, No. 1, pp. 1�13. doi: 10.1109/34.67626.

Bonaccorso, F.J. 1979. Foraging and reproductive ecology in a Panamanian

bat community. Bulletin of the Florida State Museum, Vol. 24, No. 4, pp.

359�408.

Calders, K., Verbesselt, J., Bartholomeus, H.M., and Herold, M. 2011.

Applying terrestrial LiDAR to derive gap fraction distribution time

series during bud break. In Proceedings SilviLaser 2011, 11th interna-

tional LiDAR forest applications conference, 16�20 Oct. 2011, Tasmania,

Hobart, Tasmania

Cleveland, C.J., Betke, M., Federico, P., Frank, J.D., Hallam, T.G., Horn, J.,

Lopez Jr, J.D., McCracken, G.F., Medellın, R.A., and Moreno-Valdez,

A. 2006. Economic value of the pest control service provided by

Brazilian free-tailed bats in south-central Texas. Frontiers in Ecology

and the Environment, Vol. 4, No. 5, pp. 238�243. doi: 10.1890/1540-

9295(2006)004[0238:EVOTPC]2.0.CO;2.

Cryan, P., Diehl, R., Kunz, T.H., and Parsons, S. 2008. Analyzing bat

migration. In Ecological and behavioral methods for the study of bats.

Edited by T.H. Kunz, and S. Parsons. Johns Hopkins University Press,

Baltimore, Maryland, USA.

Culvenor, D.S., Jupp, D.L.B., Lovell, J.L., and Newnham, G.J. 2005.

Evaluation and validation of canopy laser radar systems for native and

plantation forest inventory � Summary Report (FWPRDC: PN 02.3902).

Forest & Wood Products Research & Development Corporation,

Melbourne.

Fenton, M.B. 1990. The foraging behavior and ecology of animal-eating

bats. Canadian Journal of Zoology, Vol. 68, No. 3, pp. 411�422. doi:

10.1139/z90-061.

Fleck, S., Obertreiber, N., Schmidt, I., Braunsa, M., Jungkunst, H.F., and

Leuschner, C. 2007. Terrestrial LiDAR measurements for analysing

canopy structure in an old-growth forest. In ISPRS WG III/3 III/4 V/3

VIII/11 Workshop on ‘‘Laser Scanning 2007 and SilviLaser 2007’’, 12�14

Sep, 2007, Espoo, Finland pp. 125�129.

FLIR 2005. The Ultimate Infrared Handbook for R&D Professionals. In

The Ultimate Resource Guide for Using Infrared in the Research and

Development Industry.

Forbriger, M., Hofle, B., Siart, C., Schittek, K., and Bubenzer, O. 2012.

High-resolution terrestrial laser scanning (TLS) on cushion peatlands-a

case study from the Peruvian Andes. In EGU General Assembly

Conference Abstracts. pp. 11485.

Gillam, E.H., McCracken, G.F., Westbrook, J.K., Lee, Y.F., Jensen, M.L.,

and Balsley, B.B. 2009. Bats aloft: variability in echolocation call

structure at high altitudes. Behavioral Ecology and Sociobiology,

Vol. 64, No. 1, pp. 69�79. doi: 10.1007/s00265-009-0819-1.

Hartley, R., and Zisserman, A. 2004. Multiple view geometry in computer

vision. 2nd edition. University Press, Cambridge.

Heithaus, E.R., Fleming, T.H., and Opler, P.A. 1975. Foraging patterns and

resource utilization in seven species of bats in a seasonal tropical forest.

Ecology, Vol. 56, No. 4, pp. 841�854. doi: 10.2307/1936295.

Hodgkison, R., Balding, S.T., Zubaid, A., and Kunz, T.H. 2003. Fruit bats

(Chiroptera: Pteropodidae) as seed dispersers and pollinators in a

lowland malaysian rain forest. Biotropica, Vol. 35, No. 4, pp. 491�502.

Holderied, M.W., and Jones, G. 2009. Flight dynamics. Ecological and

behavioral methods for the study of bats. Johns Hopkins University Press,

Baltimore, MD, USA, pp. 459�475.

Holderied, M.W., Korine, C., Fenton, M.B., Parsons, S., Robson, S., and

Jones, G. 2005. Echolocation call intensity in the aerial hawking bat

Eptesicus bottae (Vespertilionidae) studied using stereo videogrammetry.

Journal of Experimental Biology, Vol. 208, pp. 1321�1327. doi: 10.1242/

jeb.01528.

Hristov, N.I., Betke, M., and Kunz, T.H. 2008. Applications of thermal

infrared imaging for research in aeroecology. Integrative and Compara-

tive Biology, Vol. 48, No. 1, pp. 50. doi: 10.1093/icb/icn053.

Hristov, N.I., Betke, M., Theriault, D.E.H., Bagchi, A., and Kunz, T.H.

2010. Seasonal variation in colony size of Brazilian free-tailed bats at

Carlsbad Cavern based on thermal imaging. Journal of Mammalogy, Vol.

91, No. 1, pp. 183�192. doi: 10.1644/08-MAMM-A-391R.1.

Jung, K., Kaiser, S., Bohm, S., Nieschulze, J., and Kalko, E.K. 2012.

Moving in three dimensions: effects of structural complexity on

occurrence and activity of insectivorous bats in managed forest stands.

Journal of Applied Ecology, Vol. 49, No. 2, pp. 523�531. doi: 10.1111/

j.1365-2664.2012.02116.x.

Jupp, D.L.B., Culvenor, D.S., Lovell, J.L., Newnham, G.J., Strahler, A.H.,

and Woodcock, C.E. 2009a. Estimating forest LAI profiles and structural

parameters using a ground-based laser called ‘Echidna†

’. Tree Pysiology,

Vol. 29, No. 2, pp. 171�181. doi: 10.1093/treephys/tpn022.

Jupp, D.L.B., Culvenor, D.S., Lovell, J.L., Newnham, G.J., Strahler, A.H.,

and Woodcock, C.E. 2009b. Estimating forest LAI profiles and

structural parameters using a ground-based laser called ‘Echidna†’.

Tree physiology, Vol. 29, No. 2, pp. 171. doi: 10.1093/treephys/tpn022.

Kalko, E.K.V., and Handley, C.O. 2001. Neotropical bats in the canopy:

diversity, community structure, and implications for conservation. Plant

Ecology, Vol. 153, No. 1, pp. 319�333. doi: 10.1023/A:1017590007861.

Kastberger, G., and Stachl, R. 2003. Infrared imaging technology and

biological applications. Behavior Research Methods, Vol. 35, No. 3, pp.

429�439. doi: 10.3758/BF03195520.

Kimes, D.S., Ranson, K.J., Sun, G., and Blair, J.B. 2006. Predicting LiDAR

measured forest vertical structure from multi-angle spectral data.

Remote Sensing of Environment, Vol. 100, No. 4, pp. 503�511.

doi: 10.1016/j.rse.2005.11.004.

Krusic, R.A., Yamasaki, M., Neefus, C.D., and Pekins, P.J. 1996. Bat

habitat use in white mountain national forest. The Journal of Wildlife

Management, Vol. 60, No. 3, pp. 625�631. doi: 10.2307/3802081.

Vol. 39, No. S1, Suppl. 1 2013

S124 # 2013 CASI

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.

Kunz, T.H. 2004. Foraging habits of North American insectivorous bats.

Bat Echolocation Research: Tools, Techniques, and Analysis. Edited by

R.M. Brigham, E.K.V. Kalko, G. Jones, S. Parsons, and H.J.G.A.

Limpens. Bat Conservation International, Austin, TX. pp. 13�25.

Kunz, T.H., Gauthreaux, S.A., Hristov, N.I., Horn, J.W., Jones, G., Kalko,

E.K.V., Larkin, R.P., McCracken, G.F., Swartz, S.M., and Srygley, R.B.

2008. Aeroecology: probing and modeling the aerosphere. Integrative

and Comparative Biology, Vol. 48, No. 1, pp. 1. doi: 10.1093/icb/icn037.

Kurta, A., and Baker, R.H. 1990. Eptesicus fuscus. Mammalian Species,

No. 356, pp. 1�10. doi: 10.2307/3504258.

Lacki, M.J., Hayes, J.P., and Kurta, A. 2007. Bats in forests: conservation

and management. Johns Hopkins University Press, Baltimore.

Meyer, C.F.J., Kalko, E.K.V., and Kerth, G. 2009. Small scale fragmenta-

tion effects on local genetic diversity in two phyllostomid bats with

different dispersal abilities in Panama. Biotropica, Vol. 41, No. 1, pp. 95�102. doi: 10.1111/j.1744-7429.2008.00443.x.

Omasa, K., Hosoi, F., Uenishi, T.M., Shimizu, Y., and Akiyama, Y. 2008.

Three-dimensional modeling of an urban park and trees by combined

airborne and portable on-ground scanning LiDAR remote sensing.

Environmental Modeling and Assessment, Vol. 13, No. 4, pp. 473�481.

doi: 10.1007/s10666-007-9115-5.

Parker, G.G., and Brown, M.J. 2000. Forest canopy stratification*is it

useful? American Naturalist, Vol. 155, No. 4, pp. 473�484. doi: 10.1086/

303340.

Parsons, S., and Szewczak, J.M. 2009. Detecting, recording, and analyzing

the vocalizations of bats. In Ecological and Behavioral Methods for the

Study of Bats. 2nd edition. Edited by T.H. Kunz, and S. Parsons. Johns

Hopkins University Press, Baltimore.

Schnitzler, H.U., and Kalko, E.K.V. 1998. How echolocating bats search

and find food. Bat biology and conservation. Edited by T.H. Kunz and

P.A. Racey. Smithsonian Institution Press, Washington, DC. pp. 183�196.

Schull, M.A., Ganguly, S., Samanta, A., Huang, D., Shabanov, N.V.,

Jenkins, J.P., Chiu, J.C., Marshak, A., Blair, J.B., and Myneni, R.B. 2007.

Physical interpretation of the correlation between multi-angle spectral

data and canopy height. Geophysical Research Letters, Vol. 34, No. 18,

pp. L18405. doi: 10.1029/2007GL031143.

Simmons, J.A. 2005. Big brown bats and June beetles: multiple pursuit

strategies in a seasonal acoustic predator�prey system. Acoustics

Research Letters Online, Vol. 6, pp. 238. doi: 10.1121/1.1985957.

Strahler, A.H., Jupp, D.L.B., Woodcock, C.E., Schaaf, C.B., Yao, T., Zhao,

F., Yang, X., Lovell, J.L., Culvenor, D.S., Newnham, G.J., Ni-miester,

W., and Boykin-morris, W. 2008. Retrieval of forest structural para-

meters using a ground-based LiDAR instrument (Echidna†

). Canadian

Journal of Remote Sensing, Vol. 34, pp. 5426�5440. doi: 10.5589/m08-

046.

Strahler, A.H., Woodcock, C.E., Schaaf, C.B., Myneni, R.B., Ni-Meister,

W., Liu, J., Newnham, G.J., Culvenor, D.S., Jupp, D.L.B., and Lovell,

J.L. 2007. Rapid retrieval of forest structure using an under-canopy,

upward-scanning LiDAR (Echina†). Eos, Transactions. American

Geophysical Union Fall Meeting, Vol. 88, p. 52.

Strahler, A.H., Yao, T., Zhao, F., and Yang, X. 2009. Vegetation structure

and 3-D reconstruction of forests using ground-based Echidna†

LiDAR. Eos, Transactions. American Geophysical Union Fall Meeting,

Vol. 90, p. 52.

Theriault, D.H., Wu, Z., Hristov, N.I., Swartz, S.M., Breuer, K.S., Kunz,

T.H., and Betke, M. 2010. Reconstruction and analysis of 3D

trajectories of Brazilian free-tailed bats in flight. In Computer Science

Technique Report BUCS-TR-2010-027 (pp. 1�7): Boston Univeristy,

Boston, Mass.

Thies, W., and Kalko, E.K.V. 2004. Phenology of neotropical pepper plants

(Piperaceae) and their association with their main dispersers, two short-

tailed fruit bats, Carollia perspicillata and C. castanea (Phyllostomidae).

Oikos, Vol. 104, No. 2, pp. 362�376. doi: 10.1111/j.0030-1299.2004.

12747.x.

Thies, W., Kalko, E.K.V., and Schnitzler, H.U. 2006. Influence of

environment and resource availability on activity patterns of Carollia

castanea (Phyllostomidae) in Panama. Journal of Mammalogy, Vol. 87,

No. 2, pp. 331�338. doi: 10.1644/05-MAMM-A-161R1.1.

Vaccari, S., van Leeuwen, M., Calders, K., Coops, N.C., and Herold, M.

2013. Bias in LiDAR-based canopy gap fraction estimates. Remote

Sensing Letters, Vol. 4, No. 4, pp. 391�399. doi: 10.1080/2150704X.

2012.742211.

Weinbeer, M., and Kalko, E.K.V. 2004. Morphological characteristics

predict alternate foraging strategy and microhabitat selection in the

orange bellied bat, Lampronycteris brachyotis. Journal of Mammalogy,

Vol. 85, No. 6, pp. 1116�1123. doi: 10.1644/BWG-206.1.

Wu, Z., Hristov, N.I., Hedrick, T.L., Kunz, T.H., and Betke, M. 2009a.

Tracking a large number of objects from multiple views. In IEEE 12th

International Conference on Computer Vision (ICCV), September/

October, Kyoto Japan. pp. 8.

Wu, Z., Hristov, N.I., Kunz, T.H., and Betke, M. 2009b. Tracking-

reconstruction or reconstruction-tracking? Comparison of two multiple

hypothesis tracking approaches to interpret 3D object motion from

several camera views. In Proceedings of the IEEE Workshop on Motion

and Video Computing (WMVC) Snowbird, December, Utah. pp. 8.

Yang, X., Strahler, A.H., Schaaf, C.B., Jupp, D.L.B., Yao, T., Zhao, F.,

Wang, Z., Culvenor, D.S., Newnham, G.J., Lovell, J.L., Dubayah, R.O.,

woodcock, C.E., and Ni-Meister, W. 2013. Three-dimensional forest

reconstruction and structural parameter retrievals using a terrestrial full-

waveform LiDAR instrument (Echidna†). Remote Sensing of

Environment, Vol. 135, pp. 36�51. doi: 10.1016/j.rse.2013.03.020.

Yao, T., Yang, X., Zhao, F., Wang, Z., Zhang, Q., Jupp, D.L.B., Culvenor,

D.S., Newnham, G.J., Ni-Meister, W., Schaaf, C.B., Woodcock, C.E.,

and Strahler, A.H. 2011. Measuring forest structure and biomass in New

England forest stands using Echidna† ground-based LiDAR. Remote

Sensing of Environment, Vol. 115, No. 11, pp. 2965�2974. doi: 10.1016/

j.rse.2010.03.019.

Zhao, F., Strahler, A.H., Schaaf, C.L., Yao, T., Yang, X., Wang, Z., Schull,

M.A., Roman, M.O., Woodcock, C.E., and Olofsson, P. 2012. Measur-

ing gap fraction, element clumping index and LAI in Sierra Forest

stands using a full-waveform ground-based LiDAR. Remote Sensing of

Environment, Vol. 125, pp. 73�79. doi: 10.1016/j.rse.2012.07.007.

Zhao, F., Yang, X., Schull, M.A., Roman, M., Yao, T., Wang, Z., Zhang, Q.,

Jupp, D.L.B., Culvenor, D.S., Newnham, G.J., Ni-Meister, W., Schaaf,

C.B., Woodcock, C.E., Strahler, A.H., and Richardson, A. 2011.

Measuring effective leaf area index, foliage profile, and stand height in

New England forest stands using Echidna†

Validation Instrument (EVI)

ground-based LiDAR. Remote Sensing of Environment, Vol. 115, No. 11,

pp. 2954�2964. doi: 10.1016/j.rse.2010.08.030.

Zimmerman, G.S., and Glanz, W.E. 2000. Habitat use by bats in eastern

Maine. The Journal of Wildlife Management, Vol. 64, No. 4, pp. 1032�1040. doi: 10.2307/3803214.

Canadian Journal of Remote Sensing / Journal canadien de teledetection

# 2013 CASI S125

Can

adia

n Jo

urna

l of

Rem

ote

Sens

ing

Dow

nloa

ded

from

pub

s.ca

si.c

a by

MU

GA

R M

EM

OR

IAL

LIB

RA

RY

on

01/1

1/14

For

pers

onal

use

onl

y.