Sensors 2014, 14, 2089-2109; doi:10.

3390/s140202089
OPEN ACCESS

sensors
ISSN 1424-8220
www.mdpi.com/journal/sensors
Article

Mobile Robot Self-Localization System Using Single Webcam

Distance Measurement Technology in Indoor Environments
I-Hsum Li 1, Ming-Chang Chen 2, Wei-Yen Wang 3,*, Shun-Feng Su 2 and To-Wen Lai 3
1
Department of Information Technology in Lee-Ming Institute of Technology, No.2-2, Lizhuan Rd.,
Taishan Dist., New Taipei 243, Taiwan; E-Mail: [email protected]
2
Department of Electrical Engineering, National Taiwan University of Science and Technology,
43 Keelung Road, Section 4, Taipei 106, Taiwan;
E-Mails: [email protected] (M.-C.C.); [email protected] (S.-F.S.)
3
Department of Applied Electronics Technology, National Taiwan Normal University, 160, He-ping
East Rd., Section 1, Taipei 106, Taiwan; E-Mail: [email protected]

* Author to whom correspondence should be addressed; E-Mail: [email protected];

Tel.: +886-2-7734-3536; Fax: +886-2-2351-5092.

Received: 10 September 2013; in revised form: 22 January 2014 / Accepted: 23 January 2014 /
Published: 27 January 2014

Abstract: A single-webcam distance measurement technique for indoor robot localization

is proposed in this paper. The proposed localization technique uses webcams that are
available in an existing surveillance environment. The developed image-based distance
measurement system (IBDMS) and parallel lines distance measurement system (PLDMS)
have two merits. Firstly, only one webcam is required for estimating the distance.
Secondly, the set-up of IBDMS and PLDMS is easy, which only one known-dimension
rectangle pattern is needed, i.e., a ground tile. Some common and simple image processing
techniques, i.e., background subtraction are used to capture the robot in real time. Thus, for
the purposes of indoor robot localization, the proposed method does not need to use
expensive high-resolution webcams and complicated pattern recognition methods but just
few simple estimating formulas. From the experimental results, the proposed robot
localization method is reliable and effective in an indoor environment.

Keywords: indoor robot localization; image-based distance measurement system;

parallel lines distance measurement system
Sensors 2014, 14 2090

1. Introduction

Autonomous robots have a wide range of potential applications in security guards, house cleaning
and even warfare. Most of them are equipped with position measurement systems (PMSs) for the
purpose of precisely locating themselves and navigating in their working fields. Three typical
techniques [1] in PMSs are triangulation, scene analysis, and proximity. The triangulation technique
uses the geometric properties of triangles to compute object locations. The most well-known technique
is the Global Positioning System (GPS). However, GPS, as it is satellite dependent, has an inherent
problem of accurately determining the locations of objects within a building [2]. A proximity
location-sensing technique entails determining when an object is “near” a known location, and
the object’s presence can be sensed via some limited range physical phenomenon. Some famous
techniques are detecting physical contact [3,4] or monitoring wireless cellular access points [5,6]. The
scene analysis location sensing technique uses features of a scene observed from a particular vantage
point to draw conclusions about the location of the observer or of objects in the scene. Some
well-known techniques are a radar location system [7] or a visual images location system [8].
In an indoor localization technique, the infrared light [6], ultrasonic [9], laser range finder [10,11],
RFID [12], and radar [13] are the most popular wireless techniques. Diffuse infrared technology is
commonly used to realize indoor locations, but the short-range signal transmission and line-of-sight
requirements limit the growth. Ultrasonic localization [9] uses the time-of-flight measurement
technique to provide location information. However, the use of ultrasound requires a great deal of
infrastructure in order for it to be highly effective and accurate. Laser distance measurement is
executed by measuring the time that it takes for a laser light to be reflected off a target and returned
back to the sender. Because the laser range finder is a very accurate and quick measurement device,
this device is widely used in many applications. In [10,11], Subramanian et al. and Barawid et al.
proposed an autonomous vehicle guidance system based on a laser rangefinder. The laser rangefinder
was used to acquire environment distance information that can be used to identify and avoid obstacles
during navigation. In [14], Thrun et al. provided an autonomous navigation method based on a particle
filter algorithm. In this study, the laser rangefinder can receive all the measurement information that it
can utilize to compute the likelihood of the particles. These papers confirm that laser rangefinders are
high performance and high accuracy measurement equipment. However, their high performance relies
on high hardware costs. RFID-based localization uses RF tags and a reader with an antenna to locate
objects, but the detection of each tag only can work over approximately 4 to 6 meter distances. To improve
the low precision on location positioning, the well-known SpotON [15] technology uses an aggregation
algorithm based on radio signal strength analysis for 3D-location sensing. However, a complete system is
not available yet. An RF-based RADAR system [7,16,17] uses the 802.11 network adapter to measure
signal strengths at multiple base stations positioned to provide overlapping coverage for locating and
tracking objects inside buildings. Unfortunately, most cases to date cannot provide overall accuracy of
systems as optimal as desired. In indoor localization for robots, most of these wireless techniques are used
to perform scans of static obstacles around the robots, and the localization is calculated by matching those
scans with a metric map of the environment [18,19], but in dynamic environments the detected
static-features are often not enough for estimating a robust localization.
Sensors 2014, 14 2091

Li et al. [20] proposed a NN-based mobile phone localization technique using Bluetooth
connectivity. In this large-scale network, mobile phones equipped with GPS represent beacons, and
others could connect to the beacon phones with Bluetooth connectivity. By formulating the Bluetooth
network as an optimization problem, a recurrent neural network is developed to distributively find the
solutions in real time. However, in general, the sampling rate of Bluetooth is relatively low, and then
accurately estimating a moving object in real time is not easy. In [21], a recurrent neural network was
proposed to search a desirable solution for a range-free localization of WSNs under the condition that
the WSNs can be formed as a class of nonlinear inequalities defined on a graph. Taking advantage of
parallel computation of the NN, the proposed approach can effectively solve the WSN localization
problem, although the limited transmission bandwidth might cause difficulty in the localization.
Recently, image-based techniques have been preferred over wireless techniques [4,5,9,22]; this is
because they are passive sensors and are not easily disturbed by other sensors. In [1], a
portable-PC capable of marker detection, image sequence matching, and location recognition was
proposed for an indoor navigation task. JongBae et al. used the augmented reality (AR) technique to
achieve an average location recognition success rate of 89%, though the extra cost must be considered
in this technique. In [23], Cheoket et al. provided a method of localization and navigation in wide
indoor areas with a wearable computer for human-beings. Though the set-up cost is lower, this method
is not easy to implement and set up if users do not know the basic concept of electronic circuit analysis
and design. Furthermore, an imaged-based method for distance measurement was proposed in [24–29].
According to the transform equations in those papers, the distance can be calculated from the ratio of
the size between the pre-defined reference points and the measured object. In recent years, we have
seen growing importance placed on research in two-camera localization systems [30,31]. From two
different images, the object distances can be calculated by a triangular relationship. However, to ensure
the measuring reliability, the photography angle and the distance between two cameras must be
maintained at the same position. Due to the use of two cameras for the measuring device, the set-up
costs of the experimental environment will be increased.
Nowadays, surveillance systems exist in most modern buildings, and cameras have been configured
around these buildings. In general, one camera covers one specific area. In order to locate an
autonomous patrolling robot using existing cameras in buildings, a single-camera localization
technique must be developed for the patrolling robots. This study aims to develop a single-webcam
distance measurement technique for indoor robot localization with the purposes of saving set-up costs
and increasing the accuracy of distance measurements. In our approach, the working area setting can
be as simplified as possible, because the existing webcams in the surveillance environment can be
utilized without any change. For a single webcam in its working coverage area, we develop an
improved image-based distance measurement system (IBDMS) and a parallel lines distance
measurement system (PLDMS) to measure the location of a robot according to a known-size rectangle
pattern, i.e., a ground tile. This measurement system uses four points, i.e., the four corners of a ground
tile, to form a pair of parallel lines in the webcam image. Referring to the pair of parallel lines, we can
measure the location of a robot within the visual range of a webcam. Because of the fixed monitoring
area of an individual webcam, few simple image processing strategies are used to search for the robots
before going through IBDMS and PLDMS. First, we use the low-pass filter and on-line background
update method to reduce background noise, and adopt the image morphology to complete prospect
Sensors 2014, 14 2092

information and to remove the slight noise. When the mobile robot is located, IBDMS and PLDMS
can obtain the real-world coordinates of a mobile robot. Finally, the localization of a mobile robot can
be shown on the two-dimensional map immediately. Thus, for the purpose of indoor robot localization,
the proposed method does not need to use complicate pattern recognition methods, but just few simple
estimation formulas.

2. Photography Methods

Before locating a robot by the proposed single-webcam localization technique, the acquired images
must go through photograph processing for removing noise and unnecessary information. These
techniques include a gray scale, a background subtraction, a morphological image processing, and a
connected components labeling technique. Next, we briefly discuss the procedures [32] of these
photographic correction techniques used in this paper.

2.1. Camera Calibration

Distortion could happen in captured images, especially is cheap webcams are used. To attenuate
distortion of the captured images and thus increase the accuracy of the robot location task, the camera
calibration should be done before the localization is attempted. OpenCV has taken into account the
radial and tangential factors for the image distortion problem. The radial factor can be calculated by
the following equations:
xcorrected = x(1+ k1r 2 + k2r 4 + k3r 6 ) (1)
ycorrected = y(1+ k1r + k2r + k3r )
2 4 6
(2)

The tangential distortion can be corrected via the equations as follows:

xcorrected = x + [2 p1xy + p2 (r 2 + 2x 2 )] (3)

ycorrected = y + [ p1 (r 2 + 2y 2 ) + 2 p2 xy] (4)

In Equations (1–4) the pixel (x, y) is the image coordinate in the input image and (xcorrected , ycorrected) is
the image coordinate in the corrected output image. The distortion coefficient vector can be
represented as cdi  [k1 k2 p1 p2 k3 ] . Moreover, the unit conversion can be represented as:

x   fx 0 cx   xbefore 
y   0  
   fy c y   ybefore  (5)
 w  0 0 1   zbefore 

where w is explained by the use of homography coordinate system (and w  zbefore ), f x and f y are the
camera focal length, and cx and c y are the optical centers expressed in pixels coordinates. After calculating
the camera Equation (5) and the distortion coefficients cdi , the functions initUndistortRectifyMap() and the
remap() can calibrate the distorted images. Figure 1a shows the images before the calibration done for
three webcams, and Figure 1b shows the images after the calibration procedure. For the 1st webcam
(HD Webcam C310, Logitech, Lausanne, Switzerland) in our experimental environment, the distortion
coefficient vector is:
Sensors 2014, 14 2093

cd 1  [0.0895 -0.7064 -0.0034 0.0020 1.3227]

(6)
and the camera conversion matrix is:
 x  518.38 0 253.56  xbefore 
 y   0  
   518.07 644.23  ybefore  (7)
 w  0 0 1   zbefore 

For the 2nd webcam (HD Pro Webcam C920, Logitech, Lausanne, Switzerland) in our
experimental environment, the distortion coefficient vector is:
cd 2  [0.2460 -1.8737 -0.0023 -0.0043 5.4119] (8)

and the camera conversion matrix is:

 x  598.36 0 318   xbefore  (9)
 y   0  
   600.76 264.19  ybefore 
 w  0 0 1   zbefore 

For the 3rd webcam (HD Webcam PC235, Ronald, Osaka, Japan) in our experimental environment,
the distortion coefficient vector is:
cd 3  [0.8433 -12.8097 -0.0339 0.0185 111.8152] (10)

and the camera conversion matrix is:

 x  1272.37 0 378.46   xbefore 
 y   0  
   1283.87 249.52  ybefore  (11)
 w  0 0 1   zbefore 

Figure 1. Calibration procedure for the webcam images. (a) Before the calibration
produces for three webcams, (a.1) Logitech HD Webcam C310; (a.2) Logitech HD Pro
Webcam C920; (a.3) Ronald HD Webcam PC235. (b) Images after the calibration
produces for three webcams, (b.1) Logitech HD Webcam C310; (b.2) Logitech HD Pro
Webcam C920; (b.3) Ronald HD Webcam PC235.

(a.1) (a.2) (a.3)

Sensors 2014, 14 2094

Figure 1. Cont.

(b.1) (b.2) (b.3)

2.2. Image Segmentation

In a grayscale image, the value of each pixel carries only intensity information. It is known as a
black-and-white image, which is composed exclusively of shades of gray. Black is at the weakest
intensity and white is at the strongest one. The gray scale technique can change a color image into a
black-and-white image. The luminance fl ( x, y) of the pixel(x, y) is described as:

fl ( x, y)  0.299R( x, y )  0.587G( x, y )  0.114B( x, y ) (12)

where R(x,y), G(x,y), and B(x,y) are color values at the pixel(x, y) . Equation (12) can create binary images
by a threshold value t from a grayscale image:
255 if f l ( x, y )  t (13)
f binary ( x, y )  
 0 else

Subtracting a binary acquired image fcurrent ( x, y) from a binary background image fbg ( x, y) , we
can obtain a binary foreground image as:
f fg ( x, y) | fbg ( x, y)  f current ( x, y) | (14)

2.3. Morphological Image Processing

After the process of image segmentation, discontinuous edges and noise may happen in a
foreground image. These will cause wrong judgments during object identification. Therefore, this
paper utilizes some morphological image processing operations, such as dilation, erosion, opening and
closing, in order to enable the underlying shapes to be identified and optimally reconstruct the image
from their noisy precursors.

2.4. Connected-Components Labeling

The aim of connected-component labeling is to identify connected-components that share similar

pixel intensity values, and then to connect them with each other. The connected-component labeling
scans an image and groups pixels into one or more components according to pixel connectivity. Once
all groups are determined, each pixel is labeled with a grey level on the basis of the component.
Sensors 2014, 14 2095

According to the aforementioned discussions, we can locate a robot in the captured image in an
image-domain. Figure 2 shows the overall schemes of the image processing, and the experimental results
are shown in Figure 3. In Figure 4, Rcenter is the center of the robot in the processed image and can be easily
calculated by the simple average method. In this paper, Rcenter stands for the center-coordinate of the robot
in the image-domain.

Figure 2. Procedures of the image processing techniques for the robot localization.
Captured image (current) Background image

Gray Scale
Transformation

Background Subtraction

Binarization

Morphological Processing

Connected Components
Labeling

Average Method

Coordinate in the image domain

Figure 3. (a) Background image. (b) Captured image. (c) Gray scale of the background
image. (d) Gray scale of the captured image. (e) Background subtraction. (f) Binarization
processing of the foreground image. (g) Morphological processing of the foreground
image. (h) Connected components labeling.

(a) (b)
Sensors 2014, 14 2096

Figure 3. Cont.

(c) (d)

(e) (f)

(g) (h)

Figure 4. Coordinates of the mobile robot in the image-domain.

Rcenter
Sensors 2014, 14 2097

3. Mobile Robot Localization System with Single Webcam

After the captured images go through image processing, we can locate the robot in the
image-domain. Then, we should calculate the coordinates of the robot in the image. That is, two distances,
the x-axis and the y-axis, should be determined: 1. di represents the distance between Rcenter and the
webcam: 2. wi represents the distance between Rcenter and the wall, as shown in Figure 4. In this paper the
IBDMS is used to calculate the distance di, and the PLDMS is used to calculate the distance wi.

3.1. Experimental Map

Figure 5 shows the map of our experimental environment. In the map, the coordinate of the first
webcam is set to ( x1 , y1 ) , the second one is set to ( x2 , y2 ) , and the third one is set to ( x3 , y3 ) .
walli (i  1, 2,3) represents the distance between the ith webcam and the wall. di (i  1, 2,3) represents
the distance between the ith webcam and Rcenter. wi (i  1, 2,3) is the distance between the wall and
Rcenter in the ith webcam covering area. Clearly, the coordinates of webcams ( xi , yi ) and the distances
walli are given. Therefore, if distances di and wi can be calculated, we can easily represent the robot
with its coordinate in the area, which is covered by one of the set-up webcams. The equations for
calculating the coordinate of the robot are defined in Equations (15–17):
 X 1  x1  d1 (15)
webcama1 : 
Y1  y1  ( wall1  w1 )
 X 2  x2  ( wall2  w2 ) (16)
webcama2 : 
Y2  y2  d 2

and:
 X 3  x3  d 3 (17)
webcama3 : 
Y3  y3  ( wall3  w3 )
where ( X i , Yi ) ( i  1,2,3) are the coordinate of the robot in the covering area of the ith webcam.

Figure 5. Map of our experimental location.

(0,0)
XX

(x2 , y2 ) 1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

2
wall2

1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

1 d1 Rcenter
( x1, y1 ) d2
wall1 w1

w2
1800.0 mm x 800.0 mm 1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

Rcenter

3
1800.0 mm x 800.0
1800.0
mmmm x 400.0 mm
( x3 , y3 ) d3 Rcenter
wall3 w3

Y
Y
Sensors 2014, 14 2098

3.2. Calculation of Distance d i with IBDMS [24–29]

IBDMS is developed in this paper for the purpose of calculating the distances di (i  1,2,3) , which
can work on a single webcam and only depends on a known-dimension rectangle, i.e., a ground tile.
The idea of IBDMS is from the triangular relationship, shown in Figure 6, that is we first capture an
image incorporating a known-dimension rectangle, and then the proportion relationship between the
real-dimension and the image-dimension of the rectangle can be found. According to the proportion
relationship, the distance d i can then be easily calculated. Figure 6 shows the IBDMS set-up. It only
requires a webcam and two given-location points A and B , which could be two corners of a ground
tile. h s is a constant parameter of the webcam. O is the intersection of the optical axis and the plane.
The targeted objects which lie on the plane and are perpendicular to the optical axis can be measured
by simple trigonometric function derivations. Hence, ho can expressed as:
DN
ho  N H maxCot H  hs (18)
N ( A)  N ( B)

where ho can be considered as any of distances di (i  1, 2, 3) for the ith webcam, N ( A) and N ( B)
are pixel values in the captured image. N H (max) is the maximal pixel width in a horizontal scan line
of the image. DN is the width between points A and B .  H is the horizontal view angle.

Figure 6. Schematic diagram of the IBDMS.

N ( P)

N ( B) N ( A)

N H (max)

hs

H
ho

O
P A B Q
DN

3.3. Calculation of Distance wi with PLDMS

Inheriting the concept of the IBDMS, a parallel-line distance measurement system (PLDMS) for
measuring the distances wi is developed. Figure 7 shows the schematic diagram of the PLDMS. In
PLDMS, four points, Ls1i , Ls2i , Ls3i , and Ls4i , are considered as the reference points. In Figure 7, we
draw a pair of parallel lines ( Ls1i Ls2i and Ls3i Ls4i ) through these points. DN is the width between
Sensors 2014, 14 2099

Ls1i Ls2i and Ls3i Ls4i . Furthermore, in the image-domain, as Figure 7, the linear proportion of the line
( Ls3i Ls4i ) and the line ( Ls1i Ls2i ) can be defined as:
y i  y1i xi  x1i (19)
Ls1i Ls2i : 
y2i  y1i x2i  x1i
y i  y3i xi  x3i (20)
Ls3i Ls4i : 
y4i  y3i x4i  x3i

and:
N H ( PQ)  Ls1i Ls2i ( y i )lPQ  Ls3i Ls4i ( y i )lPQ (21)

where ( x j , y j ) ( j  1,2,3,4) are, respectively, the image coordinate of the points Ls j , and the pixel
i i i

( xi , yi )(i  1, 2,3) is any image points laying at the line Ls1i Ls2i or Ls3i Ls4i . In Figure 7, N H ( PQ) is the
number of pixels between P and Q, which can be obtained by Equation (21).

Figure 7. Diagram of the PLDMS.

Webcam

Ls3i

Q
Ls4i

S WSR

Ls1i R
WRS

DN
WPR

Ls2i

In Figure 8, points R and S could be two corners of a known-dimension rectangle, i.e., a ground
tile, in a captured image (image-domain), and P and Q are the cross points between the scan line lPQ
and the lines Ls1i Ls2i and Ls3i Ls4i .For the reason that P , Q , R , and S are laying at the same scan
line lPQ , we can easily calculate the points P and Q by Equations (19) and (20), where R , S and ,
( xij , y ij ) ( j  1,2,3,4) have been given by the image processes. In our experiments, the width between
R and S means the width of the robot. Furthermore, the width WRS can be calculated by the proportion
relationship, which is defined as:
N ( RS ) (22)
WRS  DN  H
N H ( PQ )
where N H ( RS ) and N H ( PQ) can be obtained by the PLDMS. With the same idea, we can calculate
the width WPR or the width WSQ .
Sensors 2014, 14 2100

Figure 8. Schematic diagram of the PLDMS.

N
Nyy

N H (TU )
NNvV_max
max

lTU K
T U
i
Ls4i4
Ls Ls2
Ls
NV ( JK ) 2

ISC

lPQ
P R J S Q
Lsi3
Ls i
Ls11
Ls
3
N H ( RS )

NNxx
0 N H ( PQ) NNHH_max
max

3.4. Overall Procedures of the Proposed Localization Method

Figure 9 shows the overall scheme of the proposed robot localization method, where t s is the
update time-interval for the background image. The moment of updating the background image is that
if the segmentation horizontal projection of the binary foreground image f fg ( x, y) is bigger than the
threshold value t1 , the captured image fcurrent ( x, y) is set to a new background image; otherwise, we
keep the original background image. In Equation (23), the function s( ) means the process of the
segmentation horizontal projection, as shown in Figure 10.

 f current if s( f fg ( x, y ))  t1
 (23)
fbg ( x, y )  

 f current ( x, y ) otherwise,
In the image processing, the coordinates of the robot can be calculated through the methods of
gray-scale transformation, background subtraction, binarization, morphological processing, connected
components labeling, and averaging method. The background image is updated in every t s seconds
with the update Equation (23). After image processing, the coordinate in the image-domain goes
through the IBDMS for calculating the distance d s (Equation (18)) and the PLDMS for calculating the
width wi (Equation (22)). Then, we can locate the robot in the real-world domain.

4. Experimental Results

4.1. Set-Up of the Experimental Environment

Figure 11 shows the map of our experimental environment. In this map three webcams, which are
explained in Section 2.1, are used to cover the possible working area of the robot. The coordinates
(x1; y1) of the first webcam are (279; 570); (x2; y2) are (755; 465); and (x3; y3) are (650; 1,030). Because
of limitation of the USB 2.0 transmission speed, the resolution of the first webcam is 640  480; the
second webcam is 640  480; and the third webcam is 480  360.The aisle widths are shown in
Sensors 2014, 14 2101

Figure 11, and are 138, 150 and 152 cm, respectively. The background image is updated in every 5 s.
Threshold value t1 of the updated background image is set to 19.5.

Figure 9. Overall scheme of the proposed localization method.

Capturing image (current) Background image

Sample in every ts
second

Gray Scale Transformation

Background Subtraction

Binarization

At the first Yes Create Parallel

Morphological Processing
image Lines
No
Connected Components
Labeling IBDMS for di

Average Method
PLDMS for wi
Image Processing Robot Localization
Coordinate of the Robot

Figure 10. Update of the background image by using segmentation horizontal projection.
35

30
Threshold value
25

20
Pixels

15

10

0
0 50 100 150 200 250 300
x-axis

Figure 11. Map of our experimental environment.

150 cm
1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

nd
2 webcam
wall2
(x2 , y2 )
1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

1st webcam
( x1, y1 ) 138 cm
wall1

1800.0 mm x 800.0 mm 1200.0 mm x 800.0 mm 1200.0 mm x 800.0 mm

3rd webcam
1800.0 mm x 800.0
1800.0
mmmm x 400.0 mm
( x3 , y3 ) 152 cm
wall3

Y
Sensors 2014, 14 2102

4.2. IBDMS and PLDMS Set-Up

Some basic set-up steps must be performed for the IBDMS and PLDMS before we start the robot
localization procedures. In order to alleviate the impact from the image noise, building the first
background image is done taking the average of 150 consecutive images, and then the background
subtraction method can much more effectively extract the foreground image. Besides, w low-pass filter
is adopted to further refine the background image. The obtained background images are shown in
Figure 12.

Figure 12. (a) Background image of the 1st webcam. (b) Background image of the 2nd
webcam. (c) Background image of the 3rdwebcam.

(a) (b) (c)

In the 1st webcam, as shown in Figures 13 and 14, four corners of the known-dimension ground tile
are used to draw a pair of the virtual parallel line ( Ls11Ls21 and Ls31Ls41 ), whose deriving linear
equations can be expressed as:

Ls11Ls21 : y11  6.0625x11  865.063 (24)

and:
Ls31Ls41 : y21  5.68 x21  992 (25)

where the coordinates of Ls11 , Ls21 , Ls31 , and Ls41 are respectively chosen as (113, 180), (129, 83),
(112, 181), and (128, 97) in the image-domain.

Figure 13. Two chosen points for the right line in the 1st webcam.

Ls11

Ls12
Sensors 2014, 14 2103

Figure 14. Two chosen points for the left line in the 1st webcam.

Ls31

Ls14

Similar to the setting procedures of the 1st webcam, the virtual parallel line ( Ls12 Ls22 and Ls32 Ls42 )
for the 2nd webcam can be expressed as:
Ls12 Ls22 : y12  8.107 x12  1500 (26)

and:
Ls32 Ls42 : y22  4.551x22  1479.61 (27)

where the coordinates of Ls12 , Ls22 , Ls32 , and Ls42 are respectively chosen as (140, 365), (168, 138),
(404, 359), and (355, 136) in the image-domain. Figures 15 and 16 show these four points.

Figure 15. Two chosen points for the right line in the 2nd webcam.

Ls12

Ls22

Figure 16. Two chosen points for the left line in the 2nd webcam.

Ls32

Ls42

The virtual parallel line ( Ls13 Ls23 and Ls33 Ls43 ) for the 3rd webcam can be expressed as:

Ls13Ls23 : y13  5.475x13  1309.65 (28)

Sensors 2014, 14 2104

and:
Ls33Ls43 : y23  5.815x23  2055.55 (29)
where the coordinates of Ls13 , Ls23 , Ls33 , and Ls43 are respectively chosen as (174, 357), (214, 138),
(415, 358), and (377, 137) in the image-domain. Figures 17 and 18 show these four points.

Figure 17. Two chosen points for the right line in the 3rd webcam.

Ls13

Ls23

Figure 18. Two chosen points for the left line in the 3rd webcam.

Ls33

Ls43

4.3. Experimental Results

In our experiments, a remote-controlled track-robot moves through the monitored areas. The path of
the robot is shown in Figure 19a. In Figure 19a, the circled locations, causing bigger errors as shown in
Figure 19b, are the coordinates of the front arms of the robot at the moment of which the robot is
moving into the covering area of the 2nd webcam, as shown in Figure 19c. In Figure 19d, a bigger
error happens in the circled locations, which are the coordinates of the front arms of the mobile robot
at the moment of which the robot is moving into the covering area of the 3rd webcam, as shown in
Figure 19e. The error function used to show the ability of our proposed localization method is:

Error  ( x2  x1 )2＋( y2  y1 )2 (30)

where ( x1 , y1 ) are the actual coordinates, and ( x2 , y2 ) are the coordinates measured by the proposed
method. In Table 1, the measurement errors range from 2.24 cm (when the robot is near the webcams)
to 12.37 cm (when the robot is far away from the webcam). According to the definition of Rcenter and
the dimensions of the robot, which are 54 cm × 54 cm, a common size for patrolling robots, we find
Sensors 2014, 14 2105

that the robot can be correctly located even though it is far away from the webcams. Limited by the
low-resolution webcams, the measurement errors are acceptable. We also can easily reduce the
measurement errors by using high-resolution CCD camera. Furthermore, the selection of a
known-dimension rectangle pattern should be clearly seen in the captured image in order to set up the
reference points of IBDMS and PLDMS. In addition, a suitable threshold value t1 in the segmentation
horizontal projection, which is used to update the background image, and the distortion coefficient
vectors and camera conversion matrix in image calibration are important factors for precisely locating
the mobile robots.
Under the conditions of the static monitored area, it is assumed that light sources and locations of
walls and furniture are given. Light influence, therefore, can be easily attenuated through choosing
appropriate factors in the image processing techniques. In this paper, we pay more attention to locating
the moving robot by using single webcam and have not yet considered the situation of partial
occlusions. In this static indoor environment, some image techniques [33–35] could be used to
overcome temporary partial occlusion.

Figure 19. Moving path of the track robot (a) Path of the robot; (b) Path captured under
the Webcam II (c) Image of the Webcam II; (d) Path captured under the Webcam III;
(e) Image of the Webcam III
600
xX

Webcam II
(755,465)

Webcam I
600
(279,570)

Webcam III
(650,1030)

y
Y
(a)

(b) (c)
Sensors 2014, 14 2106

Figure 19. Cont.

(d) (e)

Table 1. Measurement errors of the proposed method.

Actual Coordinate Measured Coordinate Error (cm) Webcams
(610,592) (613,588) 5.00 Webcam I
(627,611) (626,609) 2.24 Webcam I
(636,560) (630,559) 8.08 Webcam I
(648,591) (641,594) 7.62 Webcam I
(684,584) (676,578) 10.00 Webcam I
(736,891) (738,883) 8.24 Webcam II
(774,912) (775,901) 11.04 Webcam II
(759,918) (756,906) 12.37 Webcam II
(763,944) (759,932) 12.65 Webcam II
(808,966) (811,952) 12.37 Webcam II
(1001,779) (997,779) 4.00 Webcam III
(1236,1076) (1225,1074) 11.18 Webcam III
(1250,1019) (1241,1021) 9.49 Webcam III
(1271,1063) (1262,1065) 11.18 Webcam III
(1250,1041) (1241,1039) 9.22 Webcam III

5. Conclusions

This paper proposes the use of IBDMS and PLDMS to locate a mobile robot in an indoor
environment. Through the image processing and according to a known-dimension ground tile, the
IBMDS and PLDMS used can calculate the coordinates of a moving tracked robot. Using this
framework, we can quickly estimate the localization of the tracked robot. Furthermore, the
experimental environment is easy to set up since only three parameters have to be defined, that is, the
maximum pixel, the perspective, and the optical distance. Because the locations of webcams are fixed,
we can utilize a simple background subtraction method to extract the data to attenuate the problem of
computational burden. In addition, we use a low-pass filter and an on-line background updating
method to reduce background noise, and we adopt the image morphology to acquire the robot’s image
information. This method does not use expensive high-resolution webcams and complex pattern
Sensors 2014, 14 2107

recognition methods to identify the mobile robot, but rather just uses a simple formula to estimate
distance. From the experimental results, the localization method is both reliable and effective.

Acknowledgment

This work was supported by the National Science Council, Taiwan, under Grants NSC 102-2221-E-
003-009, NSC 102-2221-E-234-001-, and NSC 102-2811-E-011-019.

Author Contributions

The work presented here was carried out in collaboration between all authors. I.-H. L. and M.-C. C.
defined the research theme. I.-H. L. and W.-Y W. designed methods and experiments, carried out the
laboratory experiments, analyzed the data, interpreted the results and wrote the paper. S.-F. S.
co-worked on associated data collection and their interpretation. T.-W. L co-designed experiments,
discussed analyses, interpretation, and presentation. All authors have contributed to, seen and approved
the manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Kim, J.B.; Jun, H.S. Vision-based location positioning using augmented reality for indoor
navigation. IEEE Trans. Consum. Electr. 2008, 54, 954–962.
2. Lin, C.H.; King, C.T. Sensor-deployment strategies for indoor robot navigation. IEEE Trans. Syst.
Man Cybernet. 2010, 40, 388–398.
3. Hinckley, K.; Sinclair, M. Touch-Sensing Input Devices. In Proceedings of the Conference on
Human Factors in Computing Systems, Pittsburgh, PA, USA, 15–20 May 1999.
4. Partridge, K.; Arnstein, L.; Borriello, G.; Whitted, T. Fast Intrabody Signaling. Demonstration at
Wireless and Mobile Computer Systems and Applications, Monterey, CA, USA, 7–8 December
2000.
5. Hills, A. Wireless andrew. IEEE Spectr. 1999, 36, 49–53.
6. Want, R.; Hopper, A.; Falcao, V.; Gibbons, J. The active badge location system. ACM Trans.
Inform. Syst. 1992, 10, 91–102.
7. Bahl, P.; Padmanabhan, V. RADAR: An In-Building RF Based User Location and Tracking
System. In Proceedings of Nineteenth Annual Joint Conference of the IEEE Computer and
Communications Societies, Tel Aviv, Israel, 26–30 March 2000.
8. Starner, T.; Schiele, B.; Pentland, A. Visual Context Awareness via Wearable Computing. In
Proceedings of the International Symposium on Wearable Computers, Pittsburgh, PA, USA, 19–20
October 1998; pp. 50–57.
9. Ni, L.M.; Lau, Y.C.; Patil, A.P. LANDMARC: Indoor location sensing using active RFID. Wirel.
Netw. 2004, 10, 701–710.
Sensors 2014, 14 2108

10. Subramanian, V.; Burks, T.F.; Arroyo, A.A. Development of machine vision and laser radar based
autonomous vehicle guidance systems for citrus grove navigation. Comput. Electr. Agric. 2006,
53, 130–143.
11. Barawid, O.C., Jr.; Mizushima, A.; Ishii, K. Development of an autonomous navigation system
using a two-dimensional laser scanner in an orchard application. Biosyst. Eng. 2007, 96, 139–149.
12. Sato, T.; Sugimoto, M.; Hashizume, H. An extension method ofphase accordance method for
accurate ultrasonic localization of movingnode. IEICE Trans. Fundament. Electr. Commun.
Comput. Sci. 2010, J92-A, 953–963.
13. Carmer, D.C.; Peterson, L.M. Laser radar in robotics. Proc. IEEE 1996, 84, 299–320.
14. Thrun, S.; Burgard, W.; Fox, D. Probabilistic Robotics; MIT Press: Cambridge, UK, 2005.
15. Hightower, J.; Want, R.; Borriello, G. SpotON: An Indoor 3D Location Sensing Technology
Based on RF Signal Strength; UW-CSE00-02-02, University of Washington: Seattle, WA, USA,
2000.
16. Lin, T.H.; Chu, H.H.; You, C.W. Energy-efficient boundary detection for RF-based localization
system. IEEE Trans. Mob. Comput. 2009, 8, 29–40.
17. Lorincz, K.; Welsh, M. MoteTrack: A robust, decentralized approach to RF-based location
tracking. Pers. Ubiq. Comput. 2007, 11, 489–503.
18. Dellaert, F.; Fox, D.; Burgard, W.; Thrun, S. Monte Carlo Localization for Mobile Robots. In
Proceedings of the 1999 IEEE International Conference on Robotics and Automation, Detroit, MI,
USA, 10–15 May 1999; pp. 1322–1328.
19. Thrun, S.; Bücken, A.; Burgard, W.; Fox, D.; Fröhlinghaus, T.; Henning, D.; Hofmann, T.;
Krell, M.; Schmidt, T. Map Learning and High-Speed Navigation in RHINO; MIT Press:
Cambridge, MA, USA, 1998.
20. Li, S.; Liu, B.; Chen, B.; Lou, Y. Lou neural network based mobile phone localization using
bluetooth connectivity. Neural Comput. Appl. 2013, 23, 667–675.
21. Li, S.; Qin, F. A dynamic neural network approach for solving nonlinear inequalities defined on a
graph and its application to distributed. Neurocomputing 2013, 117, 72–80.
22. Martins, M.H.T.; Chen, H.; Sezaki, K. Otmcl: Orientation Tracking-Based Monte Carlo
Localization for Mobile Sensor Networks. In Proceedings of the Sixth International Conference
on Networked Sensing Systems, Pittsburgh, PA, USA, 17–19 June 2009; pp. 1–8.
23. Cheok, A.D.; Yue, L. A novel light-sensor-based information transmission system for indoor
positioning and navigation. IEEE Trans. Instrum. Meas. 2011, 60, 290–299.
24. Hsu, C.C.; Lu, M.C.; Wang, W.Y.; Lu, Y.Y. Three dimensional measurement of distant objects
based on laser-projected CCD images. IET Sci. Meas. Technol. 2009, 3, 197–207.
25. Mertzios, B.G.; Tsirikolias, I.S. Applications of Coordinate Logic Filters in Image Analysis and
Pattern Recognition. In Proceedings of the 2nd International Symposium on Image and Signal
Processing and Analysis, Pula, Croatia, 19–21 June 2001; pp. 125–130.
26. Lu, M.C.; Wang, W.Y.; Chu, C.Y. Image-based distance and area measuring systems. IEEE Sens.
J. 2006, 6, 495–503.
27. Wang, W.Y.; Lu, M.C.; Kao, H.L.; Chu, C.Y. Nighttime Vehicle Distance Measuring Systems.
IEEE Trans. Circ. Syst. II: Exp. Brief. 2007, 54, 81–85.
Sensors 2014, 14 2109

28. Wang, W.Y.; Lu, M.C.; Wang, T.H.; Lu, Y.Y. Electro-optical distance and area measuring
method. WSEAS Trans. Syst. 2007, 6, 914–919.
29. Hus, C.C.; Lu, M.C.; Wang, W.Y.; Lu, Y.Y. Distance measurement based in pixel variation of
CCD image. ISA Trans. 2009, 40, 389–395.
30. Osugi, K.; Miyauchi, K.; Furui, N.; Miyakoshi, H. Development of the scanning laser radar for
ACC system. JSAE Rev. 1999, 20, 549–554.
31. Nakahira, K.; Kodama, T.; Morita, S.; Okuma, S. Distance measurement by an ultrasonic system
based on a digital polarity correlator. IEEE Trans. Instrum. Meas. 2001, 50, 1478–1752.
32. Kawaji, H.; Hatada, K.; Yamasaki, T.; Aizawa, K. An Image-Based Indoor Positioning for Digital
Museum Applications. In Proceedings of the 16th International Conference on Virtual Systems
and Multimedia (VSMM), Seoul, Korea, 20–23 October 2010.
33. Cucchiara, R.; Grana, C.; Tardini, G. Probabilistic People Tracking for Occlusion Handling. In
Proceedings of the 17th International Conference on Pattern Recognition, Cambridge, UK, 23–26
August 2004; pp. 132–135.
34. Eng, H.L.; Wang, J.X.; Kam, A.H. A Bayesian Framework for Robust Human Detection and
Occlusion Handling Human Shape Model. In Proceedings of the International Conference on
Pattern Recognition, Cambridge, UK, 23–26 August 2004, pp. 257–260.
35. Tsai, C.Y.; Song, K.T. Dynamic visual tracking control of a mobile robot with image noise and
occlusion robustness. Image Vis. Comput. 2007, 27, 1007–1022.

© 2014 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article
distributed under the terms and conditions of the Creative Commons Attribution license
(http://creativecommons.org/licenses/by/3.0/).