The Technology Research of Camera Calibration Based On LabVIEW

International Journal of Research in Engineering and Science (IJRES)
ISSN (Online): 2320-9364, ISSN (Print): 2320-9356
www.ijres.org Volume 3 Issue 8 ǁ August. 2015 ǁ PP.08-17
www.ijres.org 8 | Page
The Technology Research of Camera Calibration Based On
LabVIEW
Xiuqin-Li, Xianyang-Du, Yawei-Li,
(College of Mechanical Engineering, Shanghai University of Engineering Science, Shanghai 201620, P.R.China)
ABSTRACT: The technology of camera calibration is most important part for machine vision detection and
location, the accuracy of calibration directly determines the processing accuracy of machine vision systems. In
this paper, we use LabVIEW and MATLAB to calibrate the internal and external parameters of the camera, at
the same time, we use dot calibration board, the circle edge is detected by Canny operator, then with the method
of circle fitting based on subpixel edge extraction, the information of dots image coordinate is extracted. The
present method reduces the difficulty of camera calibration and shortens the software development cycle, the
most important is that it has a high calibration accuracy, which can meet the actual industrial detection accuracy,
the results of experimental show that the method is feasible.
Keywords- camera calibration, machine vision, LabVIEW, MATLAB, Canny
I. INTRODUCTION
Machine vision is to use the computer and the visual sensor to replace the human eye to do the
measurement and judgment [1]
, the technology of machine vision is more and more widely used in the field of
industrial inspection, automotive manufacturing and medical image analysis, because of its flexible, rapid and
non-contact characteristics, which can improve the intelligence of the industrial field.
Camera calibration is the most important part of machine vision applications, its purpose is through obtain
the information of the acquisition of the two-dimensional image to get the information of the three-dimensional
coordinates[2]
. The camera calibration technology has been quite mature both at home and abroad, which can be
divided into two categories: traditional calibration techniques and self-calibration techniques [3]
. The
disadvantages of the traditional calibration method are not flexible enough, the self-calibration of the camera
can only be achieved by using the camera's corresponding relationship between the image and the image in the
moving process, which does not depend on the reference object, the calibration accuracy and robustness of
self-calibration method are poor compared to traditional calibration methods.
The calibration method of Zhang Zhengyou is a kind of method between the traditional calibration and
self-calibration, its technique is flexible, at the same time, and the calibration accuracy is high, which is a
common method used in camera calibration. In this paper, we achieve the camera calibration with the help of
Zhang Zhengyou calibration method, LabVIEW and MATLAB tools.
II. LabVIEW Development Platform
LabVIEW is a graphical programming language (G language) introduced by NI, it can be able to provide a
graphics programming, which is simple, intuitive and easy to users, what’s more , it can save more than 85%
of the program development time, compared with the traditional programming language, while the running
speed is almost unaffected, which makes high efficiency. At the same time, it provides a wide range of
interfaces and has great flexibility, which can be called with a variety of software such as DLL, Visual Basic
and MATLAB.

The Technology Research of Camera Calibration Based on Lab VIEW
LabVIEW Vision IMAQ provides a good platform for the development of machine vision, it contains
various function library of image processing functions and integrates more than 400 functions for image display,
processing, analysis, and other operations, which provides a complete development for the visual system [4-5]
, it
can further shorten the development cycle of the visual system combined with the use of LabVIEW, which is
widely accepted by industry, academia and research laboratories.
III. Mathematical Model of Camera Imaging
In the process of image acquisition of CCD camera, the geometrical position of the space is determined by
the model of camera imaging. At present, the most of camera imaging model is based on the principle of pinhole
imaging. This article is based on the pinhole model, considering the radial lens distortion to establish the model
of imaging camera.
3.1 Pinhole linear model of camera
),,( ccc ZYXP
o x
y
co
cX
cY
O
cZ
u
v
 ,u u uP x y
( , , )w w wX Y Z
Yw
Zw
Xw
Fig3.1 Schmatic model of camera imaging
f
(1) Image coordinate system ( , )u v
The image which is captured by CCD camera is stored in the computer with the form of an array. After the
CCD camera acquisition of the image is stored in the form of an array of computer. Among those, ( , )u v is the
image coordinate system unit of pixels, while the actual application needs to be converted it into a physical unit
of mm in the image coordinate system ( , )x y ，The relationship between the two coordinate systems is shown in
figure 3.2:
O u
v
o
x
y
0 0( , )u v
Fig. 3.2 Iamge coordinate system

The origin of o is defined in the intersection of the optical axis of the camera and the imaging plane, which
is generally located in the center of the image. If the o’s coordinates is 0 0( , )u v , which is in the coordinate
system ( , )u v , the physical size of each pixel in the x-axis and y-axis direction is ,dx dy , then the any point in
the two coordinates ( , )x y and ( , )u v have the following relations, as equation( 3-1):
0
0
1
0
1
0 (3 1)
1 1
0 0 1
u
dxu x
v v y
dy
 
 
    
         
       
 
  
(2)Camera coordinate system ),,( ccc ZYX
Select the optical center co of CCD camera lens as the origin of the camera coordinate system, the axis of
cX and cY is parallel to the axis of x and y , which is the image physical coordinate system, cZ is the
direction of camera’s optical axis, and the intersection of the imaging plane is 0 0( , )o u v . The distance of
co o is the effective focal length of the camera, which is f, according to pinhole imaging model, the position of
the spatial points on the image can be expressed as equation(3-2):
(3 2)
fXc
x
Zc
fYc
y
Zc



 

Written the 2-2 in the type of matrix form is as follow equation(3-3):
0 0 0
0 0 0 (3 3)
1 0 0 1 0
1
Xc
x f
Yc
Zc y f
Zc
 
    
                
 
According to the equation of 3-1and 3-3, we can establish the relationship between the image coordinate system
and the camera coordinate system, as shown in follow equation3-4:
0
0
0 0 1
1
0
0 00 0 0
1
0 0 0 0 0 0 (3 4)
1 0 0 1 0 0 0 1 0
0 0 1
x
y
u
dx f uu f Xc Xc Xc
Zc v v f Yc f v Yc M Yc
dy
Zc Zc Zc
 
 
           
                         
                     
 
  
Among them, 1M is the matrix of 3×4, which is the internal parameter matrix of camera. By collecting at
least three images of calibration, we can obtain the inner parameter matrix of 1M , according to the establish
equation of 3-4.
(3) World coordinate system
( , , )w w wX Y Z
The world coordinate system can be set at any position, which is to describe the position of the camera.
The position relationship between the world coordinate system and the camera coordinate system can be

determined by the rotation matrix of R and translation vector of T, the coordinates of space point P in world
coordinate system and camera coordinate system are respectively ( , , )w w wX Y Z and ),,( ccc ZYX . Can be
expressed as a matrix:
2 (3 5)
0 1
1 1 1
Xc Xw Xw
Yc R t Yw Yw
M
Zc Zw Zw
     
     
               
     
     
Among them, R is an orthogonal unit matrix of 3 ×3, which is called rotation matrix, t is a column vector
of 3×1, which is to express the translation between the two coordinates system, 2M is matrix of 4×4, which is
the external parameter matrix of camera, and it is related to the position of the camera in the world coordinate
system. According to the equation of 3-3, 3-4 and 3-5, we can get the relationship between the image coordinate
system and the world coordinate system, which is shown in the following formula 3-6:
0
0 1 2
1
0
0 0 0
1
0 0 0 0 (3 6)
0 1
1 0 0 1 0
1 1 10 0 1
u Xw Xw Xwdxu f
R t Yw Yw Yw
Zc v v f M M H
Zw Zw Zwdy
 
       
          
                                       
       
  
Among them, H is a single stress matrix, which describes the position of the spatial points and the image
plane, by means of a sufficient number of known points of the world coordinates and the corresponding image
coordinates, according to the 3-6 formula, we can solve the single stress matrix of H, then in the case of Hand
the internal parameter matrix of 1M defined, we can get the external parameter matrix of 2M ,from the above
process, we can complete the linear model of the camera calibration.
3.2 Nonlinear model of camera
Since there will be some errors in the process of manufacturing and assembly of the camera lens, so in the
practical applications, we should take the camera lens distortion into account. There are three types of lens
distortion, which are radial distortion, eccentric distortion and thin prism distortion [6-7]
. The practice shows that
the radial distortion has the most influence on the imaging model of the camera, so sometimes, we only consider
the radial distortion, if we consider too much distortion, it will increase the number of nonlinear parameters,
which will lead to instability in solving results. In this paper, in the condition of the radial distortion, we
establish the nonlinear model of the camera, the radial distortion equation of Zhang zhenyou is shown as
follows:
2 4
1 2(1 ) (3 7)
d u
d u
x x
k k
y y
 
   
      
   
 ,d dx y is the coordinates of ideal imaging points, ( , )u ux y is the actual imaging points of the coordinates,
which is considered as the distortion, among them,
2 2 2
u ux y   . Finally, we use the nonlinear least
squares algorithm to solve the camera's parameters matrix of internal and external and distortion factor, take
these parameters as the results of the camera calibration.

IV. Calibration Process
The calibration of the camera internal parameters is accomplished by using the Vision module of LabVIEW.
After installing the Vision module of Development Module 2012 and NI Vision Builder for Automated
Inspection 2012, it has added the independent calibration module of Calibration Training, which makes the
calibration work more convenient and efficient[8]
. The calibration process can be divided into three steps, the
first is to study the template, followed by the calibration, and the last is to read and store the calibration
information. There are two kinds of calibration template, the one is the board square, and another one is the
circular calibration target,
The extraction algorithm of the center of circle’s stability and positioning accuracy is superior to the
checkerboard calibration plate extraction of corner, so in this paper, we use calibration template of 6 x 8 dot
array, as shown in figure 4.1
Fig.4.1 6×8 dot array calibration board
Sticking the dot array to a smooth surface calibration board, the distance of center between the two adjacent
dots of horizontal spacing and vertical spacing respectively is 24mm.Calibration algorithm flow is shown in
Figure4.2.
Canny filter to extract
the circular contour
Circle fitting realize
the positioning
Input calibration
picture information
Solving camera
intrinsic parameters
Nonlinear optimization
solution
Load calibration
image
Solving camera
external parameters
Storage
Calibration
information
Complete the
calibration
Fig. 4.2 Camera calibration process

4.1 Camera Internal Parameters Calibration Process
(1) In order to correct the error caused by the camera's nonlinear distortion and obtain the internal parameters of
the camera, in this paper, we select the type of camera calibration, the Molel Camera (Grid) type.
Fig.4.3 The First Step of Camera Calibration
(2) To load the pictures of the calibration plate .In this paper, we load the 8 calibration pictures, which are taken
from the different angles of the camera.
Fig.4.4 The Second Step of Camera Calibration
(3) To use Canny filter to extract the circular edge of the contour.
Fig.4.5 The Third Step of Camera Calibration

(4) To input the horizontal and vertical distance of the two adjacent center of the calibration board and
measurement unit.
Fig4.6 The Forth Step of Camera Calibration
(5) To extract feature point coordinates by using circle fitting method, calculate the internal parameters of the
camera and the camera calibration error.
Fig4.7 The Fifth Step of Camera Calibration
(6)We establish a coordinate system to calibrate the pictures, which facilitates the transformation of coordinate
next step.
Fig4.8 The Sixth Step of Camera Calibration

(7)Calibration template learning, the calibration of the picture is saved to the PNG file format, which is easy to
read and call the calibration template. As long as the relative position of the camera and the working plane is
constant, it is not to be repeated. The internal parameters of the camera can be solved by the above calibration
procedure:
1
5246.25 0 1247.889
0 5200.63 955.033
0 0 1
M
 
   
  
The physical size of the camera pixel is 0.0022mm * 0.0022mm, and we can get the effective focal length
of the camera according to the internal parameter matrix which has been already obtained.
(5246.25 0.0022 5200.63 0.0022) 2 11.492f      
4.2 Camera external parameters calibration process
The camera external parameters calibration can be achieved by the MATLAB script node provided by
LabVIEW, which can improve the processing speed. Complete the internal parameters of the camera calibration
process, then we call calibration pictures, read the calibration dots coordinate information and serve the image
coordinates ( , )u v and the world coordinate ( , )x y as the MATLAB script input and external parameter matrix
as the output, and the program flow chart is shown in Figure4.9:
Fig.4.9 The solution Process of External Camera Parameter Matrix
Because the external camera parameter matrix is related to the position where camera is located in world
coordinates, so 8 different angle calibration images have different parameter matrix, it means that we should
determine it according to the specific circumstances.
V. The Experiment Result Analysis
After calibration, we can see calibration information. The table 5-1 is calibration of 6×8 dot array
calibration parameters.
Table 5-1 The dot array calibration parameters (the average error is 0.0329 mm, the distortion is 0.1007%)
Object Radius
(pixels)
Radius
(mm)
Center u
(pixels)
Center v
(pixels)
Center X
(mm)
Center Y
(mm)
1 38.62 4.8177 797.42 1175.7 0 120.0399
2 38.84 4.8418 990.58 1176.5 23.9907 120.0275
3 38.34 4.7912 798.83 792.49 -0.01 72.0019
4 38.71 4.829 991.32 984.38 24.0056 95.9729
5 38.55 4.8233 800.48 410.68 0.0305 24.05
6 38.95 4.8505 1184.25 1177.6 48.0221 120.0518

7 38.73 4.8292 1378.9 602.52 71.9943 47.9524
8 38.59 4.8025 1766.64 412.68 119.9734 24.0654
9 38.73 4.8364 1186.28 410.64 48.026 23.9388
10 38.63 4.8052 989.65 1559.67 24.0346 167.9068
11 38.65 4.8294 992.59 601.57 24.0156 47.9643
12 39.01 4.8521 1377.62 1178.48 71.9842 120.0471
13 38.74 4.8276 1572.51 412.01 95.9459 24.0206
14 38.82 4.8229 1571.61 1179.27 95.9874 120.029
15 38.86 4.8515 991.83 792.9 23.9934 71.9709
16 38.83 4.838 1378.7 794.09 72.02 71.9566
17 38.67 4.8129 1572.42 794.85 96.0157 71.9721
18 38.88 4.8599 993.46 410.88 24.0526 24.0195
19 38.43 4.798 798.22 983.74 0.003 95.9902
20 38.81 4.8224 1377.14 1370.58 71.9732 144.0363
21 38.77 4.8325 1185.01 985.04 48.0531 95.9574
22 38.78 4.8126 1766.32 1180.25 120.0372 120.0334
23 38.47 4.7896 1766.49 221.62 119.9199 0.131
24 38.54 4.8237 801.31 219.82 0.0563 0.0557
25 38.72 4.8207 1378.15 985.85 72.0009 95.9609
26 38.7 4.798 1570.65 1563.54 95.9297 167.943
27 38.48 4.7859 1766.89 604.04 120.0364 48.023
28 38.61 4.7856 1766.01 1372.57 120.0152 144.0104
29 38.32 4.7911 799.53 601.37 -0.01 48.0083
30 38.78 4.8133 1571.21 1371.5 95.9696 144.016
31 38.81 4.8437 1185.84 601.63 48.0321 47.9051
32 38.65 4.8241 1379.45 219.41 71.9581 -0.1
33 38.61 4.8288 994.01 219.92 24.0514 0.0264
34 38.55 4.7919 1766.83 795.67 120.0562 71.9947
35 38.56 4.7953 1572.08 986.61 96.0103 95.9581
36 38.57 4.8017 796.54 1558.18 0.0765 167.8593
37 38.77 4.8233 1183.79 1369.47 48.0282 144.0292
38 38.7 4.8195 990.09 1368.24 24.0084 144.0041
39 38.87 4.8499 1379.13 411.08 71.9719 23.9463
40 38.56 4.774 1765.47 1564.76 119.9616 167.9368
41 38.46 4.7928 796.96 1367.12 0.0324 143.9901
42 38.83 4.8196 1376.55 1562.34 71.9464 167.946
43 38.5 4.7818 1766.78 987.34 120.0743 95.9494
44 38.48 4.8085 1187.12 219.7 48.0689 -0.04
45 38.55 4.8005 1572.55 603.23 95.9929 47.98
46 38.46 4.7946 1572.5 221.11 95.9021 0.0861
47 38.87 4.848 1185.39 793.47 48.0391 71.9594
48 38.73 4.8126 1183.19 1561.09 48.0177 167.9397

In metrology, the accuracy of the measured value is evaluated by using standard deviation. In the paper, we
use standard deviation to evaluate the calibration accuracy of the camera, which is shown by equation 5-1:
2
1
1
( ) (5 1)
n
i
i
x x
n


  
Among these,
ix is measurement value, x is the average value of a group of data, according to the
above measured dot array data information, let’s take the radius of the circle measured for an example, we can
calculate the value of a radius of standard difference, which unit in pixel is 0.1563pixel  and unit in mm
is 0.0215mm  , through the results of the calculation, we can see the camera calibration with high
precision, which can meet the requirements of the practical industrial measurement accuracy.
VI. CONCLUSION
In the paper, the calibration is based on the Zhang Zhengyou calibration method. First of all, we establish
the camera calibration model, then we use the method combining LabVIEW and MATLAB to complete the
intrinsic and extrinsic parameters of the camera calibration, the calibration method is fast and convenient, which
shortens the period of software development and improves the data processing ability, at the same time, it
reduces the camera calibration difficulty. The experimental results show that the calibration accuracy of the
calibration method is 0.0215mm, the average physical error is 0.0329mm, which meets the requirements of the
actual industrial measurement.
REFERENCE
[1] Carsten Steger, Markus Ulrich. Machine Vision Algorithms and Applications[M]. Berlin: Wiley-VCH, 2007.
[2] Shen Zunbing. Research on key technologies of machine vision image inspection and positioning system[D]. Harbin Institute of
Technology. 2009(9).
[3] Wang kai. The coplanar calibration of CCD and its application[D]. Jilin University.2009(5).
[4] National Instruments Corporation. IMAQ Vision for Lab⁃VIEW user manual[Z]. 2000.
[5] National Instruments. LabVIEW User Manual. 2005.
[6] Xiong ting. Research on Camera Calibration Technology of Vehicle Driving Wandering Test System[D]. Wuhan University of
Technology. 2014(5).
[7] Zhenyou Zhang. A Flexible New Technique for Camera Calibration [J]. IEEE TRANSACTIONS ON PATTERNAN ALYSIS
AND MACHINE INTELLIGENCE.2000.22 (11):1330-1334.
[8] Shi Kang, Ye Hong. A New Method for Machine Vision Calibration and Rectification Based on LabVIEW [J]. Laser &
Optoelectronics Progress. 2014(11):101501-101510.

The Technology Research of Camera Calibration Based On LabVIEW

More Related Content

What's hot (15)

Viewers also liked (20)

Similar to The Technology Research of Camera Calibration Based On LabVIEW (20)

More from IJRES Journal (20)

Recently uploaded (20)

The Technology Research of Camera Calibration Based On LabVIEW