WHITE PAPER [Link].

com

Color Calibration of Basler Cameras

Eye Sensitivity
This White Paper describes in detail what color is, 2.0
x (λ)
how it can be described in figures and how cameras
y (λ)
can be color-calibrated. z (λ)
1,5

The topic of color is relevant for many applications.

1.0
One example is post-print inspection. Today, thanks to
high technology, post-print inspection is an automated
process. Sophisticated inspection systems based on 0.5

cameras check things like labels and packaging. Often

this is combined with high demands for the quality of 0.0
color reproduction, for example, when food packaging 400 500 600 700

is checked, where the imaged foodstuff has to appear λ nm

“crisp” and “fresh” in order to increase the incentive to Figure 1: Color sensitivities of the three receptor types in the
buy. For the inspection systems for color-monitoring an human eye
optimized, standardized color representation is impor-
tant, so that different devices and installations produce
the same result when presented with the same color But how can colors be described so that they can be
information. For example, in the printing industry the handled in technical applications? In the course of time,
printed image is often checked on a monitor by a more than one answer has been found to this question.
human observer. To do this, the color on the monitor
has to be as realistic as possible in order to prevent In 1931 the International Commission for Illumination
incorrect decisions. (CIE = Commission Internationale de l’Éclairage)
defined the Yxy color space, in which all colors visible
This is only one example of many applications where to the human eye can be described. The commission is
high color accuracy is essential. an independent organization with the aim of interna-
tional cooperation for information exchange, for
example in illumination matters. Graphically, the Yxy
What is Color? color space is illustrated in the standard color chart for
the CIE color system in the shape of a shoe print. In this
When talking about a specific color it often happens color chart every color is represented by a point in a
that it is described differently by different people. The plane. The position of the point is described with the
color perception is created by the brain and the human variables x, y and Y. x represents the red/purple axis
eye together. For this reason, the subjective perception and y the green axis. The Y axis contains information
may vary. By effectively correcting the human sense of on the brightness of the color shade not included in the
sight by means of effective methods, this effect can standard color chart (the latter is a projection of the
even be observed for the same person. Physical color entire color space onto a plane).
sensations are created by electromagnetic waves of
wavelengths between 380 and 780 nm. In our eyes, One disadvantage of the Yxy color space is that it
these waves stimulate receptors for three different includes colors which cannot be perceived by the
color sensitivities. Their signals are processed in our human eye. This makes the color space somewhat
brains to form a color sensation. unwieldy. On the other hand all perceivable colors can
be represented in this color space.

1
 Advantages of the RGB color Space

In 1996 HP and Microsoft introduced the sRGB (stan-

dard RGB) color space. This is the most important
standard color space which can be deployed on just
about every monitor or printer and in every operating
system, browser and every software or internet presen-
tation. It can be realized relatively easily and is widely
used, however it is somewhat limited with regard to the
colors which can be represented.

If, in contrast, the much larger color space “wide gamut

RGB” (e.g. Adobe RGB) is used, the calibration of all of
the technical recording and representation equipment
is considerably more complex and susceptible to faults.

Figure 2: Standart color chart of the CIE color system

Different Color Spaces

For different technical applications typical alternative

color descriptions and methods are available, the
representation range of which can be illustrated in
various partial spaces in the Yxy color space. Examples
are the television color methods such as NTSC, PAL or
SECAM, the subtractive color model CMYK (cyan,
magenta, yellow, key) for color printing applications,
the wide gamut RGB and Adobe RGB for professional
color reproduction and sRGB (standard RGB) for CRT
monitors and the Internet. The described color spaces
for all of these methods can be indicated within the Yxy
color space as partial spaces and compared with each
other. Figure 3: Comparison of different color systems in the standart
color chart of the CIE color system
0.9

520

Advantages of a Calibrated Camera

0,8

540

Wide Gamut RGB

0,7

Adobe RGB If a calibrated camera is used, the colors on the sRGB

PAL/SECAN
560
monitor largely correspond to the colors as they
0,6

NTSC
appear in reality. With such a calibrated camera you
500 sRGB
Euroscale Coated CMYK
can measure colors and pass on standardized color
0,5
580
values. This is particularly important for the printing
3000 K industry. When using a calibrated camera the colors in
0,4 A 2000 K
D50
D55
D65
4000 K
1500 K 600 the image can be compared to the target values. This is
B
D75
D9300
C
E
1000 K
620 the precondition for achieving a color image true to the
0,3
10000 K
15000 K
640
700 original.
Unendlich

0,2

480
0,1

460
0,1 380 0,2 0,3 0,4 0,5 0,6 0,7 0,8

Figure 4: Position and size of the sRGB color space (Triangle)

within the standart color chart of the CIE color system

2
Different Color Systems

A color system is a specific representation of every

color within the color space. Despite the large number
of different color systems, some features are common
to all of them: In every color system at least three
values are needed for characterizing a specific color,
e.g. the values R (red), G (green) and B (blue) in the
sRGB system, or the values Y (yellow), M (magenta),
and C (cyan) in the YMC system. In other systems the
color is characterized by two values and the brightness
with the third value, e.g. in the YUV (luminance Y and
chrominance U and V) in the Yxy system described
above or similarly in the HSV (hue, saturation, value)
system.

Why are there different systems? The reason is that the

different systems are optimized for different aspects of
colorful images. The representation of the RGB color
space (fig. 5) is, for example, suitable for representing
the RGB values for conventional recording and display
devices. The representation in the HSV system (fig. 7), Figure 6: Chrominance Component of the YUV System
however, is informative, if one wishes to consider the
three parameters, brightness, saturation and color
contrast separately. The YUV model (fig. 6) is widely
used in television technology and offers advantages
with respect to the transmission of images and compat-
ibility to the older black-and-white receivers.

Figure 7: Representation of the HSV system

Figure 5: The representation of the RGB color space with R, G, and

B axis

3
Four Steps of Color Calibration Used by
Basler

A color camera uses several pixels with different color

filters in order to reproduce the color impression of the
real situation for every location in the image by means
of three color values. As, without correction, these
color values not only depend on the technical detail of
the camera but also on the illumination and other
factors, some effort is required before one is able to
reliably work with the color value determined by the
camera. In order to realistically record colors with a
camera, color calibration is required.

The powerful color calibration at Basler is a well

considered and standardized process consisting of four
steps: white balancing, gamma correction, matrix Color Calibration
correction and correction by the six-axis operator.
These four steps will be explained in more detail below.

The color calibration is carried out by recording as

unaltered as possible all of the 24 fields (18 colored and
6 grey fields) on the Greta Macbeth ColorChecker®, a
standard tool for the color calibration.

The target is that the camera records the colors as

precisely as the human eye

Figure 8: Greta Macbeth ColorChecker® , recorded with uncalibrated

camera (top) and shown with the theoretical target values (buttom)

Start
1. Step 2. Step

White Balance Gamma

Correction

3. Step

Aim
Matrix Correction

4. Step

Six-Axis-
nearly perfect
Operator

Figure 9: Illustration of the four steps towards a good color calibration: White balancing, gamma correction, matrix correction
and correction by the six-axis operator

4
Step 1 – White Balancing Step 3 - Matrix Correction

The calibration focuses in the first step - the white The name of the correction refers to the calculation for
balance - on the grey squares in the bottom line of the realizing this operation, in which the input values are
Greta Macbeth ColorChecker®. The white balancing transformed to the corrected values by means of a
matches the three color channels to the illumination so so-called matrix multiplication. The matrix correction is
that they appear grey, in just the same way as the an attempt to match the spectral sensitivity of the color
human eye does this. pixels in the sensor (color filter of the sensor) to the
spectral sensitivity of the receptors in the eye. The
Step 2 – Gamma Correction correction step works with just six free parameters. For
this reason the optimum result is not normally achieved
After the grey fields, thanks to the white balancing, are in this operation. The final gap is closed in the next
now actually grey and no longer show a color tinge, the correction step.
lightness of the fields is matched to the perception of
the human eye. In contrast to a camera sensor, the Step 4 - Correction by means of the Six-Axis-Operator
human eye does not perceive differences in brightness
linearly. Instead, the differences in brightness in the The six-axis-operator is used for the fine correction of
dark areas are perceived to be greater while they the color representation. In contrast to the matrix
appear weaker in bright areas. This also influences the correction, its adjustment is intuitive. The user can
color perception very strongly. Without correct gamma therefore use this correction step in addition to match
correction, color representation is not possible because the representation to his personal color sensitivities.
otherwise the color saturation would depend on the
brightness. For Basler cameras, a color calibration can be carried
out in a few steps by means of the Pylon software
In
1,0

gamma
correction
1
2.2

0.5 0.5
0,5

CRT
0.218
gamma
0.218
2.2

0 Out
0 0,5 1,0

Figure 10: Brightness perception of the human eye and the

camera (upper curve with gamma 1/2.2 ) And gamma correction
of the display units ( lower curve with gamma 2.2 )

Figure 11: Color calibration of basler cameras via the pylon GUI

5
Summary

How good the color representation of a camera is can

be checked by means of the color fault (∆E).

∆E is a measure of the color difference perceived by a

person between two colors. A color difference of ∆E < 1
is barely perceivable; a color with ∆E > 5 is seen as a
different color. Industrial cameras without color correc-
tion have a ∆E of 10 to 20.

Thanks to the color calibration Basler cameras achieve

an ∆E of 3-4 which corresponds to the high standard of
the printing industry.

3. Step

Matrix Correction

Figure 12: ∆E before (left-hand side) and after color calibration (right-hand side)

06/2012
Basler AG
Germany, Headquarters USA Asia
Tel. +49 4102 463 500 Tel. +1 610 280 0171 Tel. +65 6425 0472
Fax +49 4102 463 599 Fax +1 610 280 7608 Fax +65 6425 0473
[Link]@[Link] [Link]@[Link] [Link]@[Link]
[Link]
6