www.crash-network.com
Efficient lighting for high-speed imaging on vehicle safety test
applications
Burkhard Severon
K.H.Steuernagel Lichttechnik GmbH
Gerauer Str. 56a, D-64546 Mörfelden-Walldorf, Germany
Part 1 Light and Radiation Sources
For the quantity and quality of data recorded by the use of High-Speed-Photography and Videography,
light plays an important role.
Beside the basic requirements concerning illuminance, spatial uniformity and light modulation, the
efficiency in respect to the sensitivity of the high-speed imaging sensor needs to be considered.
On this subject the paper will focus on the effective illumination for the latest technology of highspeed video systems.
1. EFFECTIVE AND BALANCED ILLUMINATION FOR CCD-SENSORS
The use of high speed photography and videography
achieves an essential contribution for the collection of
relevant data at automotive safety testing. Today
especially the high-speed video systems are gaining on
importance.
Figure 1 shows a typical test configuration for this
application.
To receive and process the test data by capturing images
requires an adequate illumination or better irradiation.
The general term for this need is “Light”. The following
expositions may clarify that light is not necessarily all
what is needed to make a high ratio of data available for the receiving sensors.
For this a detailed characterization of the radiation sources and the technical detectors must be
undertaken.
1.2
Definition of Light
V (λ)
Relative spectral sensitivity of photopic vision / V (λ)
100%
80%
rel. sensitivity
60%
40%
20%
0%
780
760
740
720
700
680
660
640
620
600
580
560
540
520
500
480
460
440
420
400
380
If we talk about light and related terms
like, luminous flux (Lumen [lm]),
luminous efficacy (Lumen per watt
[lm/W]), luminous intensity (Candela
[cd] or illuminance (Lux [lx]), it is
important to understand how these
values are measured. The illuminance
given in Lux is the most common value
to define "light" intensity requirements
and specifications. Devices to measure
illuminance have a spectral sensitivity
that is well adjusted to the spectral
sensitivity of the human eye, V (λ), as
shown at figure 2.
nm
Page 1 of 5
Fig. 2
www.crash-network.com
So only the irradiance being
effective to the human eye is
called
and
measured
as
illuminance.
If radiation sources are judged by
there luminous efficacy, they are
not necessarily an efficient
source in respect to the technical
sensors used for photography and
videography. Figure 3 shows an
extreme example on this matter.
The Sodium vapor lamp on the
right, found as a common light
Fig. 3
source for street lighting, has an
extreme high luminous efficacy of about 200 lm/W. The reason for this is given by the sources
spectral power distribution in relation to the V(λ) sensitivity. A "full" spectrum source, like the HMI
lamp on the left, still offers high luminous efficacy, but about only half compared to the Sodium vapor
lamp.
HMI Lamp ~ 100 lm/W
Sodium Vapour Lamp ~ 200 lm/W
Looking to the results given by the two images below the curve of each lamp, proofs how much
information could get lost if ones concentrate just on luminous efficacy. Even transferred to a gray
scale the images clearly show the effect.
The reason for this is found taken
Sensitivity of the Human Eye
again a closer look to the "light"
The Additive Colour System
sensor, the human eye at figure 4.
Here we found that the total
sensitivity is splitted into three (3)
sensitivity ranges, the blue, green
and red shown from left to right.
The additive colour system than
enables our colour vision.
An unbalanced irradiation as
shown by the example above will
therefore lead to divergences or
even losses on colour information.
Same is true if one of the three
ranges is not correctly developed.
Fig. 4
100%
90%
80%
rel. Sensitivity
70%
60%
50%
40%
30%
20%
10%
800
780
760
740
720
700
680
660
640
620
600
580
560
540
520
500
480
460
440
420
400
380
360
340
0%
Wavelength (nm)
1.2
Technical Sensors
Technical sensors, which need to be discussed when searching for effective and balanced irradiation
on
high-speed
imaging
Sensitivity of Video Color Sensor
applications,
are
likewise
characterized by their spectral
sensitivity. Gaining more and
more importance we take a look
to the CCD-Sensor of a highspeed video system (Fig. 5)
Same as for the human eye we
find three independent ranges of
sensitivity. The blue, the green
and the red channel are forming
the total sensitivity s (λ), which is
given by Sblue (λ) + Sgreen (λ) +
Sred (λ ).
Fig. 5
100%
90%
80%
60%
50%
40%
30%
20%
10%
W avelength (nm )
Page 2 of 5
800
780
760
740
720
700
680
660
640
620
600
580
560
540
520
500
480
460
440
420
400
380
360
0%
340
rel. Sensitivity
70%
www.crash-network.com
Εeff = ∫ Eλ ∗ s (λ ) dλ ,
To define now the total effective irradiance for the CCD-Sensor, which is
λ
a look to the spectral power distribution of various radiation sources needs to be taken.
1.3
Radiation Sources
Radiation sources are technical described by the following values:
Radiant Flux Φ =
Q
, measured in W. It indicates which radiation energy per unit of time is leaving
t
the radiation source.
Φ
, measured in W/m². It indicates which radiant flux is impinges an area F.
F
Φλ
Spectral irradiance Ελ =
, measured in W/m² / nm. It indicates the amount of irradiance at an area
F
Irradiance Ε =
F in the wavelength range ∆λ, which is usually 1 nm.
Most important to evaluate the efficiency is the knowledge of the spectral distribution.
As representatives of radiation sources that found frequent use for the application of high-speed
imaging, Daylight (global solar radiation), Tungsten Halogen, Xenon and special Metal Halide lamps
had been selected for evaluation.
Figure 6 shows the spectral distribution of those sources when standardized to the same irradiance
(280 – 3.000nm).
3.000
2.500
mW/m²*nm
2.000
Daylight
MHG
Xenon
Tungsten Halogen
1.500
1.000
500
Wavelength (nm)
Fig. 6
Page 3 of 5
2500
2420
2340
2260
2180
2100
2020
1940
1860
1780
1700
1620
1540
1460
1380
1300
1220
1140
980
1060
900
820
740
660
580
500
420
340
-
www.crash-network.com
1.4
Effective Irradiance
Effective Spectral Irradiance for Video Color Sensor / Total Sensitivity
Total Effective Spectral Irradiance for Color Sensor
2.000
1,20
1.800
1.600
1,00
1.400
0,80
mW/m²*nm
1.200
Daylight
MHG
1.000
0,60
Xenon
Tungsten Halogen
800
0,40
600
400
0,20
200
0,00
Daylight
800
780
760
740
720
700
680
660
640
620
600
580
560
540
520
500
480
460
440
420
400
380
360
340
-
MHG
Xenon
Tungsten
Halogen
Wavelength (nm)
Fig. 7
Fig.8
Figure 7 shows the effective spectral irradiance and figure 8 the total effective irradiance for the CCDSensor given by the four different radiation sources. It shows that for "Daylight" and the MHG lamps
the effective irradiance is about 4 times higher than for Xenon or Tungsten Halogen.
1.5
Effective and Balanced Irradiance
As discussed under 1.1 and 1.2 the
spectral balance of the radiation is even
more important than the total effective
irradiance in respect of colour imaging.
Spectral Power Distribution of different radiation Sources
Standardized to the same Total Irradiance (300-3000nm)
2.000
1.800
1.600
1.400
1.200
mW/m²*nm
Figure 9 shows again the spectral power
distribution of the four selected
radiation sources, but narrowed to the
spectral range of importance, from 340
to 800nm.
The standardization to the same total
irradiance had been kept for this graph.
Daylight
MHG
Xenon
1.000
Tungsten Halogen
800
600
400
200
800
780
760
740
720
700
680
660
640
620
600
580
560
540
520
500
480
460
440
420
400
380
360
340
-
Different from the first approach the
calculation on effective irradiance had
been made now more selective for the three (3) colour channels (blue-green-red).
The individual results for each channel are shown on figure 10.
Wavelength (nm)
Fig. 10 Effective irradiance for blue – green – red CCD sensor channels
Page 4 of 5
Fig. 9
www.crash-network.com
Using this information a comparison of the effective and balanced irradiance could be made.
Figure 11 first shows the summary for blue, green and red channel (from top to bottom). At figure 12
the ratio of total (100%) sensitivity for the three (3) channels is visualized.
Ratio of Total Sensitivity - Video Sensor
Effective Spectral Irradiance for Blue - Green - and Red Channel
100%
1,20
90%
100%
80%
1,00
70%
0,80
60%
50%
0,60
40%
0,40
30%
28,1%
20%
0,20
35,7%
36,2%
10%
0,00
Daylight
MHG
Xenon
0%
Tungsten
Halogen
Total
Fig. 11
Blue
Green
Red
Fig. 12
The ratio of effective irradiance is than given in relation to the ratio of the video sensor at Figure 13,
showing the increasing red and decreasing blue ratio for the Xenon and especially for the Tungsten
Halogen sources.
Using the calculation on the effective irradiance for the three (3) channels as shown in figure 11 and
recalculate the results to achieve a effective irradiance balanced to the ratio of the CCD sensor, will
finally give the effective and balanced irradiance offered by the different radiation sources (Fig. 14).
Effective and Balanced Spectral Irradiance for
Blue - Green - and Red Channel
Ratio of Effective Irradiance for Blue - Green - and
Red Channel
1,12
1,20
100,0%
80,0%
35,7%
34,4%
34,3%
36,9%
1,00
1,00
47,4%
0,80
60,0%
36,2%
38,9%
36,7%
20,0%
0,60
37,0%
40,0%
35,4%
28,1%
26,7%
29,0%
26,1%
17,2%
0,33
0,40
0,17
0,20
0,00
0,0%
Video Sensor
Daylight
MHG
Xenon
Tungsten
Halogen
Fig. 13
Daylight
MHG
Xenon
Tungsten
Halogen
Fig. 14
So only radiation sources offering a close match to the spectral sensitivity of the sensors will enable
the highest collection of test data at high-speed imaging processes.
Page 5 of 5