Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/366290110

A Grid in Perspective for Road Lighting Calculations

Article in LEUKOS The Journal of the Illuminating Engineering Society of North America · December 2022
DOI: 10.1080/15502724.2022.2135529

CITATIONS READS

3 208

5 authors, including:

Florian Greffier Vincent Boucher


Centre d'études et d'expertise sur les risques l'environnement la mobilité et l'amé… Centre d’études et d’expertise sur les risques, l’environnement, la mobilité et l’am…
31 PUBLICATIONS 130 CITATIONS 52 PUBLICATIONS 228 CITATIONS

SEE PROFILE SEE PROFILE

Valerie Muzet Sandy Buschmann


Centre d’études et d’expertise sur les risques, l’environnement, la mobilité et l’am… Technische Universität Berlin
47 PUBLICATIONS 233 CITATIONS 3 PUBLICATIONS 8 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Valerie Muzet on 08 February 2023.

The user has requested enhancement of the downloaded file.


LEUKOS
The Journal of the Illuminating Engineering Society

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/ulks20

A Grid in Perspective for Road Lighting


Calculations

Florian Greffier, Vincent Boucher, Valérie Muzet, Sandy Buschmann &


Stephan Völker

To cite this article: Florian Greffier, Vincent Boucher, Valérie Muzet, Sandy Buschmann &
Stephan Völker (2022): A Grid in Perspective for Road Lighting Calculations, LEUKOS, DOI:
10.1080/15502724.2022.2135529

To link to this article: https://doi.org/10.1080/15502724.2022.2135529

© 2022 The Author(s). Published with


license by Taylor & Francis Group, LLC.

Published online: 15 Dec 2022.

Submit your article to this journal

Article views: 164

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=ulks20
LEUKOS
https://doi.org/10.1080/15502724.2022.2135529

A Grid in Perspective for Road Lighting Calculations


a a b
Florian Greffier , Vincent Boucher , Valérie Muzet , Sandy Buschmannc, and Stephan Völkerc
a
Light and Lighting Research Team, Cerema, Les Ponts-de-Cé, France; bENDSUM Research Team, Cerema, Strasbourg, France; cLighting
Technology, Technische Universität Berlin, Berlin, Germany

ABSTRACT ARTICLE HISTORY


In road lighting, calculations are made within the rectangular CIE grid while measurements Received 13 April 2022
conducted with Imaging Luminance Measurement Devices (ILMD) have a trapezoidal shape. We Revised 6 October 2022
propose to build a grid that takes the observer’s perspective into account by projecting ellipses of Accepted 8 October 2022
adjustable dimensions than can match human vision or ILMD angular aperture and allow direct
KEYWORDS
comparison with luminance images. Lighting standard criteria definitions remain unchanged and Road lighting quality;
values of average luminance and longitudinal uniformity are equivalent. However, as the grid in calculation grid
perspective describes more finely the luminance distribution, the overall uniformity is lower than
in the classical grid but more representative of observer’s perception. We present in detail our
approach, give the corresponding code, expose calculations performed with this new grid and
show an example of direct comparison between calculations and measurements.

1. Introduction
Discrepancies are often observed between simu­
The design of road lighting installations is per­ lations and experimental measurements regarding
formed according to the specifications of CIE the quality criteria (average luminance, overall and
documents and relevant standards (ANSI/IESNA, longitudinal uniformities) and luminance distribu­
2000; CEN 2015; CIE 2019). To calculate lumi­ tions. These deviations may be due to various
nance-based quality criteria, a rectangular grid of factors:
the relevant area has to be defined. There are
usually three transversal points and a minimum ● uncertainties on light sources distribution
of ten longitudinal points per traffic lane, i.e. typi­ and power,
cally 60 points for two lanes (Fig. 1). This standard ● the use of a standard r-table in calculations
methodology is efficient, very pragmatic for on- instead of a measured r-table,
site punctual measurements and does not need ● planarity defaults,
advanced calculation tools. For more than ● uncertainty of the luminance-measuring
60 years, manual measurements were conducted device,
by luxmeters, then replaced by punctual lumi­ ● methodology of exploitation of the experi­
nance meters and recently by Imaging mental luminance measurement.
Luminance Measurement Devices (ILMD’s).
Nowadays in road lighting design, ILMD’s are In this paper, we will address this last possible
able to produce experimental luminance maps source of discrepancy. Having a common exploita­
like the one in Fig. 2. Their angular resolution tion of ILMD is a first step, available to all and
can be much finer than that of the human eye easy to implement.
(Buschmann et al. 2017; Greffier et al. 2019; On the one hand, the camera provides
Rossi et al. 2017). These on-site images are often a perspective image of the measured area, which
compared with calculations obtained by common has a trapezoidal shape. On the other hand, the
lighting calculation software (Radiance, Dialux, calculations are carried out on a rectangular grid
Relux) that still use the standard grid. defined by the CIE. To match the ILMD

CONTACT Vincent Boucher vincent.boucher@cerema.fr Light and Lighting Research Team, Cerema, 23 Avenue Amiral Chauvin, 49136 Les Ponts-de-
Cé, France
© 2022 The Author(s). Published with license by Taylor & Francis Group, LLC.
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-
nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built
upon in any way.
2 F. GREFFIER ET AL.

Fig. 1. Standard grid definition.

Fig. 2. Luminance measurement map from the standard observer’s perspective.

experimental results and the calculations with the The method a. involves transforming the
standard CIE 6 × 10 grid, several methodologies imaged field from a trapezoidal shape to
are possible, like: a rectangular one (Buschmann et al. 2020). This
mathematical transformation is called
a. Stretching the trapezoidal shape to transform a rectification and requires interpolations that
it in a rectangular one. Then, the grid points have an impact on the measurements (Boucher
are calculated by a rectangular shape, or et al. 2021).
b. Projecting the grid points in the ILMD tra­ The method b. does not transform the measured
pezoidal shape. field but needs to define the grid points according
LEUKOS 3

to the parameters as height, line of sight, pixel size, transverse direction (Fig. 3). Angular aper­
focal length. It is finally a way to take into account tures can be chosen to simulate the observer’s
the observer’s perspective at the price of many vision, then ellipses represent the retinal cells
geometrical calculations. projection on the road surface. Whether they
The final step, common for both methods, is are known, apertures can fit the ILMD ones.
to average luminances within areas of dimen­ For ILMD measurements, (CEN 2015)
sions about 2’x20’ (arcminutes) centered on the recommends an angular aperture in the
grid points. These dimensions are a requirement range [1’, 2’].
of (CIE 2011) but its justification is related to ● Taking observer’s position as a reference, the
(AFNOR 2006) for tunnel lighting that does not first line of the grid is constructed by ellipses
provide additional justification nor references. projection in front of the observer. It corre­
Whatever the method, the fine angular resolu­ sponds to the center of the driving lane. This
tion of ILMD is lost because the luminance dis­ first line (star markers in Fig. 4) is used to the
tribution is finally described by only 6 × 10 longitudinal uniformity calculation as defined
values. in standards.
In this paper, we expose a method of direct ● The whole grid is generated adding ellipses
comparison between experimental luminance next to the first line according to the ellipses
image and calculation. This method will not width. As the ellipses width increases with
resolve the discrepancies often observed due to distance, the grid owns more ellipses at the
lack of knowledge of some parameters (real source beginning of the mesh than at the end, trans­
power, exact source tilt, planarity default, road lating the perspective view.
surface r-table, . . .). But our aim is to propose
and diffuse a method that do not induce deviations The grid in perspective owns some hundreds
due to the mathematical process. Therefore, in our of points describing more finely the luminance
approach, the measured luminance map is not distribution than the classical grid. It adapts the
transformed to match a calculation grid, but con­ calculation to the field of view and generates
versely adapts the calculation to the measurement. a trapezoidal grid more convenient to fit ILMD
The general principle is the following. Instead of images. Moreover, the three classical road light­
the classic grid, a grid based on observer’s perspec­ ing criteria definitions are unchanged.
tive and a given angular aperture is used. Our goal here is to describe in detail our
approach and to give the corresponding code
● The positions of the grid points are defined as so that it becomes available to all. We also pre­
the centers of ellipses projected from the sent results on calculations under several light­
observer’s position with given angular aper­ ing situations and an example of comparison
tures in the longitudinal direction and in the between measurement and calculation.

Fig. 3. Ellipses projection from observer’s eye.


4 F. GREFFIER ET AL.

Fig. 4. Points coordinates within the grid in perspective (ApL = 2’, ApT = 20’).

2. Perspective grid construction %abscissas of ellipses’ centres


Xp = Obs.Z/tand(atand(Obs.Z/abs(Obs.X))-1/
In the following we describe step by step the con­ 2*ApL) ; % first centre abscissa
struction of the grid. All the variables implied and
the Matlab® code lines are explained. The entire Then abscissas Xp(i) are calculated recursively
code is available in Appendix A. with a decreasing projection angle substracting the
First we need to define some scene parameters: the longitudinal aperture each time. Results are stored
road width, the spacing between two luminaires, the in a vector Xp while the calculated center stands in
observer’s position and the apertures of ellipses. In this the calculation area.
example, the axis origin states at 60 m from the obser­ while Xp(end)<(Spacing + abs(Obs.X)) % abscissas
ver (see Fig. 3) and angular apertures are fixed to 2’. into the calculation area
Xp = [Xp ; Obs.Z/tand(atand(Obs.Z/Xp(end))-ApL)] ;
%variables definition end
Spacing = 20 ; % spacing between two luminaires Xp = Xp(1:end-1) ;% delete the last value outside the
in m area
Wr = 6; % road width in m
Obs.X = -60 ; % longitudinal observer’s position Output vectors are instantiated. They corre­
in m spond to centers’ coordinates for two separate
Obs.Y = [4 2] ; % transversal observer’s positions observer’s positions, one in the center of the
in m fast lane and the other in the center of the
Obs.Z = 1.5 ; % observer’s eye height in m slow lane.
ApL = 2/60 ; % Longitudinal Aperture in degree
ApT = 2/60 ; % Transversal Aperture in degree Xp_FL = [] ; % Abscissas of points for fast lane
observer’s position
Yp_FL = [] ; % Ordinates of points for fast lane
First, we calculate the abscissas of ellipses’
observer’s position
centers. The first center abscissa Xp(1) is shifted Xp_SL = [] ; % Abscissas of points for slow lane
from the 60 m line by a half of the longitudinal observer’s position
aperture. A bit of trigonometry leads to the Yp_SL = [] ; % Ordinates of points for slow lane
following expression: observer’s position
LEUKOS 5

Ordinates of ellipses’ centers can now be calcu­ a direct comparison with experimental measure­
lated for each abscissa. As they depend on observer’s ments. The image is initialized once the first row
transverse position, the procedure for the fast of centers is calculated. The image is defined as
lane will be repeated for the slow lane (given in an array of dimensions: number of abscissas by
Appendix A). The first ordinate corresponds to the number of ordinates at the first abscissa. As the
center of the given lane. A counter is also initialized number of ellipses decreases with distance by the
to keep track of the number of centers on each side of perspective effect, the image will finally look like
the center of the lane. It will be used to construct a trapezoid surrounded by zeros.
a mask representing the grid seen by the observer.
if x==1
%ordinates of ellipses’ centres % image seen from observer’s position
for x = 1:length(Xp) center_lane_FL=c1+1; % index of central row
% fast lane Mask_FL=zeros(length(Xp),length(Yp),‘logical’);
Yp = Obs.Y(1) ;% first centre ordinate % mask initialisation
c1=-1;% count centres on the left side of the first end
centre Mask_FL(length(Xp)-x+1,center_lane_FL-c1:
% (seen from the observer) center_lane_FL+c2)=1;
%fill image at given indices
Ordinates Yp(i) are calculated recursively from
the first center adding the transversal aperture The same procedure is repeated for the slow
each time. They stand on the left side of the first lane observer’s position.
center seen from the observer. Results are stored in Calculations can now be done at (Xp_SL,
a vector Yp in descending order while the calcu­ Yp_SL) positions for an observer in slow lane
lated ordinate is inside the road width. and at (Xp_FL,Yp_FL) positions for an observer
while Yp(1)<Wr % ordinates into the road width in fast lane. Figure 4 presents results of previous
Yp = [Xp(x)*tand(atand(Yp(1)/Xp(x))+ApT) ; Yp] ; calculations for the slow lane observer’s position
c1=c1+1;
end
(with ApT = 20’ to be drawable). Star markers
Yp = Yp(2:end) ;% delete last value outside the road represent the line in front of the observer.
width Mask_SL and Mask_FL matrices can be filled
with calculated luminances to compare with ILMD
Then ordinates Yp(i) corresponding to the right images. Note that the longitunal and transverse
side of the first center are calculated by subtracting aperture parameters, ApL and ApT, can be
the transversal aperture each time. adjusted to exactly match the imaging system if
c2=-1;% count centres on the right side of the first needed. Figure 5 shows the image Mask_SL with
centre (seen from the observer) ApL = 2’ and ApT = 20’ where (Xp_SL, Yp_SL)
while Yp(end)>0 % ordinates into the road width
Yp = [Yp ; Xp(x)*tand(atand(Yp(end)/Xp(x))-
crosses are superimposed.
ApT)] ;
c2=c2+1;
end
3. Calculations with the grid in perspective
Yp = Yp(1:end-1) ; % delete last value outside the
road width To assess the impact of using the grid in perspective
instead of the classical CIE one, simulations were
At this point, all ordinates have been calcu­
conducted with both approach on 10 different light­
lated for each abscissa. They can be stored in the
ing situations. The seven CIE situations proposed in
output vectors taking care of vectors dimensions
(CIE 2019) with standard r-table R3 (CIE 2001) and
and of the axis origin.
three French experimental situations evaluated during
Xp_FL = [Xp_FL ; Xp(x)*ones(length(Yp),1)+Obs.X] ;
Yp_FL = [Yp_FL ; Yp] ;
the Lumiroute® experiment (Muzet et al. 2019) with
LED or Metal Halide luminaires and experimental
Besides the grid points coordinates used for r-tables obtained on site with Coluroute device
further calculations, an image (in pixels) seen by (Muzet et al. 2008). The parameters of the 10 lighting
the observer is also constructed. It will allow situations are given in Appendix B.
6 F. GREFFIER ET AL.

Fig. 5. Mask of the grid in perspective seen by the observer (ApL = 2’, ApT = 20’).

All calculations are made with Ecl_R, uniformity is also not very sensitive to the grid
a lighting calculation engine developed by change as about the same number of longitudinal
Cerema (Greffier et al. 2021). points are used. On the other hand, the overall uni­
For both approaches, the lighting quality formity is always higher with the classical grid. The
criteria of (CEN 2016) were performed: perspective grid containing a larger number of cal­
culation points, the luminance distribution is ana­
● The average luminance is calculated as the lyzed more finely than with 60 points. This explains
arithmetic mean of the luminances obtained why the extreme values (min and max) are different
at the grid points. and lead to lower overall uniformities. If angular
● The overall uniformity, Uo, is calculated as apertures are chosen to simulate the vision, this
the ratio of the lowest luminance to the overall uniformity seems more representative of the
average luminance in the field of calculation. observer’s perception because it is closer to the eye
● The longitudinal uniformity, Ul, is calculated as angular resolution.
the ratio of the lowest to the highest luminance
in the longitudinal direction along the center-
line of each lane. The observer’s position should 4. Measurement vs calculation
be in line with the row of calculation points. A comparison of simulated luminance and mea­
surement with ILMD has been conducted on the
All the results are presented in the Table 1. The test case LUMIROUTE® 4. Luminaires stand at
deviations in percentage between the two grids are a height of 9 m with a spacing of 29 m. The road
computed using (Xperspective – Xclassical)/Xclassical * 100, width is 5.7 m. The road reflection characteristics
where X stands for average luminance and unifor­ have been measured by taking drill core samples
mities. As expected, the average luminance varies and measuring the r-table in laboratory according
very little between the classical and perspective to (CIE 2001). Road surface exhibits a Q0 = 0.066
grids. Indeed, the average value is not very sensitive and S1 = 1.10 and the corresponding r-table is
to the number of calculation points. The longitudinal used for the following ECL_R calculations.
LEUKOS 7

Table 1. Lighting quality criteria calculations with classical and perspective grid.
Classical Grid Grid in Perspective Deviation
fast lane slow lane fast lane slow lane fast lane slow lane
Situation CIE 1 Laverage 0.26 0.26 0.25 0.25 −4% −4%
Uo 0.14 0.15 0.12 0.12 −16% −19%
Ul 0.56 0.64 0.55 0.64 −2% 1%
Situation CIE 2 Laverage 0.26 0.26 0.26 0.26 −2% −2%
Uo 0.32 0.35 0.25 0.26 −23% −26%
Ul 0.33 0.33 0.34 0.34 3% 1%
Situation CIE 3 Laverage 0.55 0.60 0.52 0.58 −5% −4%
Uo 0.52 0.54 0.5 0.51 −5% −6%
Ul 0.68 0.70 0.65 0.71 −3% 2%
Situation CIE 4 Laverage 0.88 0.87 0.87 0.83 −1% −4%
Uo 0.54 0.51 0.47 0.46 −13% −10%
Ul 0.53 0.53 0.53 0.53 0% 1%
Situation CIE 5 Laverage 0.44 0.41 0.42 0.39 −3% −3%
Uo 0.45 0.67 0.43 0.63 −5% −6%
Ul 0.79 0.75 0.79 0.74 0% −2%
Situation CIE 6 Laverage 1.17 1.31 1.09 1.24 −7% −5%
Uo 0.49 0.49 0.47 0.46 −3% −6%
Ul 0.62 0.72 0.61 0.72 −1% 0%
Situation CIE 7 Laverage 0.53 0.58 0.53 0.58 −1% 0%
Uo 0.45 0.46 0.41 0.41 −8% −11%
Ul 0.84 0.91 0.82 0.88 −2% −3%
LUMIROUTE® 1 Laverage 0.89 0.95 0.88 0.94 −1% 0%
Uo 0.56 0.57 0.49 0.48 −12% −14%
Ul 0.84 0.79 0.85 0.77 1% −2%
LUMIROUTE® 2 Laverage 1.21 1.28 1.2 1.26 −1% −1%
Uo 0.61 0.62 0.54 0.54 −11% −13%
Ul 0.82 0.82 0.82 0.79 1% −3%
LUMIROUTE® 4 Laverage 0.71 0.78 0.70 0.77 −1% −1%
Uo 0.66 0.66 0.58 0.56 −13% −15%
Ul 0.77 0.78 0.73 0.77 −4% −1%

Image luminance measurements were conducted calculations are able to reproduce measurements. It
using an HDR-ILMD (Boucher et al. 2017) based on should be noted that this level of knowledge is rarely
four CMOS cameras to provide a wide range of lumi­ achieved in operational conditions.
nance. The standard observer’s position according to Applying the perspective method with angular
(CIE 2019) was used and luminance maps were apertures, ApT and ApL equal to 1.5’, the grid con­
recorded with the above mentioned setup. The focal tains 215 points (ellipse centers) at 60 m in transverse
length of the imaging system is 12.5 mm and the pixel direction, 148 points at 89 m and 19 points in long­
pitch of the sensor is 5.86 μm, thus the pixel angular itudinal. Then the image dimension is 19 × 215 pix­
aperture is 1.5’ (averaged over the whole sensor area) els with a trapezoidal luminance map surrounded by
and the measurement field is imaged by 18 × 226 zeros (see Fig. 7). This dimension is near but not
pixels (black trapeze in Fig. 2). exactly the experimental image (18 × 226). So the
The method b. is first applied in averaging the measured image is resized to fit exactly the calcula­
luminance over 1.5’ × 20’ (i.e. 1 × 13 pixels) tion. We have verified that there is no incidence on
around each grid point. Figure 6 (left) presents results (up to third digit). Figure 7 (left) presents the
results of method b. applied on the ILMD image. ILMD trapezoidal area of Fig. 2 resized to fit exactly
Figure 6 (right) presents result of calculation. the calculated grid size. Figure 7 (right) presents
In this test case, for the classical 6 × 10 grid, the result of calculation with ApT and ApL equal to 1.5’.
measured and calculated road lighting quality criteria For the perspective method with a large number
are in very good agreement, with a difference of 1.3% of measurement considered (3444 points), the mea­
for luminance and a maximum difference of 7.6% for sured and calculated road lighting quality criteria are
overall uniformity. Since all the parameters (light also in very good agreement, with a maximum dif­
sources, scene geometry, road surface and measured ference of 4%. One minor drawback of this method
area) have been carefully measured, it explains why is illustrated in Fig.7 (left): it includes the central
8 F. GREFFIER ET AL.

Fig. 6. Measured grid by method b. (left) and calculated grid (right) for the 60 CIE points.

Fig. 7. Measured grid (left) and calculated grid (right) using the grid in perspective.

marking whereas it is not present in the CIE grid. one hand, the camera provides a perspective image of
Here, markings are imaged by 30 pixels representing the area to be measured, on the other hand, the calcu­
less than 1% of the imaged area and their influence lations are carried out on a rectangular grid defined by
can be considered negligible. If it is not the case, the CIE. To avoid the drawbacks of a comparison
markings should be removed. between trapezoidal measurements and rectangular
Whatever the number of points considered, average
calculations, we propose to build a grid that takes the
luminance and longitudinal uniformity are not sensi­
observer’s perspective into account by projecting
tive to the grid change. However the overall unifor­
mity decreases with the perspective grid as the ellipses of adjustable dimensions than can match
luminance distribution is much more finely described human vision or imaging system angular aperture
(but remains over the standard threshold of U0 ≥0.4). and allow direct comparison with images provided
We have shown with this example that the CIE light­ by an ILMD.
ing criteria are still consistent between calculation and In this paper, we present both the methodology
measurement with the perspective method. Finally, there and a software implementation, providing the
is no more need to project points in an image frame. code in appendix.
Benefits of the perspective grid are to not induce errors To test the implementation, lighting quality cri­
due to a stretching of the measured data (to fit a rectangle teria calculations have been carried out using the
from a trapeze) and to allow a direct comparison as the 60 points CIE grid and our approach for ten situa­
calculation fits exactly the measured image. tions: seven CIE situations and three experimental
ones. Whatever the methodology, the lighting
standard criteria definitions remain unchanged
and we have shown that the values of average
5. Conclusion
luminance and longitudinal uniformity are equiva­
The comparison of luminance maps cannot be made lent. The luminance distribution is more finely
directly between experiments and calculations: on the described using more calculation points, thus
LEUKOS 9

calculated and measured overall uniformity tends (HDR-ILMD) and applications in motion. Proceedings of
to decrease but it is more representative of the CIE 2017 Midterm Meetings and Conference on Smarter
Lighting for Better Life; Jeju Island, Republic of Korea:
driver’s perception. A trapezoidal mask is also
International Commission on Illumination, CIE. p. 923–32.
generated and is used to fit ILMD images. doi: 10.25039/x44.2017.
The example of the comparison between calcu­ Buschmann S, Steblau J, Voelker S. 2017. New image based
lation and measurement shows the advantages of measurement method of reflective properties of road
the perspective grid. Increasing the resolution surfaces. CIE 2017 Midterm Meetings and Conference on
could allow other quality criteria to be considered. Smarter Lighting for Better Life. p. 284–93. doi:10.25039/x44.
2017.OP39.
For example, the standard deviation of the lumi­
Buschmann S, Völker S, Schumacher H. 2020. Steigerung der
nance distribution could be used to represent Energieeffizenz in der Straßenbeleuchtung durch
a more representative overall uniformity than the Entwicklung und Evaluierung einer nutzflächenbezogenen
ratio of two values. Gradient calculations in the Beleuchtung : schlussbericht StEffi. Technische Universität
longitudinal and transverse directions could also Berlin. doi:10.2314/KXP:1751371913.
become uniformity criteria in these dimensions. CEN. 2015. EN 13201-3:2015. Road lighting—Part 3 : calcu­
lation of performance. Brussels (Belgium): CEN.
They might be more representative of visual com­
CEN. 2016. EN 13201-4:2016. Road lighting—Part 4 : methods
fort than the track axis alone. This work now of measuring lighting performance. Brussels (Belgium): CEN.
needs to be consolidated by multiple comparisons CIE. 2001. CIE 144:2001—Road surface and road marking
between calculations and measurements to con­ reflection characteristics. Vienna (Austria): International
firm the trends. We hope that researchers in the Commission on Illumination (CIE).
field will use this new grid to explore its potential. CIE. 2011. CIE 194:2011—On site measurement of the photo­
metric properties of road and tunnel lighting. Vienna
(Austria): International Commission on Illumination (CIE).
Disclosure statement CIE. 2019. CIE 140:2019 road lighting calculations. 2nd ed.
International Commission on Illumination (CIE). doi:10.
No potential conflict of interest was reported by the 25039/TR.140.2019.
authors. Greffier F, Charbonnier P, Tarel J-P, Boucher V, Fournela F.
2015. An automatic system for measuring road and tunnel
lighting performance. Proceedings of 28th Quadrennial
Funding Session of the CIE, CIE 216, Manchester, UK. p. 1647–56.
Greffier F, Muzet V, Boucher V, Fournela F, Dronneau R. 2019.
This work was supported by the Cerema “Tremplin Europe & Use of an imaging luminance measuring device to evaluate
International” under Grant Savaphor-2020. road lighting performance at different angles of observation.
PROCEEDINGS OF the 29th Quadrennial Session of the
CIE. p. 553–62. doi:10.25039/x46.2019.OP75.
ORCID Greffier F, Muzet V, Boucher V, Fournela F, Lebouc L,
Florian Greffier http://orcid.org/0000-0001-7745-6629 Liandrat S. 2021. Influence of pavement heterogeneity and
Vincent Boucher http://orcid.org/0000-0003-2299-4168 observation angle on lighting design : study with new metrics.
Valérie Muzet http://orcid.org/0000-0002-0026-6592 Sustainability. 13(21):11789. doi:10.3390/su132111789.
Muzet V, Greffier F, Nicolaï A, Taron A, Verny P. 2019.
Evaluation of the performance of an optimized road sur­
References face/lighting combination. Light Res Technol. 51
(4):576–591. doi:10.1177/1477153518808334.
AFNOR. 2006. FD CEN/CR 14380—Éclairagisme—Éclairage Muzet V, Paumier J-L, Guillard Y. 2008. COLUROUTE :
des tunnels. Plaine Saint-Denis (France): Association a mobile gonio-reflectometer to characterize the road surface
française de normalisation. photometry. 2nd CIE Expert Symposium on « Advances in
ANSI/IESNA. 2000. American national standard practice for Photometry Abd Colorimetry », CIE x033. https://www.
roadway lighting. New-York (USA): Illuminating researchgate.net/publication/279258920_COLUROUTE_a_
Engineering Society. RP-8-00. mobile_gonio-reflectometer_to_characterize_the_road_sur
Boucher V, Buschmann S, Greffier F, Muzet V, Voelker S. 2021. face_photometry.
Keeping the benefit of ILMD’s high resolution in measuring Rossi G, Iacomussi P, Radis M. 2017. Metrological characteriza­
road lighting quality parameters. CIE 2021 Midterm Meeting tion of ILMD for smart lighting applications. Proceedings of
and Conference. p. 431–41. doi:10.25039/x48.2021.OP55. CIE 2017 Midterm Meetings and Conference on Smarter
Boucher V, Dumont E, Dronneau R, Fournela F, Greffier F. 2017. Lighting for Better Life. p. 218-27. doi:10.25039/x44.2017.
High dynamic range imaging Luminance measuring device OP31.
10 F. GREFFIER ET AL.

APPENDICES
Appendix A. Grid calculation code
Yp = Yp(1:end-1) ; % delete last value outside the road
%variables definition width
Spacing = 20 ; % spacing between two luminaires in m Xp_FL = [Xp_FL ; Xp(x)*ones(length(Yp),1)+Obs.X] ;
Wr = 6; % road width in m Yp_FL = [Yp_FL ; Yp] ;
Obs.X = -60 ; % longitudinal observer’s position in m if x==1
Obs.Y = [4 2] ; % transversal observer’s positions in m % image seen from observer’s position
Obs.Z = 1.5 ; % observer’s eye height in m center_lane_FL=c1+1; % index of central row
ApL = 2/60 ; % Longitudinal Aperture in degree Mask_FL=zeros(length(Xp),length(Yp),’logical’); % mask
ApT = 2/60 ; % Transversal Aperture in degree initialisation
%abscissas of ellipses’ centres end
Xp = Obs.Z/tand(atand(Obs.Z/abs(Obs.X))-1/2*ApL) ;% Mask_FL(length(Xp)-x+1,center_lane_FL-c1:
first centre abscissa center_lane_FL+c2)=1; %fill image at given indices
while Xp(end)<(Spacing + abs(Obs.X)) % abscissas into % slow lane
the calculation area Yp = Obs.Y(2) ; % first centre ordinate
Xp = [Xp ; Obs.Z/tand(atand(Obs.Z/Xp(end))-ApL)] ; c3=-1;% count centres on the left side of the first centre %
end (seen from the observer)
Xp = Xp(1:end-1) ;% delete the last value outside the area while Yp(1)<Wr
Xp_FL = [] ; % Abscissas of points for fast lane observer’s Yp = [Xp(x)*tand(atand(Yp(1)/Xp(x))+ApT) ; Yp] ;
position c3=c3+1;
Yp_FL = [] ; % Ordinates of points for fast lane observer’s end
position Yp = Yp(2:end) ; % delete last value outside the road
Xp_SL = [] ; % Abscissas of points for slow lane observer’s width
position c4=-1;% count centres on the right side of the first centre
Yp_SL = [] ; % Ordinates of points for slow lane observer’s % (seen from the observer)
position while Yp(end)>0
%ordinates of ellipses’ centres Yp = [Yp ; Xp(x)*tand(atand(Yp(end)/Xp(x))-ApT)] ;
for x = 1:length(Xp) c4=c4+1;
% fast lane end
Yp = Obs.Y(1) ;% first centre ordinate Yp = Yp(1:end-1) ; % delete last value outside the road
c1=-1;% count centres on the left side of the first centre % width
(seen from the observer) Xp_SL = [Xp_SL ; Xp(x)*ones(length(Yp),1)+Obs.X] ;%
while Yp(1)<Wr % ordinates into the road width abscissas coord.
Yp = [Xp(x)*tand(atand(Yp(1)/Xp(x))+ApT) ; Yp] ; Yp_SL = [Yp_SL ; Yp] ;% ordinates coord.
c1=c1+1; if x==1
end % image seen from observer’s position
Yp = Yp(2:end) ;% delete last value outside the road width center_lane_SL=c3+1; % index of central row
c2=-1;% count centres on the right side of the first centre Mask_SL=zeros(length(Xp),length(Yp),’logical’); % mask
% (seen from the observer) initialisation
while Yp(end)>0 % ordinates into the road width end
Yp = [Yp ; Xp(x)*tand(atand(Yp(end)/Xp(x))-ApT)] ; Mask_SL(length(Xp)-x+1,center_lane_SL-c3:
c2=c2+1; center_lane_SL+c4)=1; %fill image at given indices
end end
LEUKOS 11

Appendix B. CIE and LUMIROUTE® Situations

Situation CIE 1

Luminaire Source Type Fluo compact


Photometry Sit1.xls
Height 4m
Spacing 20 m
Overhang 2m
Tilt 0°
Power 24 W
Flux 1 800 lm
Arrangement Single-sided Left
Maintenance Factor 0.8
Road Surface Standard R3
Width 10 m
Lanes 2
Observer Observation angle 1°

Situation CIE 2

Luminaire Source Type Fluo compact


Photometry Sit2.xls
Height 4m
Spacing 20 m
Overhang 2m
Tilt 0°
Power 24 W
Flux 1 800 lm
Arrangement Staggered
Maintenance Factor 0.8
Road Surface Standard R3
Width 10 m
Lanes 2
Observer Observation angle 1°
12 F. GREFFIER ET AL.

Situation CIE 3

Luminaire Source Type High-pressure Sodium


Photometry Sit3.xls
Height 10 m
Spacing 45 m
Overhang 0m
Tilt 5°
Power 100 W
Flux 10 500 lm
Arrangement Single-sided Left
Maintenance Factor 0.8
Road Surface Standard R3
Width 7.5 m
Lanes 2
Observer Observation angle 1°

Situation CIE 4

Luminaire Source Type High-pressure Sodium


Photometry Sit4.xls
Height 8m
Spacing 35 m
Overhang 0m
Tilt 5°
Power 100 W
Flux 10 500 lm
Arrangement Staggered
Maintenance Factor 0.8
Road Surface Standard R3
Width 7m
Lanes 2
Observer Observation angle 1°
LEUKOS 13

Situation CIE 5

Luminaire Source Type High-pressure Sodium


Photometry Sit5.xls
Height 12 m
Spacing 54 m
Overhang 0m
Tilt 5°
Power 100 W
Flux 10 500 lm
Arrangement Opposite
Maintenance Factor 0.8
Road Surface Standard R3
Width 10.5 m
Lanes 3
Observer Observation angle 1°

Situation CIE 6

Luminaire Source Type High-pressure Sodium


Photometry Sit6.xls
Height 18 m
Spacing 75 m
Overhang −2.25 m
Tilt 5°
Power 250 W
Flux 32 000 lm
Arrangement Central
Maintenance Factor 0.8
Road Surface Standard R3
Width 7m
Lanes 2
Observer Observation angle 1°
14 F. GREFFIER ET AL.

Situation CIE 7

Luminaire Source Type High-pressure Sodium


Photometry Sit7.xls
Height 10 m
Spacing 40 m
Overhang 0m
Tilt 5°
Power 90 W
Flux 13 000 lm
Arrangement Single-sided Left
Maintenance Factor 0.8
Road Surface Standard R3
Width 8m
Lanes 2
Observer Observation angle 1°
LEUKOS 15

Situation LUMIROUTE® 1

Luminaire Source Type LED


Photometry Lumiroute1.xls
Height 9m
Spacing 30 m
Overhang −0.62 m
Tilt 0°
Power 50 W
Flux 3 500 lm
Arrangement Central
Maintenance Factor 1
Road Surface S1Lumi1P1BrT36
Width 6.5 m
Lanes 2
Observer Observation angle 1°

Situation LUMIROUTE® 2

Luminaire Source Type LED


Photometry Lumiroute2.xls
Height 9m
Spacing 30 m
Overhang −0.62 m
Tilt 0°
Power 66 W
Flux 4 600 lm
Arrangement Central
Maintenance Factor 1
Road Surface S2Lumi2P1BrT36
Width 6.5 m
Lanes 2
Observer Observation angle 1°
16 F. GREFFIER ET AL.

Situation LUMIROUTE® 4

Luminaire Source Type Metallic iodide


Photometry Lumiroute3&4.xls
Height 9m
Spacing 29 m
Overhang −1.42 m
Tilt 5°
Power 140 W
Flux 6 950 lm
Arrangement Central
Maintenance Factor 1
Road Surface S4TemoinP2BrT36
Width 5.7 m
Lanes 2
Observer Observation angle 1°

View publication stats

You might also like