OSA's Digital Library

Optics Express

Optics Express

  • Editor: Andrew M. Weiner
  • Vol. 21, Iss. 22 — Nov. 4, 2013
  • pp: 26812–26819
« Show journal navigation

A single analytical model for sparkle and graininess patterns in texture of effect coatings

A. Ferrero, J. Campos, A. M. Rabal, and A. Pons  »View Author Affiliations


Optics Express, Vol. 21, Issue 22, pp. 26812-26819 (2013)
http://dx.doi.org/10.1364/OE.21.026812


View Full Text Article

Acrobat PDF (1361 KB)





Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

Flakes in effect coatings, which are responsible for its color shift, are not perfectly horizontally oriented, producing a non uniform texture when observed from a distance of around one meter or less. If the coating is illuminated by a diffuse source, a granular appearance is observed, called graininess. But when the coating is illuminated by unidirectional light, some luminous spots on a dark background appear, effect that is known as sparkle. The characterization of these two textures is getting more important for automotive industry because of the high percentage of cars with effect coatings and because improvements in imaging technology allows this effect to be measured by, for instance, commercial instruments as BYK-mac. A single analytical model to understand and radiometrically characterize both sparkle and graininess in effect coatings is presented and studied in this work. It allows both patterns to be explained and, despite its simplicity, includes variables related to the optical system (Point Spread Function (PSF) and size of the entrance pupil), its distance to the coating, the diffusion grade of the illumination, the illumination and observation directions, and coating parameters.

© 2013 OSA

1. Introduction

This kind of coating consists of a transparent substrate having embedded metallic or interference pigments that are horizontally arranged [3

3. G. Pfaff and P. Reynders, “Angle-Dependent Optical Effects Deriving from Submicron Structures of Films and Pigments,” Chem Rev. 99, 1963–1982 (1999). [CrossRef]

]. The flakes, which are responsible for the observed color shift, are not perfectly oriented within a horizontal plane, contributing to the non uniform texture when observed from a distance of around one meter or less. The characterization of this non uniform texture is getting more important for automotive industry because of the high percentage of cars with effect coatings [4

4. C. C. Esposito, “Metallic pigments: New finishes on the rise,” Coat World 8, 21–23 (2003).

] and because improvements in imaging technology allows this effect to be measured by, for instance, commercial instruments as BYK-mac [5

5. E. Kirchner, G. J. van den Kieboom, L. Njo, R. Sùper, and R. Gottenbos, “Observation of visual texture of metallic and pearlescent materials,” Col. Res. Appl. 32, 256–266 (2007). [CrossRef]

]. A typical problem is, for instance, matching the appearance of the repaired part of a car after an accident.

The non uniform texture of effect coatings is usually described as the composition of two distinctly separated perceptions: sparkle (or glint) and graininess (or diffuse coarseness) [5

5. E. Kirchner, G. J. van den Kieboom, L. Njo, R. Sùper, and R. Gottenbos, “Observation of visual texture of metallic and pearlescent materials,” Col. Res. Appl. 32, 256–266 (2007). [CrossRef]

]. The former refers to the star-like appearance of luminous points on a dark background, and it is observed under the extreme condition of very directional illumination. On the other hand, graininess refers to the granular appearance that it is seen under other extreme condition, that is, under diffuse illumination. When any of these conditions is not completely fulfilled an intermediate texture is observed. These both extreme patterns are very different and only separated phenomenological models to characterize them have been proposed so far in the literature [6

6. I. van der Lans, E. Kirchner, and A. Half, “Accurate appearance-based visualization of car paints,” Proceedings of the CGIV conference (Amsterdam, May 2012) 17–23.

, 7

7. E. Kirchner and J. Ravi, “Predicting and measuring the perceived texture of car paints,” Proceedings of the 3rd International Conference on Appearance “Predicting Perceptions” (Edinburgh, April 2012) 17–19.

], although the physical origin of both sparkle and graininess are obviously the reflection and transmission characteristics of flakes. Physical- and statistical-based Ershov’s model [8

8. S. Ershov, A. Khodulev, and K. Kolchin, “Simulation of sparkles in metallic paints,” Proceeding of Graphicon (August, 1999) 121–128.

] allows sparkle to be simulated, but it is too complex to be used for sparkle characterization and it does not consider graininess.

2. Description of the model

The proposed model is based in the hypothesis that the bright spots in sparkle patterns and the grains in graininess patterns are both produced by the luminous flux specularly reflected in flakes. Since these flakes are not perfectly horizontally arranged in the bulk, the proportion of this luminous flux collected by the entrance pupil varies depending on the flake, giving rise to the intensity variation on the image plane.

The value of the collected luminous flux from a single flake can be calculated considering the geometrical and radiometric variables represented in Fig. 1.

Fig. 1 Geometrical and radiometric variables involved in the calculation of the collected luminous flux from a single flake.

The coating is illuminated by a source S of radius RS at a distance dS and from a direction given by the spherical coordinates (ϑS, φS) or by the unit vector rS respect to the reference system, defined by its normal unit vector N. The higher the ratio RS/dS the more diffuse is the illumination on the surface. Reciprocally, the observer O is located at a distance dO, the radius of its entrance pupil is RO and its direction is (ϑO, φO) or rO. The half-angles of the solid angle of illumination and observation (αS and αO, respectively) are calculated from the corresponding radius and distances. The orientation of a given flake inside the coating is given by (ϑF, φF) or by its normal unit vector NF. It is assumed that the size of the flake is very small respect to the dimensions of the illumination and observation geometries.

The luminous flux collected by the entrance pupil (Φep) (confined within the cone Ωep) is composed by two different fluxes: The luminous flux specularly reflected by the flake (Φep,sp) and the luminous flux directed to the entrance pupil from other reflection mechanisms (Φep,dif), for instance, diffuse light from the bulk, specular reflection on the clear coat or light from several reflections between metallic or interference pigments.

The luminous flux specularly reflected by the flake (Φsp) is proportional to the incident luminous flux Φi for a given illumination/observation geometry. The proportional constant, hereafter called ρsp, is wavelength dependent and includes Fresnel coefficients, absorption through the bulk and size of the flake.

Φsp is confined within a cone Ωsp centered around the specular direction of the flake, with a half-angle αsp close to αS, lying the slight difference in the fact that the flake is not completely flat. Therefore, a parameter denoted flatness (f) is added to the model, whose effect is equivalent to modify αS (the directionality of the illumination):
αsp=fαS
(1)

The proportion of Φsp arriving to the entrance pupil of the observer Φep,sp depends on the angular distance δ between the direction of Φsp (rsp) and the observation direction rO. It is calculated from the unit vector rO and the unit vector rsp:
cosδ=rOrsp
(2)
where rsp is derived, by previously calculating r′sp from the relation:
rS+rsp=(2NFrS)NF
(3)
where r′S represent the direction of the illumination from the flake position (within a medium of refractive index n) and, in the same way r′sp represent the direction rsp from the flake position (see Figure 1). The transformation from rS to r′S is accomplished by transforming the polar spherical coordinate according to Snell’ law:
ϑS=arcsin(sinϑSn)
(4)
and similarly to transform from r′sp to rsp:
ϑsp=arcsin(nsinϑsp)
(5)

Φep,sp is proportional to the area of intersection I between the circular section of the cone Ωep and the circular section of the cone Ωsp. This area was accounted as the intersection of two circles of radii S and αO separated a distance δ. This intersection can be calculated using the formula for the area Acs of the circular segment of radius R and chord-center distance h:
Acs(R,h)=R2arccos|h|R|h|R2h2
(6)

Given two circles 1 and 2 with radii R1 and R2 with centers located at (0,0) and (0,D), their respective h1 and h2 are deduced as:
h1=D2+R12R222D;h2=D2R12R222D
(7)

If R1 > R2 and the circumferences of both circles intersect, the intersection of these two circles is:
I(R1,R2,D)={1πR22[Acs(R1,h1)+Acs(R2,h2)]ifh201πR22[Acs(R1,h1)+πR22Acs(R2,h2)]otherwise
(8)
where I is normalized to have the unit as maximum.

A similar calculation can be done if the source is not circular but rectangular.

Bearing in mind these expressions, it can be written for the proposed model:
Φep,sp={I(fαS,αO,δ)ΦspiffαSαOI(αO,fαS,δ)Φspotherwise
(9)

Therefore, the total luminous flux collected by the entrance pupil from the direction of a single flake can be written as:
Φep=Φep,dif+Φep,sp=Φi[ρdif+ρspI(fαS,αO,δ)]
(10)

3. Rendering texture images

The texture can be rendered by calculating Φep at different positions of the focal plane of the receptor, assuming a certain orientation distribution of the flakes. In addition, the Point Spread Function (PSF) of the eye can be applied to resemble the diffraction-limited effect of its optical system. A nonlinearity or a saturation function can be used to resemble the acquisition by a camera with saturation at some extent.

Application of Eq. (10) is shown in Fig. 2. The resolution of these images is 500 pixels × 1000 pixels. Since the model for the collected flux from a flake is applied for every pixel, the scale of the image is as one pixel representing one flake. This way, for instance, for an average size of flake of 50 μm, the image size is 2.5 cm × 5 cm. This is only an idealization, since flakes are not equally separated each other. We assume that this simplification does not change in great extent the texture effect. The rendering time for one image is around 0.9 s, operated by Matlab in a conventional workstation (2.6 GHz).

Fig. 2 Sparkle/graininess patterns as a result of applying Eq. (10) to a surface with flakes oriented at different directions. αS increases from top-left to bottom-right, whereas the other variables are kept constant.

It was assumed that both observer and illumination are far away enough to consider that all flakes in the image are observed and illuminated at approximately the same angles, allowing these angles to be taken as constants. A slightly more complex way of modeling these images is required if this assumption is not valid.

The orientation distribution of the flakes is taken as Gaussian for the angle of their surfaces and the horizontal (ϑF) (with a standard deviation of σF) and uniform regarding azimuthal orientation (φF).

Images shown in Fig. 2 were created using the following fixed values: ϑS = 20°, φS = 0°, ϑO = 25°, φO = 180°, RO/dO = 5 × 10−3, σF = 7°, ρsp/ρdif = 50, n = 1.5 and f = 1. The value of αS is doubled for every image, from top-left to bottom-right, from 0.11° to 14.4° (values shown on the top of each image). The calculated images were convolved with a Gaussian PSF with a standard deviation of 8 pixels (0.4 mm). It is clearly observed the transition from sparkle texture to graininess texture when αS is increased (from very directional to less directional illumination), according to the appearance of effect coatings.

4. Discussion

The presented model explains sparkle and graininess as two extremes of the same reflection phenomenon. If the sparkle spots are clearly distinguishable in a specific observation condition, the sparkle texture is observed, whereas if they are not, graininess can be perceived, but not always. Two conditions have to be fulfilled to clearly distinguish sparkle spots. First, a meaningful proportion of them has to be resolvable, that is, the mean distance between them must be at least similar to their apparent size in the image, defined by the PSF. Second, contrast C between sparkle spots and background has to be high enough. We defined this contrast as the increase of luminous flux produced by a very bright sparkle spot (Φep,sp) respect to a constant level or background (Φep,dif):
C=max(Φep)Φep,difΦep,dif=max(Φep,sp)Φep,dif=ρspρdif
(11)
Contrast C depends on illumination/observation geometry and on the wavelength via ρsp and ρdif.

Proper contrast allows the texture to be observed, but the density of sparkle spots determines whether it appears as sparkle or graininess. Graininess is perceived when this density is high and the resolvability condition is not completely fulfilled, whereas sparkle corresponds to the case of low density of sparkle spots.

The density of sparkle spots in the image depends on the illumination/observation geometry, and on the parameters and orientation distribution of the flakes. The more flakes directing specular reflection on the flake in the entrance pupil, the more sparkle spots, which is favored with a wide distribution of illumination and observation directions (higher αS and αO). As a matter of fact, illumination and observation solid angles are exchangeable in their effects, and the smaller one limits the sparkle effect due to the variation of the other. The closer the observation direction to the specular direction, the higher the density of sparkle spots too, as long as the average of the orientation distribution of the flakes is parallel to the coating’s normal.

When the illumination solid angle is higher than the entrance pupil solid angle, the density of sparkle spots for a given geometry can be formally defined as the probability that the luminous flux specularly reflected by any flake was partially or completely collected by the entrance pupil. Using the proposed model it is possible to calculate it for any condition in percentage units by counting the pixels with Φep,sp > 0. The density of sparkle spots was calculated by the model and plotted in Fig. 3 for values of αS between 0° and 30°, and using the values quoted in section 3 for the other variables. It increases toward less directional illumination, as expected. Notice that the image corresponding to αS = 3.7° in Fig. 2, with graininess texture, has a density of sparkle spots of less than 10 %.

Fig. 3 Density of sparkle spots calculated by the model for values of αS between 0° and 30°.

5. Conclusions

Despite its simplicity, this model includes the essential concepts and parameters to understand sparkle/graininess of effect coatings and should be very helpful for its radiometric characterization and for the design of sparkle/graininess psychophysics experiments.

Acknowledgments

Authors are grateful to “ Plan Nacional de Física” for funding this work FIS2010-19756-E, to CSIC’s JAE Program and ” European Social Fund” for awarding us a research trainee.

References and links

1.

F. J. Maile, G. Pfaff, and P. Reynders, “Effect pigments–past, present and future” Progress in Organic Coatings 54, 150–163 (2005). [CrossRef]

2.

H. J. Streitberger and K. F. Dössel, Automotive Paints and Coatings (Wiley-VCH, Weinheim, 2008). [CrossRef]

3.

G. Pfaff and P. Reynders, “Angle-Dependent Optical Effects Deriving from Submicron Structures of Films and Pigments,” Chem Rev. 99, 1963–1982 (1999). [CrossRef]

4.

C. C. Esposito, “Metallic pigments: New finishes on the rise,” Coat World 8, 21–23 (2003).

5.

E. Kirchner, G. J. van den Kieboom, L. Njo, R. Sùper, and R. Gottenbos, “Observation of visual texture of metallic and pearlescent materials,” Col. Res. Appl. 32, 256–266 (2007). [CrossRef]

6.

I. van der Lans, E. Kirchner, and A. Half, “Accurate appearance-based visualization of car paints,” Proceedings of the CGIV conference (Amsterdam, May 2012) 17–23.

7.

E. Kirchner and J. Ravi, “Predicting and measuring the perceived texture of car paints,” Proceedings of the 3rd International Conference on Appearance “Predicting Perceptions” (Edinburgh, April 2012) 17–19.

8.

S. Ershov, A. Khodulev, and K. Kolchin, “Simulation of sparkles in metallic paints,” Proceeding of Graphicon (August, 1999) 121–128.

9.

F. E. Nicodemus, J. C. Richmond, and J. J. Hsia, “Geometrical considerations and nomenclature for reflectance,” Natl. Bur. Stand. Monogr. 160, (1977).

OCIS Codes
(120.5700) Instrumentation, measurement, and metrology : Reflection
(240.6700) Optics at surfaces : Surfaces

ToC Category:
Thin Films

History
Original Manuscript: July 29, 2013
Revised Manuscript: September 19, 2013
Manuscript Accepted: September 27, 2013
Published: October 30, 2013

Citation
A. Ferrero, J. Campos, A. M. Rabal, and A. Pons, "A single analytical model for sparkle and graininess patterns in texture of effect coatings," Opt. Express 21, 26812-26819 (2013)
http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-21-22-26812


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. F. J. Maile, G. Pfaff, and P. Reynders, “Effect pigments–past, present and future” Progress in Organic Coatings54, 150–163 (2005). [CrossRef]
  2. H. J. Streitberger and K. F. Dössel, Automotive Paints and Coatings (Wiley-VCH, Weinheim, 2008). [CrossRef]
  3. G. Pfaff and P. Reynders, “Angle-Dependent Optical Effects Deriving from Submicron Structures of Films and Pigments,” Chem Rev.99, 1963–1982 (1999). [CrossRef]
  4. C. C. Esposito, “Metallic pigments: New finishes on the rise,” Coat World8, 21–23 (2003).
  5. E. Kirchner, G. J. van den Kieboom, L. Njo, R. Sùper, and R. Gottenbos, “Observation of visual texture of metallic and pearlescent materials,” Col. Res. Appl.32, 256–266 (2007). [CrossRef]
  6. I. van der Lans, E. Kirchner, and A. Half, “Accurate appearance-based visualization of car paints,” Proceedings of the CGIV conference (Amsterdam, May 2012) 17–23.
  7. E. Kirchner and J. Ravi, “Predicting and measuring the perceived texture of car paints,” Proceedings of the 3rd International Conference on Appearance “Predicting Perceptions” (Edinburgh, April 2012) 17–19.
  8. S. Ershov, A. Khodulev, and K. Kolchin, “Simulation of sparkles in metallic paints,” Proceeding of Graphicon (August, 1999) 121–128.
  9. F. E. Nicodemus, J. C. Richmond, and J. J. Hsia, “Geometrical considerations and nomenclature for reflectance,” Natl. Bur. Stand. Monogr.160, (1977).

Cited By

Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.

Figures

Fig. 1 Fig. 2 Fig. 3
 

« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited