Temukan buku favorit Anda berikutnya

Jadilah anggota hari ini dan baca gratis selama 30 hari
Colour Measurement: Principles, Advances and Industrial Applications

Colour Measurement: Principles, Advances and Industrial Applications

Baca pratinjau

Colour Measurement: Principles, Advances and Industrial Applications

840 pages
Aug 31, 2010


The measurement of colour is important in many commercial operations and professions, such as bleaching and colouration of textiles, applications of paints, dentistry and colouration of food products. This book will discuss colour measurement theories, the latest technological and scientific developments of measuring colour and the applications of colour measurement.

Part one reviews the underlying theories, principles and methods of how to measure colour. It includes topics such as expressing colours numerically, camera based colour measurement, colour shade sorting and determining and improving the accuracy of colour measurement. Part two presents a selection of industrial applications illustrating the use of colour measurement in textiles, paint, teeth, hair and food.

With its international range of contributors, Colour measurement: Principles, advances and industrial applications is beneficial to a variety of readers such as colour technologists, colour quality inspectors, product developers, dentists, cosmetologists and anyone who uses colour in their work. It will also be a valuable reference for academics and students studying design, fashion or colour related subjects.
  • Discusses colour measurement theories and the latest technological and scientific developments of measuring colour
  • Case studies illustrate camera based colour measurement and review visual and instrumental evaluation of whiteness and yellowness applications in industries including cosmetics and dentistry
  • Motivations for colour measurement are explored to answer questions raised as to why colours do not match and explain factors such as wet and dry fabric differences
Aug 31, 2010

Terkait dengan Colour Measurement

Judul dalam Seri Ini (173)
Buku Terkait
Artikel Terkait

Pratinjau Buku

Colour Measurement - Elsevier Science


Part I

Theories, principles and methods of measuring colour


Colour vision: theories and principles

V. Viqueira Pérez, D. De Fez Saiz and F. Martinez Verdú,     University of Alicante, Spain


When viewing any scene, the human visual system is able to extract information regarding light wavelength, which is why we see in colour. This chapter discusses the mechanisms of human colour vision. The chapter first reviews the anatomy and the physiology of the visual system, and then describes the generic ATD models of colour vision. From these models, the chapter discusses the topics of colour appearance, colour constancy, and defective colour vision.

Key words

ATD models

colour appearance

defective colour vision

colour constancy

mechanisms of chromatic adaptation

1.1 Introduction

1.1.1 Advantages of colour vision

When viewing any scene, the human visual system is able to extract information regarding light wavelength, which is why we see in colour. But what advantages does this ability give to human vision? In evolutionary terms, the advantages for our ancestors are clear: seeing in colour makes it easier to detect food, such as the colour of fruit against the green of a leafy background, and the ability to detect animals hidden from view, be they predators or possible prey.

Today, considering the fact that people with chromatic anomalies or defi ciencies are able to lead a normal life, we would tend to think that colour vision is not a relevant factor. But consider the sheer amount of information that surrounds us, and how much of it is based on colour – is it not surprising to realise that most information is actually colour-coded? Traffic signs, advertising, graphic design, the internet. … Not only will people who have difficulties seeing in colour not be qualified for certain jobs, it could also mean that they do not properly recognise information surrounding them in their normal life.

1.1.2 Anatomy and physiology of the human visual system (rods; L, M and S cones LGN and cortical areas)

The visual system is the part of the brain that makes us see. It does this by interpreting information from available light to build a representation of the outside world. The process begins when light passes through the eye’s transparent elements and hits the retina. All the elements involved before the retina form the ocular optic system (optically, the eye consists of a series of refractive surfaces defined by transitions between air, fluid and solid tissues, the study of which is not the purpose of this book).

The human eye is roughly spherical in shape (Fig. 1.1), and is made up of three distinct layers of tissue: sclerotic coat, choroid coat and retina. The sclerotic coat is the outer layer. It is white and extremely tough, except in the front where it forms the transparent cornea which contributes to the image-forming process by refracting light entering the eye. The surface of the cornea is kept moist and dust-free by the secretion from the tear glands. The intermediate layer is the choroid coat. This layer is deeply pigmented with melanin that reduces reflection of stray light within the eye. The choroid coat forms the iris, a diaphragm of variable size whose function is to adjust the size of the pupil to regulate the amount of light admitted into the eye. The pupil contraction is under the control of the autonomic nervous system: in dim light, the pupil opens wider letting more light into the eye; in bright light the pupil closes down. The retina is the inner layer of the eye. It contains the light receptors, the rods and cones.

1.1 Human eye structure.

Inside the eye, the cavity between the lens and cornea – called the anterior chamber – is filled with a gel-like fluid called aqueous. The lens is located just behind the iris. The lens is a flexible unit that consists of layers of tissue enclosed in a tough capsule. It is suspended from the ciliary muscles by the zonule fibres. Behind the lens is the vitreous: a thick, transparent substance that fills the back of the eye. It is composed mainly of water and comprises about two-thirds of the eye’s volume, giving it form and shape. The next element is the retina, a neurosensory layer that initiates the neural processes of vision.

An inverted image of the outside world is projected onto the retina, and the visual system then has the complex task of rebuilding a three-dimensional image from this two-dimensional projection. The end result of this process is visual perception.

The retina is in fact a prolongation of the brain. Anatomically, it is structured in ten layers for different types of neurones: photoreceptors (cones and rods), which are in fact modified neurones, and horizontal, bipolar, amacrine and ganglion cells.

However, the photoreceptor’s mosaic is not evenly distributed along the retina. In physiological terms, it has a central zone and an outer area. The central zone includes the fovea, a small pit that ensures best visual acuity. The fovea is located in an area known as the macula lutea (Latin for yellow spot), which takes its name from the yellow pigment that covers it. The central retina includes another region, the optic disc, at the exit of the optic nerve, formed by the ganglion cell axons leaving the eyeball to form the optic nerve. This elongated pink disc is located in the nasal area and usually measures around 1.5 mm². It has no photoreceptors, and as a result forms a blind spot in our vision, where we cannot see (since there are no cells to detect light, this part of the field of vision is not perceived). The rest, named the peripheral retina, has a low concentration of photoreceptors.

The retina is inverted, in such a way that light passes through its entire structure before hitting the photosensitive layer. A photoreceptor is a specialist neurone capable of performing phototransduction, by which light is converted into electrical signals, which are then transmitted from neurone to neurone. Cones and rods are anatomically and physiologically different. Cones are much more abundant in the central retina, so the fovea contains only cones and, more importantly, the cones connect with the bipolar and ganglion cells at a proportion of 1:1:1, whereas in the peripheral retina area, several hundred or thousands of cones converge to a single bipolar cell. This explains why visual acuity is so much sharper in the fovea than at any other point on the retina.

In terms of light sensitivity, rods can function in low light, whereas cones need much higher levels of light. So, rods are responsible for night vision, and cones are used for daytime vision, seeing colours and visual acuity.

When light hits a photoreceptor, it sends a proportional synaptic response to a bipolar cell, which in turn sends the signal to a ganglion cell. Furthermore, there are two substrata at a synaptic level, with horizontal cells and amacrine cells (see Plate I in colour section between pages 42 and 43). These modify the initial signal to an extent that, after two or three synapses, the signal contains much more complex information than a simple point-to-point visual representation. From the ganglion cells, the signal reaches the appropriate left and right lateral geniculate nuclei (LGNs). The function of the LGNs is complex. We know that there is a three-system separation of fibres (PC cells, MC cells and KC cells), relating to Magno, Parvo and Konio systems. The LGNs have six layers, numbered from 1 (innermost layer) to 6 (outermost layer). Layers 1 and 2 contain the largest cell bodies and constitute the magnocellular system, whereas layers 3, 4, 5 and 6, which have smaller neurones, are part of the parvocellular system. Layers 2, 3 and 5 receive fibres from the homolateral eye, and layers 1, 4 and 6 from the contralateral eye. Between each layer are the KC cells, which belong to the koniocellular system. Eighty per cent of the cells form part of the parvo system, ten per cent belong to the magno system, and ten per cent to the konio system.

The LGN signals are sent to the primary visual cortex (V1), located at the back of the brain. The V1 fibres are then projected to area V2, and from V2 to V3, V4 and V5. Another important fact is that more than half of the LGN and V1 neurones process information from the fovea. A process therefore exists that gives priority to the part of the scene that is projected onto the fovea (the fixation point).

1.2 Human colour vision

1.2.1 Chromatic stimulus and perceived colour

When we observe a scene, our visual system creates a perception of the outside world from the radiating energy that reaches our eyes from the objects within that scene. One of the elements of that perception is colour. But where is the colour? Is it external, or is it inside our visual system? The answer is clear: there are different chromatic stimuli in any scene that we view, and our visual system is able to capture that information, relating to the wavelength of light, which is how colour ‘emerges’. Colour, therefore, is something internal. Colour is perception.

The chromatic stimulus is electromagnetic radiation from sources and objects that hits the optic system and triggers the visual process. Perceived colour, therefore, is a sensation produced by the chromatic stimulus that makes it possible to differentiate that stimulus from others with the same area, duration, shape and texture.

As we shall see, chromatic stimuli have three variables, which are directly related to three perceptual variables.

1.2.2 Models of chromatic vision

The trichromatic theory (Young–Helmholtz–Maxwell)

In the early nineteenth century, Thomas Young (1801) suggested that the retina contained three types of nerve fibres that can be stimulated to a greater or lesser extent by the different wavelengths that correspond to red, green and violet colours. Several years later, James C. Maxwell would demonstrate that any colour in the spectrum could be matched with three monochromatic primary colours: red, green and blue. These two ideas were concurrent, and gave a clear, simple explanation of colour vision. Thus, a colour that stimulates the red and green particles in equal measure will be perceived as yellow; if it stimulates green and violet, we see blue. Chromatic vision is simply a matter of additive mixing. At around the same time, the German physicist Hermann V. Helmholtz also suggested that subjects with deficient colour vision (dichromats) had reduced forms of normal vision, which lacked one of the three receptor types. This explains the fact that dichromats accept as equal the same kind of matchings as do subjects with normal vision.

However, there were weak points in the theory in terms of colour appearance: if everything can be explained through mixes of colours, why are there no reddish greens or bluish yellow colours?

Theory of chromatic opponency

Ewald Hering (1872–4) was more concerned with the appearance of colours. In his reports to the Imperial Academy of Sciences in Vienna, Hering proposed the existence of three opponent processes generated at some point in the visual process: red-green, blue-yellow and black-white mechanisms.

The idea of opponency is that the red-green categories (and, similarly, the blue-yellow and light-shadow categories) are opposed and represent two extremes of variations in a continuum. More red necessarily implies less green. Despite having no experimental proof, Hering challenged both the established trichromatic theory and the prevailing physiological theory (Müller’s doctrine of specifi c nerve energies) head-on: ‘a nerve fibre is capable of conducting only one type of qualitative information’. A nerve fibre, therefore, could not transmit both red and green information, as they are qualitatively different.

These two theories, which in principle seemed to take opposing views, are in fact compatible; the three retina cone types would in fact correspond to trichromacy, and the sensor responses by these cones would then be combined in the three red-green, blue-yellow and light-shadow mechanisms (although, as we shall see, dark-light is an additive rather than an opposing mechanism). It was not until 1957, a hundred years later, that D. Jameson and L. M. Hurvich would prove the existence of the opposing mechanisms. Paradoxically, Hering himself had suggested that these two theories did not have to be necessarily incompatible.

Today we know that the parvo cells are responsible for the red-green mechanisms (L – M cells and M – L cells), and the konio cells are responsible for the blue-yellow mechanisms S – (L + M) cells. Magno cells, on the other hand, are not opposing cells, but rather add the inputs from the L, M and S cells. Therefore it is an additive mechanism.

ATD models (two-stage models)

Each of these previous theories can explain a series of phenomena: the trichromatic theory reproduces many psychophysical experiments, but it does not predict appearance, which the colour opponency theory does. Combining the two theories leads to models that combine an initial trichromatic stage at the receptor level (first stage), with a stage of neural processing ruled by colour opponency mechanisms (second stage), comprising a non-opposing luminance mechanism (A) and the two opposing chromatic mechanisms: green-red (T) and blue-yellow (D). To a certain extent, the CIELAB values, which are very useful for industrial colorimetry, represent these concepts as Lightness (L) and can be linked to the achromatic channel (A); the chromatic coordinate a* to the greenness-redness channel (T) and the chromatic coordinate b* to the yellowness-blueness channel (D).

The model as a whole is summed up in an ATD matrix resulting from the combination of the L, M and S cone signals, as shown in equation 1.1. The three independent channels are listed as A (achromatic), T (tritan) and D (deutan):


The Boynton model (1986) is a good representation of these two-stage models. It proposes three cone types, the spectral response curves of which are the Smith and Pokorny fundamentals. The coefficients for the resulting matrix are shown in equation 1.2, and the general functioning for this model is summarised in colour Plate II. The illustration also shows how the S cones contribute to the achromatic channel, though this value is very low and in the matrix is worth zero.


Why does the matrix have these coefficients? If we look at the single 575 nm yellow in the Smith and Pokorny fundamentals (Fig. 1.2), (L – M) would be a positive value, i.e. red would predominate over green, and we would see a reddish orange colour. However, the M cones’ response becomes balanced by applying a factor of two, so that in 575 nm they are cancelled out (T = 0) and a pure yellow colour is perceived. The red-green channel gives no signal (T = 0), and the blue-yellow channel gives a response tending towards yellow.

1.2 Smith and Pokorny fundamentals.

Another justification for this factor of two is in the proportion of cones in the retina: L:M = 10:5. According to this model, there is also a small intrusion of the S cones in the RG channel, though generally it is not taken into consideration.

The oponnent process yellow-blue receives an amplified signal with a negative sign from relatively small values of the S cone, and a positive signal from the L and M cones: (L + M) – S. If S > L +  M, the channel’s response is negative and is interpreted in upper states of the visual process as blue. But if S < L + M, the channel’s response is positive and the colour is interpreted as yellow. For lengths above 520 nm, the S cones give no signal; the channel signal will always be positive and interpreted as yellow. Finally, we can state that the L + M signal is transmitted by different nerve fibres to those that transmit the L – 2 M signal.

Almost all chromatic vision models largely coincide in this general structure that we have just seen for the Boynton model: there are two opposing channels (channels T and D), and one additive channel (no opponent) for luminance (channel A). The differences are in the matrix coefficients. However, there are certain points that have not been fully clarified:

Channel T: There is L – M opponency, but there are doubts over contribution from the S cones. If there is a contribution, it would be of the (L + S) – M type.

Channel D: There is L – S opponency, but there are doubts over contribution from the M cones to this channel.

Channel D: The achromatic channel is L + M, but there are doubts over contribution from the S cones to this be of the L + M + S additive type. Nevertheless, most of the experimental results point to such a contribution not existing.

Other more complex models, also exist, such as the De Valois-De Valois (1993) model. This considers a second opposing stage of cones (ATDinterm) that generates three colour opponent signals (rather than two) at the LGN cell level, corresponding approximately to L – M, M – L and S – (L + M). A third perceptual (linear) opposing stage would then occur, generating one achromatic and two chromatic channels. Finally, there are other non-linear models that succeed in explaining effects resulting from adaptation to a light source, and the influence of the background and the environment surrounding the stimulus, etc. In this case, non-linearities need to be introduced into the ATD response calculation. The 1990 Guth model, and subsequent modifications up to 1995, are a good example of these models.

The general outline for these models is as follows (equation 1.3): from LMS an initial non-linearity is introduced, obtaining LMSs that are adapted to those conditions. Following this, the ATD exchange matrix and a second non-linearity are applied to obtain the adapted ATDs.


As always, the objective with this model is to explain chromatic vision by reproducing the behaviour of the physiological A, T and D mechanisms.

1.3 Chromatic perception

1.3.1 Chromatic discrimination

Colour differences are of enormous importance in industry. Understanding the human visual system’s degree of tolerance to colour differences is fundamental in knowing the maximum tolerable error in formulating a paint or printing a fabric. The most important aspect is not in fact the differences of the various components of colour (hue, colourfulness and lightness), but rather the perceptive differences of colour as a whole.

In 1934, Wright evaluated colour differences with constant illumination by means of a 2° bipartite field, with 100 td retinal illumination throughout the whole spectrum and for five directions of the chromaticity diagram. In this way, he determined intervals or segments with edges that represent a constant difference in chromaticity. As a result of this experiment, it was deduced that the CIE1931xy space had very little uniformity, as two very distant points in the area of greens differ in colour just as two very close points in the area of blues and purples do. In 1942, MacAdam took these studies further. Using a bipartite field, a fixed reference stimulus was placed in one of the two fields, and colour matching was carried out in the other; the test was for 2° and was surrounded by an adaptation field of white light of the same luminance (200 td). After 50 matchings, two symmetrical points were defined in the same direction. By repeating this same operation for various directions and joining the points (within the standard deviations), an ellipse was generated. MacAdam showed the results in the CIE1931xy space for 25 reference colours.

The result of joining the experimental colour discrimination points is an ellipse. These ellipses are of different sizes, orientation and shape throughout the chromaticity diagram. The ellipses are smaller in the area of blues, of medium size in the reds and larger in the greens. By considering the distance from the centre of the ellipse to the edge as the unit of colour difference, it can be confirmed that this distance is not the same from one ellipse to another; it can thus be deduced that the CIE1931xy space is not uniform. In a space of uniform representation, representing colour differences would obtain circumferences of equal radius, regardless of the centre colour chosen, rather than ellipses of different sizes and orientation. As a result of these works, attempts have been made to find a more uniform colour representation space. The most commonly used are CIELAB, CIELUV, Guth’s ATD and SVF, among others.

1.3.2 Appearance of colour: chromatic effects

Colour is trivariant, which means that in an isolated colour stimulus, we can distinguish three separate qualities: hue, brightness and saturation. When one studies a related colour, i.e. one that forms part of a scene, the descriptors refer to the environment, and one thus speaks of hue, lightness and colourfulness. Similarly, any colour is defined by means of three physical variables that correspond to these three perceptive variables: wavelength to hue, luminance to brightness, and colorimetric purity to saturation (see Table 1.1). However, these relationships do not display absolute independence, but rather there is interference between parameters. A variation in a single parameter (L, Pc, λ) can affect not only the perceptual attribute to which it is associated, but the other two as well. On the other hand, environmental modifications also affect colour. These effects cannot be explained by a linear model, as occurs with characterisation by three-way stimulus values or by means of opponency. We thus know that there must be some subsequent process, and that it cannot be linear.

Table 1.1

Physical and perceptual colour descriptors for isolated and related colours

Studies of these chromatic effects are many and varied, and are beyond the scope of this work; we can merely provide a brief explanation of some of the better known and more widely considered of these chromatic effects.

 Bezold-Brücke effect: ‘A variation in luminance can alter the tone, thus changing its colour appearance.’ Bezold and Brücke (1873 and 1878) discovered this effect independently, and both indicated that at high luminance levels, reds and yellowish greens tended towards yellow, whereas greenish blues and violets became blue. However, they also confirmed that there are three hues that do not vary: yellow (571 nm), green (506 nm) and blue (474 nm). This effect shows that tone and luminosity attributes are not completely independent.

 Aubert-Abney effect: ‘By adding white to a purple or a monochromatic colour, not only is saturation reduced, but hue changes also occur.’ This effect can be seen in Munsell’s Atlas colour samples, with equal tone and different chroma in the xy diagram. Theoretically, they should be straight lines of constant λd, but instead a curve is obtained that becomes greater the nearer the samples are to the locus. Again, there are some exceptions to this behaviour: for 572 nm yellow and for a purple (0.240,0.035) (λc = 559 nm).

 Helmholtz-Kohlrausch effect: ‘In heterochromatic Matching and with identical luminance, Brightness varies with the chromaticity of the stimulus.’ With identical luminance levels, achromatic colours display lower luminosity than chromatic colours. And equally, the other way round, with identical luminosity levels, achromatic colours display greater luminance than chromatic colours. In any matching of a white with any colour, then (Lwhite / Lcolour) > 1 applies. This ratio nears 1 when the colour is more desaturated, and reaches values higher than 1.6 for monochromatic colours.

 Simultaneous contrast: An increase in the brightness of a stimulus with the simultaneous decrease of background luminance. A dark environment makes one grey seem lighter than another with a white environment.

 Crispening effect: An increase in the difference between two stimuli occurs if the background is similar to the stimuli; in this case the brightness differential threshold is lower.

 Apparent mix: When the spatial frequency of an object is high, a chromatic mix of background and stimulus is produced. This colour quality is applied in artistic painting, with its maximum expression in the pointillist movement. George Seurat used this technique; his brushstrokes are tiny points, with no merging between them on the canvas, yet when viewed from a certain distance they create the desired combinations on the retina. The same occurs with the pixels of a television screen.

1.4 Defective colour vision

1.4.1 Anomalies and deficiencies in colour vision

Most people see colour in the same way, which we can call normal chromatic vision. However, other people’s sight behaves abnormally. In most of these cases, they are able to differentiate colours, but their chromatic vision is much poorer than that of someone with normal sight. There are many colours which for a normal observer are clearly different and which the person with chromatic deficiency views as the same.

The reason can be explained with any ATD model, by considering the hypothesis that chromatic deficiency is due to the lack of one of the three cone types. This lack of a cone type means a failure in either the T or the D chromatic mechanisms.

Consider the Boynton model. The chromatic channels are T = L – 2 M and D = (L + M) – S. If one of the two terminals is cancelled out, the channel response will always be of the same sign. What happens if the L cone is missing? Channel D would function, but in an anomalous way (reduced to D = M – S), and its response would differ to that of the normal channel, but a positive or negative response would still be possible; however, channel T (T = – 2 M) would always respond in the same direction. Clearly, chromatic vision would thus be greatly reduced. In the hypothetical case of two cone types being missing (L and M), neither chromatic channel could function, and the subject would have no chromatic vision of any kind. A person lacking one cone type is known as a dichromat. Given that there are three types (L, M and S), there can be three separate categories, known as protanopia (lacking L cones), deuteranopia (lacking M cones) and tritanopia (lacking S cones). Individuals with these defi ciencies have sight that differs between the three types. Table 1.2 includes all types of reduced colour vision currently reported.

Table 1.2

Types of colour vision depending on the cones

As well as chromatic deficiencies, another anomaly type also exists: people who have all three cone types, one of which has a slightly displaced curve response for that pigment. In this case, the person will also suffer from poor colour discrimination. These cases are known as protanomalous, deuteranomalous and tritanomalous.

In people with protanomaly, maximum pigment L absorption is not at 560 nm, but instead is displaced slightly to shorter wavelengths and is thus closer to the M pigment. If we analyse the response to a red 650 nm wavelength, for example, sensitivity for this colour is reduced when compared with someone with normal sight, and as a result, in a yellow match with a sum of red and green, people with protanomaly have to add more red than a normal observer.

People with deuteranomaly suffer a similar problem, but in their case it is the M pigment that is abnormally displaced towards red. Maximum absorption is no longer 540 nm, but rather is displaced towards a high wavelength and is thus closer to the L

Anda telah mencapai akhir pratinjau ini. Daftar untuk membaca lebih lanjut!
Halaman 1 dari 1


Pendapat orang tentang Colour Measurement

0 peringkat / 0 Ulasan
Apa pendapat Anda?
Penilaian: 0 dari 5 bintang

Ulasan pembaca