Inspection of Micro Optical Effect Images in Thin Substrates

Technical Field

[0001] The invention relates to the production of optical devices of the type used to authenticate articles. In particular, the invention relates to testing micro optical devices for defects.

Background of Invention

[0002] To guard against counterfeits, security features are often applied to articles of value such as security documents or tokens. These security features will often generate an“optically variable effect”, which is an image that changes at different viewing angles. These types of security features are often applied to valuable documents such as identity cards, passports, credit cards, banknotes, cheques, vouchers and certificates to provide an overt feature that is difficult to replicate.

[0003] One category of these security features are lens-based devices that use an array of micro lenses to magnify an underlying array of micro images to generate the optically variable effects. These devices require precision manufacturing as small defects have significant impact on the optical effects generated. In view of this, the optical characteristics of security features will be monitored during production, typically by an automated vision system (i.e. camera and feedback control). However, checking the optical characteristics of a lens-based security device is problematic. In particular, it is difficult to measure the image contrast of magnified images produced by a micro lens array applied to thin substrates such as banknotes. Similarly, it is difficult to detect defects in the micro image array, particularly an array of micro-images applied to thin substrates such as banknote substrates. In the production environment, these difficulties are exacerbated due to the need for testing to occur at normal production speeds. Individual inspection of all micro images produced, also known as a 100% inspection rate, is exceptionally difficult to perform at production speeds. Even if 100% inspect rates were practical in the production environment, they do not ensure maximum quality or reduced spoilage due to image defects and/or poor image contrast.

[0004] Micro lenses on security documents such as banknotes have a lens focal length limited by the thickness of the security document. The transparent substrate of the banknote is often used as the optical spacer between the lens array and the micro images. The tiny micro lenses are normally embossed into a layer of clear lacquer material that is subsequently hardened using UV radiation. The layer of UV curable lacquer is relatively thin which in turn limits the maximum depth available for embossing the micro lenses. In practice there is a limit to the maximum lens‘sag’ (i.e. the height of the lens above its planar base) that can be embossed. There is also a corresponding limit to the width of the lens in view of the need to focus on the micro-images on the other side of the substrate.

[0005] The limited width of the lenses on a thin substrate places limits on the size of the micro images on the other side of the substrate if optical effects such as moire magnification, image flips, animations, integral images and so on are to be achieved. In short, the geometry of the micro-lenses and the type of optical effect will determine the size of the micro-images.

[0006] Of all the different optical effects, the so called“contrast switch” optical effect involves the largest printed image element size. This is because the contrast switch optical effect requires the least amount of image information to be contained in the image layer beneath the micro lens array. Figure 1 is a diagrammatic representation of a contrast switch optical effect device. A single cell 10 (i.e. a single cylindrical lens 14 from the lens array and its corresponding image element 18 from the opposing side 16 of the transparent substrate 12) is shown being viewed from three different viewing angles (20, 22 and 24). The distance from the image plane 16 to the top of the micro lens 14 is less than the focal length of the lens 14. Each image element 18 is a printed line magnified by the corresponding lenticular lens 14 (from the one-dimensional lenticular lens array) to project a lenticular image pixel seen by the viewer.

[0007] The colour strength of the pixel seen by the viewer depends on the viewing angle as well as the amount of offset of the printed line 18 with respect to the central vertical plane (i.e. the plane through the vertex) of lens 14. At viewing angle 24, the colour strength is a maximum while at viewing angle 20 the colour strength is a minimum. This effect is optimised if the printed feature or element is sized to be approximately half the width of the micro-lens 14. Sufficient image contrast is also possible with a printed feature 18 more than, or less than half the width of the lens 14. However, the amount that the printed line 18 is larger or smaller than the optimal width (typically half the width of the micro-lens) determines the reduction in the maximum image contrast projected to the viewer. Therefore, a modest variation in the size or rather width of the image element 18 can greatly vary the contrast in the manufactured security device (in this case, the contrast switching device). This creates a perception of inconsistent product quality in the end user. Furthermore, a significant change in contrast can reduce the counterfeit resistance of the security device as relatively low quality counterfeits will replicate the same optical effect.

[0008] Similarly, missing portions of the image elements 18 will generate defects in the magnified image which creates the perception of inconsistent quality and hence lowers the inherent counterfeit resistance.

[0009] If a contrast switch optical device is incorporated onto a thin substrate (such as the transparent substrate of a banknote) with micro-lenses (in particular micro lenses around 64 microns wide), it will be necessary to control the printed image element size 18 to a width of about 30 microns to 40 microns. The measurement of lines this narrow is difficult to achieve towards a micron level accuracy using an inline camera based inspection system. Typically, a camera system such as this employs line-scan cameras that have a scanning resolution about 100 microns to 250 microns per pixel (although the actual resolution will also depend on the speed of the substrate web relative to the camera), and hence scanning resolution will decrease as the media feed speed increases. This means the smallest size change in a printed feature that can be theoretically detected (at a typical production speed of at least 100 metres per minute) is approximately 100 microns to 250 microns. This assumes that the repeating pattern of printed features are presented to the camera in exactly the same relative position. Unfortunately, in practice this is not possible. The continuous web of media printed in a gravure process can move up to +/- 0.25mm. This increases the smallest detectable size change in the printed features. Therefore, even for a simple micro optical effect such as a contrast switch, which uses the largest printed feature size of all micro optical effect devices, measuring the contrast of individual printed features using camera based inspection systems can be prone to significant inaccuracies.

[0010] Offline testing, using optical microscopes and the like, provide more accurate image width measurements but this does not prevent the production of printed material with contrast and defects outside the specified acceptance range. Offline inspection necessarily occurs after printing, whereas a 100% inspection technique operating inline provides corrective feedback during the print process.

[0011] Direct scanning or acquisition of the magnified images generated by the optical effect can be used to measure contrast. However, there is an inherent production tolerance varying the printed micro-imagery relative to the associated micro lenses creating a relative phase shift between the lenses and the imagery in each individual optical device. These variations cause corresponding variations in the magnified image projected to the camera system and therefore the acquired image cannot be meaningfully compared to a single reference image (a so called‘golden template’). In other words, even if the acquired image is printed with perfect contrast, it may well significantly differ from the reference image and therefore be falsely identified as spoilage.

[0012] Referring to Figures 2 and 3, a contrast switch optical effect of the design shown in Figure 2 produces the selection of projected images shown in Figure 3. The image is printed with perfect contrast. The image projected to the inspection camera can be the positive image 18, created when viewed from a first viewing angle and exhibiting a first contrast measurement, or it can be the negative image 26, created when viewed from a second viewing angle and which exhibits a second contrast measurement (complementary to the first contrast measurement). Likewise, it may be any intermediate transition image 28 to 34 that is generated from other viewing angles which exhibit a contrast that differs from that of the first or second contrast measurement. These inherent variations make attempts to compare the acquired image to a single golden template image pointless for the purposes of inspection. [0013] Similarly, using acquired camera images to detect image defects in individual print features such as missing image portions (due to missing ink, for example from dried ink clogging the cylinder cells) is also not possible due to variation in the acquired image from the inherent variation in the lens-to-imagery phase associated with production tolerances.

[0014] In an effort to mitigate these problems, it may be possible to use multiple cameras acquiring the projected magnified images from a range of different viewing angles. All the acquired images are compared to the golden template and if any one of the acquired images matches the golden template, the magnified image is classified as acceptable. However, this technique relies on capturing a sufficiently large number of images from different viewing angles to ensure that a viewing angle close to that of the golden template is captured. Capturing multiple images of individual print features from different viewing angles is complex, expensive and requires powerful processing. An effective inspection system of individual print features using multiple cameras would only be practical for printing processes that operate at low production speeds.

[0015] An alternative approach is to use a single camera view to acquire images of the micro-imagery that are then compared with multiple golden templates. This type of technique has been disclosed in US 5850466, entitled‘Golden Template Comparison for Rotated and/or Scaled Images’. While this technique reduces the number of cameras used, it still requires powerful image processors to compare the acquired image with each of the set of golden templates. Once again, this technique is only practical for processes operating at low production speeds.

[0016] In yet another approach to address the issue of contrast measurement, a printed swatch of colour is used as a control or reference for the magnified image contrast. For example, an optical device for producing a contrast switch effect printed via gravure on a banknote substrate can also simultaneously print a swatch of the micro image colour on the substrate. The control swatch brightness/colour density can then be measured using an inline camera based inspection system. The swatch colour density is correlated to the contrast of the magnified image to test whether contrast is within the acceptable range. However, in practice, any changes in the size of the micro- imagery that ultimately results in adverse changes to the magnified image contrast, are not correlated to changes in the colour density of the colour swatch (which is detected using the online inspection camera). The control colour swatch carries a much higher ink density than that of the micro imagery. Therefore, the ink composition has a large influence on the swatch brightness but the contrast of the magnified image will be influenced more strongly by other factors that have no impact on the brightness of the colour swatch. For example, the depth of the fine gravure cells in the surface of the gravure cylinder used to print the micro-imagery will have a far greater influence on the contrast in the magnified image. For example, a 10% change in cell depth has an insignificant change in the brightness of the colour swatch (simply because the colour swatch cells are much larger) whereas this change produces a significant variation in the contrast of the magnified image. Skilled workers will understand that a small change (+/- 10%) in the depth of the micro-imagery cells can dramatically increase or decrease the contrast in the magnified image.

[0017] In light of the above, an image acquisition system capable of 100% inspection, in line and at normal production speeds is not a realistic option in most commercial printing facilities. Flence, there is an ongoing imperative to develop other techniques for detecting micro-imagery defects and accurately measuring image contrast using in-line systems during the printing process.

[0018] Any reference herein to a patent document or other matter which is given as prior art is not to be taken as an admission that that document or matter was known or that the information it contains was part of the common general knowledge as at the priority date of any of the claims.

Summary of Invention

[0019] Accordingly, the invention provides a system for detecting defects in printed imagery used for generating optically variable images, the system comprising: a print feed path for media substrate to be printed with the imagery; a print station for printing the imagery and a test pattern on the media substrate, the test pattern having image elements spaced apart by gaps; and an image sensor for capturing images of the test pattern; such that, brightness levels acquired from the test pattern by the image sensor are used to derive a contrast measurement for comparison to a reference contrast level.

[0020] In another aspect, the invention provides a method of detecting defects in printed imagery used for generating optically variable images, the method comprising: feeding a media substrate along a print feed path; printing the imagery and a test pattern on the media substrate with a print station, the test pattern having image elements spaced apart by gaps; capturing images of the test pattern with an image sensor; and, recording brightness measurements of the test pattern for comparison to a reference brightness level to detect defects.

[0021] In yet another aspect, the invention provides a printing assembly configured for detecting defects in printed imagery used for generating optically variable images, the printing assembly comprising: a print station for printing the imagery and a test pattern on the media substrate, the test pattern having image elements spaced apart by gaps; and an image sensor for capturing images of the test pattern; such that, brightness levels are acquired from the test pattern to derive a contrast measure for comparison to reference image contrast levels.

[0022] The invention is based on acquiring brightness levels from a test pattern to derive a contrast measure that is indicative of the contrast of the imagery, instead of a direct contrast measurement of the imagery itself. This is well suited to measuring the contrast of a magnified image in a micro-lens based security feature and is a technique that can be applied as an inline test at production speeds. The test pattern may be printed inside or outside the‘product area’, that is, the printed micro-imagery or the printed article itself. The print station can print the test pattern as well as the micro- imagery and/or any other image element to be shown on the media substrate. The print station is any equipment used for printing the substrate bearing in mind the term ‘printing’ encompasses more than just the application of ink with a print tool (see the “Definitions” sub-section below). Typically, the image sensor will be part of a camera system arranged to acquire the brightness levels of the test pattern and process that image data to provide desired measure of the image contrast.

[0023] In one form, the printed imagery is micro imagery used in a security device with a lens array to generate optically variable magnifications; and the image elements of the test pattern each have an associated micro-lens such that at least some of the image elements are printed at different displacements relative to the respective micro lenses to generate respective magnified images with different levels of image brightness when viewed from a predetermined viewing angle, the different levels of image brightness defining a range for comparison to a predetermined acceptable brightness range to provide a measure indicative of contrast in the optically variable magnifications of the security device.

[0024] Preferably, the image elements and the gaps of the test pattern are a uniform size. Preferably, the image elements and the gaps of the test pattern are sized to substantially correspond to that of the printed imagery used to generate the optically variable image. Preferably, the associated micro-lenses for viewing the image elements of the test pattern and the micro-lenses within the micro-lens array for viewing the micro-imagery have the same lens geometry and orientation.

[0025] Preferably, the micro-lens array is a lenticular lens array.

[0026] Preferable, the micro-imagery and the image elements are spaced from the respective vertices of the associated lenticular lenses by a distance less than the focal length.

[0027] Preferably, the micro-imagery is an array of micro images that optically interacts with the lens array to create a contrast switch optical effect.

[0028] Preferably, the print station is a gravure cylinder and the media substrate is a continuous web extending from a feed roll to a collection roll along the print feed path. [0029] Preferably, the image elements are each a single line printed on a surface of the continuous web opposite the surface supporting the lenticular lens array.

[0030] Preferably, the micro-imagery and the image elements are printed in more than one colour.

[0031] Preferably, the single line has a width between 30% and 70% of, the width of each lenticular lens. More preferably, the single line width is approximately half the width of one of the lenses.

[0032] Preferably, the test pattern has a number of regions in which the single lines are at incrementally increasing offsets from the vertices of the associated lenticular lenses respectively.

[0033] Preferably, the test pattern and the associate lenticular lenses generate a moire magnification such that a maximum brightness level and a minimum brightness level are measured from the images captured by the camera for a comparison to the reference brightness levels indicative of the contrast.

Definitions

Security Document or Token

[0034] As used herein the term security document includes all types of documents and tokens of value and identification documents including, but not limited to the following: items of currency such as banknotes and coins, credit cards, cheques, passports, identity cards, securities and share certificates, driver's licenses, deeds of title, travel documents such as airline and train tickets, entrance cards and tickets, birth, death and marriage certificates, and academic transcripts.

[0035] The invention is particularly, but not exclusively, applicable to security documents such as banknotes or identification documents such as identity cards or passports formed from a substrate to which one or more layers of printing are applied. The diffraction gratings and optically variable devices described herein may also have application in other products, such as packaging. Substrate

[0036] As used herein, the term substrate refers to the base material from which the security document or token is formed. The base material may be paper or other fibrous material such as cellulose; a plastic or polymeric material including but not limited to polypropylene (PP), polyethylene (PE), polycarbonate (PC), polyvinyl chloride (PVC), polyethylene terephthalate (PET); or a composite material of two or more materials, such as a laminate of paper and at least one plastic material, or of two or more polymeric materials.

[0037] The use of plastic or polymeric materials in the manufacture of security documents pioneered in Australia has been very successful because polymeric banknotes are more durable than their paper counterparts and can also incorporate new security devices and features. One particularly successful security feature in polymeric banknotes produced for Australia and other countries has been a transparent area or“window”.

Transparent Windows and Half Windows

[0038] As used herein the term window refers to a transparent or translucent area in the security document compared to the substantially opaque region to which printing is applied. The window may be fully transparent so that it allows the transmission of light substantially unaffected, or it may be partly transparent or translucent partially allowing the transmission of light but without allowing objects to be seen clearly through the window area.

[0039] A window area may be formed in a polymeric security document which has at least one layer of transparent polymeric material and one or more opacifying layers applied to at least one side of a transparent polymeric substrate, by omitting least one opacifying layer in the region forming the window area. If opacifying layers are applied to both sides of a transparent substrate a fully transparent window may be formed by omitting the opacifying layers on both sides of the transparent substrate in the window area. [0040] A partly transparent or translucent area, hereinafter referred to as a“half window,” may be formed in a polymeric security document which has opacifying layers on both sides by omitting the opacifying layers on one side only of the security document in the window area so that the“half-window” is not fully transparent, but allows some light to pass through without allowing objects to be viewed clearly through the half-window.

[0041] Alternatively, it is possible for the substrates to be formed from a substantially opaque material, such as paper or fibrous material, with an insert of transparent plastics material inserted into a cut-out, or recess in the paper or fibrous substrate to form a transparent window or a translucent half-window area.

Printing

[0042] As used herein, the terms‘printed’ or‘printing’ refer to a surface treatment or technique to form images or indicia by changing the characteristics of reflected or transmitted incident radiation. The incident radiation need not be within the visible spectrum, and the wavelength of the transmitted or reflected radiation may be different to, or a subset of the incident radiation. The surface treatment or technique may involve the deposition of material such as ink, metal or polymer including radiation curable resin. Similarly, the surface treatment or technique may involve changes to a surface structure to alter the optical properties such as embossing, etching and laser ablation. The surface structure may be embossed or otherwise restructured to create structures, including diffractive or holographic structures, forming the images or indicia. Of course, the term printing also refers to more conventional and well known printing techniques such as intaglio, gravure, offset printing, inkjet printing, laser printing, screen printing, flexographic printing and so on.

Opacifying Lavers

[0043] One or more opacifying layers may be applied to a transparent substrate to increase the opacity of the security document. An opacifying layer is such that LT LO where Lo is the amount of light incident on the document, and LT is the amount of light transmitted through the document. An opacifying layer may comprise any one or more of a variety of opacifying coatings. For example, the opacifying coatings may comprise a pigment, such as titanium dioxide, dispersed within a binder or carrier of heat- activated cross-linkable polymeric material. Alternatively, a substrate of transparent plastic material could be sandwiched between opacifying layers of paper or other partially or substantially opaque material to which indicia may be subsequently printed or otherwise applied.

Security Device or Feature

[0044] As used herein the term security device or feature includes any one of a large number of security devices, elements or features intended to protect the security document or token from counterfeiting, copying, alteration or tampering. Security devices or features may be provided in or on the substrate of the security document or in or on one or more layers applied to the base substrate, and may take a wide variety of forms, such as security threads embedded in layers of the security document; security inks such as fluorescent, luminescent and phosphorescent inks, metallic inks, iridescent inks, photochromic, thermochromic, hydrochromic or piezochromic inks; printed and embossed features, including relief structures; interference layers; liquid crystal devices; lenses and lenticular structures; optically variable devices (OVDs) such as diffractive devices including diffraction gratings, holograms and diffractive optical elements (DOEs).

Embossable Radiation Curable Ink

[0045] The term embossable radiation curable ink used herein refers to any ink, lacquer or other coating which may be applied to the substrate in a printing process, and which can be embossed while soft to form a relief structure and cured by radiation to fix the embossed relief structure. The curing process does not take place before the radiation curable ink is embossed, but it is possible for the curing process to take place either after embossing or at substantially the same time as the embossing step. The radiation curable ink is preferably curable by ultraviolet (UV) radiation. Alternatively, the radiation curable ink maybe cured by other forms of radiation, such as electron beams or X-rays. [0046] The radiation curable ink is preferably a transparent or translucent ink formed from a clear resin material. Such a transparent or translucent ink is particularly suitable for printing light-transmissive security elements such as sub-wavelength gratings, transmissive diffractive gratings and lens structures.

[0047] In one particularly preferred embodiment, the transparent or translucent ink preferably comprises an acrylic based UV curable clear embossable lacquer or coating,

[0048] Such UV curable lacquers can be obtained from various manufacturers, including Kingfisher Ink Limited, product ultraviolet type UVF-203 or similar. Alternatively, the radiation curable embossable coatings maybe based on other compounds, e.g. nitro-cellulose.

[0049] The radiation curable inks and lacquers used herein have been found to be particularly suitable for embossing microstructures, including diffractive structures such as diffraction gratings and holograms, and microlenses and lens arrays. However, they may also be embossed with larger relief structures, such as non-diffractive optically variable devices.

[0050] The ink is preferably embossed and cured by ultraviolet (UV) radiation at substantially the same time. In a particularly preferred embodiment, the radiation curable ink is applied and embossed at substantially the same time in a Gravure printing process.

[0051] Preferably, in order to be suitable for Gravure printing, the radiation curable ink has a viscosity falling substantially in the range from about 20 to about 175 centipoise, and more preferably from about 30 to about 150 centipoise. The viscosity may be determined by measuring the time to drain the lacquer from a Zahn Cup #2. A sample which drains in 20 seconds has a viscosity of 30 centipoise, and a sample which drains in 63 seconds has a viscosity of 150 centipoise.

[0052] With some polymeric substrates, it may be necessary to apply an intermediate layer to the substrate before the radiation curable ink is applied to improve the adhesion of the embossed structure formed by the ink to the substrate. The intermediate layer preferably comprises a primer layer, and more preferably the primer layer includes a polyethylene imine. The primer layer may also include a cross-linker, for example a multi-functional isocyanate. Examples of other primers suitable for use in the invention include: hydroxyl terminated polymers; hydroxyl terminated polyester based co-polymers; cross-linked or uncross-linked hydroxylated acrylates; polyurethanes; and UV curing anionic or cationic acrylates. Examples of suitable cross linkers include: isocyanates; polyaziridines; zirconium complexes; aluminium acetyl acetone; melamines; and carbodi-imides.

Brief Description of the Drawings

[0053] Preferred embodiments of the present invention will now be described by way of example only with reference to the accompanying drawings in which:

Figure 1 schematically illustrates a prior art optically variable device for generating a contrast switch effect.

Figure 2 shows the positive image generated by the contrast switch effect device shown in Figure 1 .

Figure 3 shows six different images generated by the contrast switch effect device of Figure 1 transitioning from the positive image to the negative image through six different viewing angles.

Figure 4 is a schematic representation of a test pattern according to the present invention having four regions at different image element offsets overlayed by a micro lens array.

Figure 5 schematically illustrates a variation of the test pattern shown in Figure 4 in which the image element offsets are positioned such that no region generates maximum or minimum image brightness for the detection camera.

Figure 6 schematically illustrates a test pattern according to the present invention in which the image element period is 25% greater than the lenticular lens array period. Figure 7 shows a bottom row of images depicting the moire magnifications of a measurement test pattern at six different image element offset locations while the top row shows the corresponding moire magnification image generated by the product area of the device when viewed from the detection camera.

Figure 8 shows a contrast measurement system in which the test pattern is viewed by the detection camera in transmitted light.

Figure 9 shows a contrast measurement system in which the test pattern is viewed by the detection camera in reflected light against a strongly contrasting background surface.

Figure 10 shows a contrast measurement system in which the test pattern is indirectly viewed through a print substrate in transmitted light.

Figure 1 1 shows a contrast measurement system similar to that of Figure 9 in which a white test pattern is indirectly viewed by the detection camera through the print substrate 12 in reflected light against a strongly contrasting dark background.

Figure 12 shows a contrast measurement system in which the test pattern is viewed by the detection camera in reflected light with the opposing lens array against a strongly contrasting background surface.

Figure 13 shows four bank notes with an area of microimagery for generating an optically variable effect and a test pattern printed outside the areas of microimagery but within periphery of the banknotes;

Figure 14 shows an alternative to Figure 13, in which the test pattern is printed within the areas of microimagery in each of the bank notes;

Figure 15 shows a further option in which the test pattern for each of the bank notes is not only printed outside the areas of microimagery but also beyond the periphery of the associated bank note; and Figure 16 shows a sheet of bank notes with the test pattern printed in the peripheral margin of the sheet surrounding the array of notes, such that each column of the notes is associated with a single test pattern respectively.

Detailed Description

[0054] Referring to Figure 4, an image contrast measurement system is schematically illustrated as a test pattern 72 having four regions 36, 38, 40 and 42. All four regions of the test pattern 72 have image elements 46, 48, 50 and 52 respectively printed on one side of the print substrate 12, optionally at a position spaced from the product area. Flowever, the lenses 14 have the same geometry and pitch as the lenses in the product area, and the test pattern image elements are printed with the same print station used to apply the micro-images in the product area.

[0055] The image elements in each region have different offsets relative to the vertices 62 of their associated lenses 14. The offset amount increases at fixed increments for each successive region with each increment being the lens pitch divided by the number of regions. In Region I, the image elements 46 are offset by an amount 64 from the vertex 62 of the associated lens 14. The geometry of the lenses 14 is such that the focal length is longer than the distance from the top of the lens to the image elements. From the camera’s viewing angle 54, each lens 14 focuses on respective strips 44, 56, 58, 60, etc. extending parallel to the lenses along the opposing side of the print substrate 12. Conveniently, the lens geometry and/or the substrate 12 thickness is selected so that the width of these focal strips 44, 56, 58 and 60 roughly corresponds to the width of the image elements 46, 48, 50 and 52 shown in the test pattern. This has the potential to provide the greatest difference in detected colour brightness across all regions.

[0056] In the example shown in Figure 4, the image elements are approximately half the width of the lenses 14. In Region I, the test pattern 36 has image elements 46 at an offset 64 from the lens vertices 62. At this offset, the focal strip 44 of the lenses 14 sees half of a print element 46. In Region II, the test pattern 38 has image elements 48 at a smaller offset 66 from the vertices 62 of the lenses 14. In this region, the focal strips 56 see all the image elements 48. Therefore, the colour brightness seen in Region II should be more than Region I and indeed the minimum brightness across all the regions (assuming that image elements are darker than the gaps and background).

[0057] In Region III, the focal strips 58 see half the image elements 50 in the test pattern 40 as the offset 68 is set to zero. The brightness detected in Regions I and III should be equal as the lenses 14 in each region are seeing the same proportion of the image elements in the test patterns.

[0058] In Region IV, the image elements 52 are at an offset 70 from the vertices 62 so that the focal strips 60 see none of the image elements 52 in the test pattern 42. Hence the colour brightness seen by the camera in Region IV is the highest of all four regions, and theoretically the highest possible brightness level. Conversely, Region II should have the lowest possible brightness level and therefore the difference between these brightness levels provides an accurate measurement of the colour contrast in the magnified image seen in the product area. If the colour contrast detected across the four regions of the complete test pattern 72, meets or exceeds a predetermined threshold, the optically variable device is suitable.

[0059] While the test pattern 72 shown in Figure 4 has four regions with image elements at fixed offsets of 0%, 25%, 50% and 75% of the lens width, it can be seen that the test pattern need only have a minimum of two regions in which the image elements are at two offsets, preferably equal to half the width of the lenses. However, the advantages of having additional regions and a larger range of offsets is discussed in more detail below.

[0060] Referring again to Figure 4, each region effectively implements a contrast switching effect in which the stages of contrast switch are seen from a single viewing angle 54 (i.e. that of the camera). The minimum and maximum brightness regions (Regions II and IV respectively) represent the contrast switching effect generated by the optically variable device in the product area when viewed from two complementary viewing angles. The detection camera views the test pattern 72 along the perpendicular 54 to the lenses 14 to simultaneously acquire the projected image. The brightness level in each region is determined and the highest and lowest brightness levels are used to calculate the contrast of the image. [0061] One common measure of contrast is Michelson Contrast, which is calculated as:

Michelson Contrast

Imax + Imin where:

Imax is the maximum brightness level detected; and

Imin is the minimum brightness level detected.

[0062] The skilled worker in this field will understand that other definitions of contrast may be used, such as:

Luminance contrast;

Weber contrast;

RMS (root mean squared) contrast;

Any defined function of the brightness levels, colour strengths or grey levels detected in the magnified image of the test pattern 72 acquired by the camera

[0063] As discussed above, a contrast measurement theoretically requires only two regions of different brightness levels, preferably at maximum brightness and minimum brightness. However, in a real world production environment, the offset of the print element lines in the test pattern, relative to their associated micro lenses, will vary from product to product due to the printing press lens-to-print registration tolerance, substrate skew, substrate deformation, and so on. Hence in some products (i.e. optically variable devices), and therefore their corresponding test patterns, no regions will have image elements at offsets that project images at the maximum and minimum brightness levels to the fixed viewing angle of the detection camera. In other words, none of the regions will have the precise offsets depicted in Regions II and IV shown in Figure 4. Instead, the maximum and minimum brightness levels projected will be less than, and more than, the true maximum and minimum values respectively. [0064] Figure 5 schematically depicts this scenario. Here the image elements 46 to 52 in Regions I to IV respectively are at actual offsets of -12.5%, 12.5%, 37.5% and 62.5%. In other words, they differ from theoretical or optimal offset by 12.5%. In this case, Regions I and IV of test pattern 74 have the highest brightness levels as their focal strips 44 and 60 see the least of the image elements 46 and 52 respectively. Likewise, Regions II and III of the test pattern 74 will project the lowest brightness levels as their focal strips 56 and 58, see more of the image elements 48 and 50 respectively. However, the brightness of Regions I and IV is less than the maximum brightness level (which would be seen at a different viewing angle) and conversely the brightness level in Regions II and III is greater than the minimum brightness levels seen at the complementary viewing angle.

[0065] On the premise that brightness is proportional to the unprinted area of the focal strips 44 to 60, then the highest and lowest brightness levels detected by the camera in test pattern 74 will differ from the true maximum and minimum brightness levels by 12.5%. More generally, the maximum difference in percent between the measured brightness and the maximum or minimum brightness level =

1

- ÷ 2 x 100

(number of fixed offsets)

[0066] This illustrates the problem of measuring the true maximum and minimum brightness levels with a camera at a fixed viewing angle, while accounting for production tolerances causing variation in the micro-lens to print element offset, can be mitigated by using a larger number of test pattern regions, each with a different print element offset. Test pattern 72 and 74 shown in Figures 4 and 5, each use four different offsets. However, it will be appreciated that more accurate contrast measurements can be obtained if eight evenly distributed offsets are used (i.e. 0%, 12.5%, 25%, 37.5%, 50%, 62.5%, 75% and 87.5%, expressed as percentages of lens width). With eight different fixed offsets in eight different regions, the maximum possible discrepancy from true maximum and minimum brightness intensity levels is equal to 1 ÷ 8 ÷ 2 x 100 = 6.25%. In other words, the maximum discrepancy is halved compared to that depicted in Figure 5, in which four offsets in four regions are used. [0067] Referring to Figure 6, another embodiment is illustrated in which the number of fixed offsets (and therefore the number of corresponding regions) is greatly increased.

[0068] The technique used in Figures 4 and 5 has test patterns 72 and 74, in which four regions have four different offsets evenly distributed across one lens width. In the test pattern 76 shown in Figure 6, each region consists of a single micro-lens and the regions are contiguously adjacent one another. The image elements 46, 48, 50 and 52 are evenly spaced at a pitch given by the equation: print element pitch = lens pitch + (lens pitch ÷ number of evenly distributed offsets)

[0069] In the schematic illustration of Figure 6, the print element pitch is: lens width + (lens width ÷ 4) = 1.25 x 1 lens width

[0070] This formula applies if the image elements are being offset to the right with each subsequent lens 14. If the image elements are offset to the left, each subsequent lens moving from left to right, then the formula for the image element pitch is given by the equation: print element pitch = lens pitch— (lens pitch ÷ the number of evenly distributed offsets)

[0071] The test pattern shown in Figure 6 provides image elements 46, 48, 50 and 52 in the form of parallel lines with a pitch different to that of the lens pitch. This mismatch between the lens pitch and image element pitch generates a moire magnification seen by the detection camera (not shown). From the captured image of the moire magnification, the highest and lowest brightness intensity levels are measured by detecting the bright and dark bands (see Figure 7). Using the highest and lowest brightness levels, an accurate measure of image contrast may be determined.

[0072] The period of the magnified moire bands can be calculated using the known moire magnification formula, that is:

Moire period =(lens pitch x image element pitch) ÷ Abs(image element pitch

— lens pitch). [0073] In the examples shown in Figure 6, the moire period equals five times one lens pitch. As the micro lenses 14 are very small, this example is a very small moire period and would be difficult to detect the highest and lowest brightness levels. Similarly, the problems outlined above regarding the insufficient fixed offsets used, the detected maximum and minimum brightness levels may differ significantly from the true maximum and minimum values. In light of this, the test pattern 76 used in an actual production environment generates a much larger moire period to enable the camera to detect the true maximum and minimum brightness levels. For example, a test pattern 76 using image elements 46 to 52 with a pitch that is 1 % greater than that of the lens pitch (which is equivalent to using 100 different fixed offsets) will produce a moire period equal to 101 times the lens width. If, for example, the lens width is 64 microns and the print element pitch is 64.64 microns, the dark and light bands generated in the moire magnification have a period of 6.464 mm. The detection of maximum and minimum brightness levels within a 6.464 mm wide portion of the moire magnification of the test pattern 76 is straight forward for the image processor receiving captured images from the detection camera. The maximum error or discrepancy between the highest and lowest brightness levels detected and the true maximum and minimum brightness levels is 0.5%:

1

0.5 = ( ÷ 2 X 100

100

[0074] The moire magnification of the test pattern 76 will consist of a repeating pattern of light and dark bands with only one dark band and one light band within each moire period. From this image brightness data, any of the different contrast measures discussed above (Michelson contrast, RMS contrast, Weber contrast, etc.) can be determined.

[0075] The moire bands need not be created by a mismatch between the pitch of the image elements and that of the lenses. Optionally, the moire magnification of test pattern 76 may be generated using relative skew between the micro-lenses 14 and the image elements 46 to 52. Skilled workers will also understand that a moire magnification may also be generated by varying both the image element pitch (relative to the lens pitch) as well as introducing a relative skew between the lens array 14 and the image elements 46 to 52. Using the known moire band configurations created in moire magnifications generated by predetermined pitch mismatches and/or skews, a defined moire band period and orientation relative to the detection camera can be controlled and set to a predetermined design. However, the contrast measurement test pattern and associated micro-lens array must be large enough to cover an area of at least one magnified moire band period. This ensures the detection camera can properly detect the highest and lowest brightness levels within the moire magnification.

[0076] Figure 7 shows a moire magnification of a test pattern and the magnified image generated in the product area of the associated optically variable device. The top row shows the magnified images 78A to 78F generated by the optically variable device in the product area, while the associated moire magnifications 80A to 80F are shown directly below in the row beneath. Each of the moire magnifications 80A to 80F show one magnified moire period. That is the test pattern size is equal to one moire period and this ensures the detection camera captures the highest 82 and lowest 84 brightness intensity levels for the moire magnified image.

[0077] Moire magnifications 80A to 80F show the particular moire magnifications seen by the detection camera for six different image offset distances between the lens vertices and the print images (as discussed above in relation to Figures 4 and 5). These different offset distances may be readily expected to occur in a production environment due to the various tolerances involved as discussed above. However, regardless of the different offset positions, the same highest 82 and lowest 84 brightness levels are captured. Only the position of the bright bands 82 and the dark bands 84 changes with the changes in offset position between the lenses and the image elements of the test pattern. Accordingly, the contrast measured will be consistent and effectively independent of the lens to print offset variation that will be inherent to some degree in the production environment. The contrast measured will differ from the true contrast by a negligible percentage that may be calculated using the following equations:

[0078]

1

÷ 2 X 100.

(number of fixed offsets) [0079] Wherein: number of fixed offsets = lens pitch ÷ Abs(image element pitch - lens pitch).

[0080] The techniques described above in relation to Figures 4 to 7, provide a contrast measurement test pattern that is printed using the same print station (e.g. gravure cylinder) that prints the micro imagery of the optically variable device product. Due to this, the contrast in the magnified image of the optically variable device product will be accurately represented in the contrast of the magnified test pattern image. These techniques described above have overcome the issue of varying lens to print element phase shift which complicates the direct measurement of the contrast in the magnified image generated by the optically variable device product. Providing a test pattern that projects consistent and accurately representative contrast enables a single detection camera viewing from a single angle to provide an accurate contrast measurement while operating inline and at normal production speeds. As discussed above, the contrast test pattern may also be placed either inside or outside the product area.

[0081] It should be noted that it is not essential that the test pattern image elements are the same size as the micro-imagery. However, similarly sized test pattern imagery and micro-imagery (and the intervening gaps), will yield a contrast measure that is more representative of the contrast in the magnified image. So while it is preferable the test pattern elements and micro-imagery are roughly equivalent, it is not essential. If the test pattern element sizes and gaps are smaller (for example 50% smaller) or larger (for example 100% larger) the contrast techniques still work. Likewise, the test pattern may consist of multiple image element sizes and multiple gap sizes, for example sizes that fall within a range or a distribution. Indeed, a variety of test pattern geometries could be used within a single test pattern (such as geometries with randomised element sizes and gap sizes) but test pattern image element sizes and gaps corresponding to that in the product imagery is preferable in most situations.

[0082] Figures 8 to 12 show further techniques for measuring contrast in accordance with the present invention. However, these techniques differ from the previously described methods in that: the detection camera views the image elements of the test pattern directly; or

the detection camera views the test pattern image elements through the substrate but not through the lens array; or

the detection camera views the test pattern image elements through one or more layers of ink applied to the test pattern.

[0083] In the measurement technique depicted in Figures 8 to 12 the image elements of the test pattern and the gaps between the image elements are preferably constant and of a similar size to the image elements and intervening gaps in the micro image array of the product area (although as discussed above, this is not essential). However, in these techniques the micro lens to print element offset is not a relevant parameter. Once again, the test pattern may be printed outside the product area encompassing the optically variable device or it may be incorporated within the product area and form part of the micro imagery used by the optically variable device.

Camera Views Test Pattern Image Elements Directly

[0084] As shown in Figure 8, if there are no micro lenses opposite the test pattern, it will be preferable to view the test pattern 36 in transmitted white light from a white light source 88 positioned opposite the field of view 92 of the detection camera 90. The camera directly views the image elements 18 and the intervening gaps 16, which are dimensioned to roughly correspond with the sizes of the micro imagery elements and intervening gaps within the product area to ensure a meaningful representation of the contrast is measured.

[0085] Figure 9 shows micro lenses 14 opposite the test pattern 36. In this situation, transmission lighting is not appropriate because it produces variable magnified images to the camera 90. Accordingly, the test pattern 36 is viewed in reflection from a white light source 88 on the camera side of the substrate 12. For more accurate contrast measurement, the lens side of the substrate 12 is held against a high contrast background surface 96, such as a white surface, which may be provided by one of the rollers in the printer press. The colour of the background surface 96 is a strongly contrasting colour to that of the test pattern image elements 18. In reflected light with the lenses 14 in direct contact with the contrasting background surface 96, magnified images are not generated or viewed by the camera 90. Similarly, the direct contact between the lenses 14 and the contrasting support surface 96 prevents magnified optical images of the image elements 18 being formed by light reflected from the internal lens to air interface focussed onto the image elements 18.

[0086] It should be noted that the use of reflection lighting as shown in Figure 9 will also allow the camera 90 to acquire a suitable image of the test pattern 36, if the lenses 14 are not present.

[0087] Camera Views Test Pattern Image Elements Through the Substrate But Not Through Lens Array

[0088] Referring to Figure 10, the camera 90 views the test pattern 36 through the substrate 12 (formed with a suitably translucent or transparent material) but not through a lens array formed on the opposing side of the substrate. In the lighting arrangement 98 shown in Figure 10, the white light source 88 is positioned on the opposing side of the substrate 12, such that the camera views the image elements 18 of the test pattern in transmitted light, however placing the light source 88 on the same side of the substrate 12 as the camera 90, such that the test pattern 36 is viewed in reflected light, would also be suitable, if the test pattern is placed in contact with a surface having a contrasting colour. An example of the latter configuration is illustrated in Figure 1 1 .

[0089] In Figure 1 1 , the test pattern 102 is a series of white image elements 18. Accordingly, the test pattern 102 is placed against a contrasting dark surface 104 (again possibly the surface of a roller) such that the camera 90 views the test pattern 102 through the substrate 12 in reflected light provided by the white light source 88.

[0090] The decision to use transmission or reflection lighting may depend on whether the test pattern 36 is overprinted with one or more contrasting colour layers. If the image elements 18 are overprinted with a contrasting ink layer viewing the test pattern 36 through the substrate 12 in reflected light is preferred, with the overprinted layer preferably placed in contact with a similarly coloured surface, in order to further enhance the contrast in the acquired image of the test pattern. [0091]

Camera Views Test Pattern Image Elements Through Overprinted Ink

Laver(s)

[0092] The arrangement 106 shown in Figure 12 shows the image elements 18 of the test pattern 36 being viewed by the camera 92 through a layer of contrasting ink 108 applied over the test pattern 36. As discussed above, the lenses 14 present on the opposing side of the substrate 12 means that transmission lighting should not be used in order to avoid generating magnified images within the camera field of view 92. Instead, the image elements 18 are viewed from the so called‘lens reverse side’ in reflection lighting provided by the white light source 88. Similar to arrangements 94 and 100 shown in Figures 9 and 1 1 respectively, the lenses 14 are in direct contact with a support surface 96 of strongly contrasting colour to that of the image elements 18. This configuration of reflection lighting and a contrasting colour support surface 96 placed against the lens side effectively eliminates any magnified optical effect images that would otherwise manifest in other lighting arrangements. As discussed above, the magnified images of the image elements 18, generated on the lens reverse side in transmission lighting, are undesirable optical effects that cause the detection camera to identify false rejects.

[0093] If lenses are not present on the arrangement 106 shown in Figure 12, the test pattern 36 can be viewed in transmission or reflection lighting. If reflection lighting is used, the substrate 12 is placed against a contrasting support surface 96. This has the effect of increasing the contrast between the test pattern 36 and its background.

[0094] To measure the contrast using the techniques shown in Figures 8 to 12, a brightness measure is determined from the image of the test pattern 36 acquired by the camera 92. For example, one possible brightness measure is the average grey-level of an acquired greyscale image of the test pattern. In other words, each pixel of the image captured by the camera has a grey level that can be used as a brightness measure for that pixel. The average grey-level is simply the sum of each pixel grey-level divided by the number of pixels. This brightness measure is compared to a predetermined threshold brightness value to ascertain whether the contrast of the optically variable device in the product area is acceptable.

[0095] The brightness measure computed from the acquired test pattern image is representative of the image contrast in the optically variable device because the size of the image elements 18 in the test pattern 36 roughly correspond to the size of the image elements in the product area. As the size of the image elements in the product area varies, so too does the contrast in the magnified images generated by the optically variable device. As the test pattern is printed with the same print station (such as a gravure roller) that prints the micro imagery in the product area, the variations in the micro imagery correspond to variations in the image elements 18 in the test pattern. As the size of the image elements 18 and the test pattern 36 increase or decrease, the size of the unprinted areas of the test pattern 36 correspondingly increase or decrease. As a result, the brightness measure of the test pattern 36 derived from the image data acquired by the camera 90 will be representative of the contrast in the magnified image of the optically variable device in the product area. Using an averaging technique or similar image data manipulation to derive a single brightness measure is necessary because the actual dimensions of the image elements 18 and the intervening gaps 16 of the test pattern 36 are below the spatial resolution of the camera 92, and a small dimensional change (that adversely impacts on product contrast) cannot be measured from the acquired camera image.

Inline Defect Detection Method

[0096] The lighting configurations and arrangements shown in Figures 8 to 12 may also be applied for the purpose of detecting defects in the micro imagery printed within the product area. In this case, missing portions of the micro imagery used by the optically variable device are directly detected using the image acquired by the camera instead of computing a measure of brightness generated by the image elements 18 of the test pattern 36. The missing portions of the micro imagery are detected in the image acquired by the camera through a process of comparison to a golden template. The missing portions detected have a minimum size corresponding to the image acquisition resolution of the camera system. [0097] The production of optically variable devices typically involves supplying the lenses to the substrate before applying the micro imagery to the reverse side, hence the inspection of the product micro imagery missing portions occurs with lenses 14 present on the substrate 12. As discussed above, the camera should acquire the images of the micro imagery in reflection lighting with the lenses in direct contact with a support surface of strongly contrasting colour to that of the micro imagery. Furthermore, the camera should acquire these images prior to any further layers being applied over the micro imagery. This ensures images are acquired with maximum contrast and therefore maximum defect detection sensitivity.

Test Pattern Location

[0098] Figures 13 to 16 show various options for positioning the test pattern relative to the printed article and/or the product area that generates the optically variable effect. For the purposes of illustration, the printed articles shown in Figures 13 to 16 are banknotes 1 10, each of which has a product area 1 12 containing microimagery used to generate an optically variable effect. The product area 1 12 is typically covered by an array of microlenses to create an optically variable synthetic magnification of the underlying microimagery. These lens-based optically variable devices provide an effective anti-counterfeiting measure that is extremely difficult to replicate. Individual banknotes may be printed on a continuous web extending between a feed roller and a take up roller, or they may be printed on individual sheets which are later guillotined into individual banknotes 1 10. The sheet of media substrate may have the banknotes arranged in a rectangular array. A gravure printing process may be used with the gravure print cylinder sized to print a single sheet of banknotes (arranged in a rectangular array) per cylinder rotation. As discussed above, the same gravure cylinder prints both the micro imagery 1 12 and the test pattern imagery 1 14.

[0099] As shown in Figure 13, the test pattern 1 14 may occupy an area within the outer periphery of each of the banknotes 1 10, but external to the micro imagery of the product area 1 12.

[0100] In Figure 14, the test pattern 1 14 is printed inside product area 1 12 of each of the banknotes 1 10. This may require careful design to incorporate or merge the test pattern imagery 1 14 and the other micro imagery of the product area 1 12, to generate the desired optically variable effect. This arrangement has the advantage that a second and separate array of microlenses is not required for the test pattern (as may be the case in Figure 13).

[0101] In Figure 15, the test patterns 1 14 for each of the banknotes 1 10 are printed external to the banknotes. In this way, the test patterns 1 14 do not form part of the final printed articles 1 10. Instead the test patterns 1 14 are used to assess the image contrast within the product area 1 12 and then discarded when the sheet or web is guillotined into individual banknotes 1 10. This option does require an additional (or extended) micro lens array for each of the test patterns 1 14 and a larger proportion of the media substrate is‘wasted’ to accommodate the test patterns 1 14.

[0102] A more efficient use of the media substrate 1 16 is shown in Figure 16. The sheet 1 16 of media substrate has peripheral areas 1 18, 120 and 122 surrounding the rectangular array of bank notes 1 1 0. The test patterns 1 14 may be positioned anywhere within these peripheral margins, however for convenience the test patterns can be located in the peripheral margin area 122 directly aligned with each column of the product areas 1 12. In this way the microlens array can be applied continuously along the entire sheet 1 16. Furthermore, Figure 16 illustrates that it is not necessary to provide a test pattern 1 14 for each and every one of the product areas 1 12.

[0103] The present invention has been described here by way of example only. Skilled workers in this field will readily recognise many variations and modifications which do not depart from the spirit and scope of the broad inventive concept.