Sensors and methods for examining valuable documents

By detecting optical or magnetic fiber signals of valuable documents using sensors and determining local feature values ​​using image detection and analysis devices, the problem of identifying counterfeit valuable documents has been solved, enabling efficient identification and sorting of counterfeit documents.

CN116686019BActive Publication Date: 2026-11-17GIESECKE & DEVRIENT CURRENCY TECHNOLOGY GMBH
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202180080323.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2020-12-01
Filing Date
2021-12-01
Publication Date
2026-11-17
Estimated Expiration
2041-12-01

AI Technical Summary

Technical Problem

Existing technologies struggle to effectively identify forgeries in valuable documents, especially constructive forgeries, and known methods cannot fully identify these forgeries.

Method used

A sensor is used to detect optical or magnetic fiber signals in valuable documents. An image detection and analysis device determines the local feature values ​​of the fibers to identify suspected forgeries. The sensor includes an image detection device and an analysis device; the analysis device locates the fibers and classifies valuable documents based on their local feature values.

Benefits of technology

It improves the detection rate of forged valuable documents, especially structural forgeries, and can automatically identify and sort out valuable documents suspected of being forged.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116686019B_ABST
    Figure CN116686019B_ABST
Patent Text Reader

Abstract

This invention relates to a sensor for inspecting valuable documents, the valuable documents having multiple fibers distributed on the document and possessing characteristic optical or magnetic signals. The sensor includes an image detection device for detecting the optical or magnetic signals of the valuable document in a point-resolved manner, the image detection device being configured to detect an image of the valuable document. The sensor also includes an analysis device configured to, for analyzing the image of the valuable document, locate the fibers in the image of the valuable document, determine at least one local fiber feature value for one or more different points in the image of the valuable document, and classify the valuable document as potentially counterfeit based on these local fiber feature values. This allows for the automatic sorting of potentially counterfeit valuable documents in the automated processing of valuable documents.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] This invention relates to sensors and methods for inspecting valuable documents.

[0002] Here, a valuable document is understood to be, for example, a sheet of paper representing monetary value or rights and therefore should not be arbitrarily manufactured by an unauthorized person. Thus, a valuable document possesses anti-counterfeiting features that are not easily forged, the presence of which is a sign of authenticity. Important examples of such valuable documents are chip cards, coupons, vouchers, checks, and especially banknotes. Valuable documents can be distinguished by their type; for example, in the case of banknotes, they can be distinguished by their face value or monetary value, or in the case of checks, by the type of check format provided by the check issuer.

[0003] To provide security against counterfeiting, it is known to equip valuable documents with magnetic fibers that provide characteristic magnetic signals or with fibers that provide characteristic optical signals, such as characteristic reflection or luminescence. Valuable documents are typically equipped with luminescent mixed-color fibers (or fibers with mixed colors), which are distributed in a specific manner or randomly on the surface of the document. These mixed-color fibers are usually added to the paper substrate of the valuable document during the manufacturing process. It is known to optically excite valuable documents equipped with mixed-color fibers and examine the luminescence of the mixed-color fibers. To distinguish genuine valuable documents from counterfeits, it is common practice to examine whether the corresponding valuable document has luminescent mixed-color fibers.

[0004] Forgeries of valuable documents have repeatedly surfaced; these forgeries are referred to below as constitutive forgeries. These structural forgeries can be produced by attaching separate portions of a valuable document (usually by pasting) to a substrate segment, such as a suitably shaped piece of paper or film, to create a structure that approximates the size of the genuine document. For example, a valuable document can be cleverly divided into two parts, which are then each spliced ​​together with a suitable substrate segment. This produces two forgeries of the valuable document that, based on their forgery design, may be difficult to distinguish from the genuine document.

[0005] Various methods are known for identifying structural counterfeits, such as those targeting adhesive strips or printed image misalignments. However, these methods can only identify some structural counterfeits, while others remain undetected.

[0006] Therefore, the technical problem to be solved by the present invention is to provide a sensor and an improved method for checking whether valuable documents are suspected of being forged, especially structural forgery.

[0007] The technical problem described herein is solved by the features of the independent claims. Advantageous designs and extensions of the invention are given in the dependent claims.

[0008] This invention relates to the inspection of valuable documents having multiple fibers distributed on the document and possessing characteristic optical or magnetic signals. These fibers may be distributed (non-uniformly or uniformly) on one or more sections of the valuable document or throughout the entire document. For example, these fibers may be distributed in or on the surface of a substrate of the valuable document. In particular, the optical and / or magnetic properties or signals of these fibers differ from those of the substrate of the valuable document. Valuable documents inspected according to the method of the invention include, for example, banknotes, checks, admission tickets, coupons, etc. To inspect these valuable documents, especially to identify counterfeits, particularly structural counterfeits, the respective valuable documents to be inspected are placed in the detection area of ​​a sensor.

[0009] The sensor includes an image detection device (or image acquisition device) configured to detect an image of a valuable document, preferably the entire valuable document. Specifically, the image detection device is configured to detect optical or magnetic signals of the valuable document in a resolvable point (or position) manner. The image of the valuable document shows a two-dimensional distribution of the optical or magnetic signals of the valuable document on the document, preferably the entire valuable document, and includes or displays characteristic (or representational) optical or magnetic signals of fibers. The sensor also includes a control device configured to manipulate the image detection device such that it detects the optical or magnetic signals of the valuable document in a resolvable point manner.

[0010] Furthermore, the sensor includes an analysis device configured to analyze valuable document images. This analysis device is configured to locate fibers contained within a given valuable document image and determine at least one local fiber feature value for one or more different points within the valuable document image. This local fiber feature value is applicable to the corresponding point and, if necessary, to the environment of the corresponding point within the valuable document image. The analysis device is also configured to classify valuable documents as potentially counterfeit based on the at least one local fiber feature value, particularly based on the local fiber feature values ​​determined for different points within the valuable document image. In the case of valuable documents being processed automatically by a valuable document processing device, valuable documents suspected of being counterfeit can be automatically sorted out by the valuable document processing device.

[0011] The analysis device is configured to receive valuable document images detected by an image detection device, which are transmitted to the analysis device (e.g., from a processor or the image detection device). The analysis device may be located within or outside the sensor housing. Alternatively, the analysis device may be a central analysis unit of a valuable document processing device that (if necessary, jointly) analyzes measurement signals from multiple sensors.

[0012] The corresponding local fiber feature values ​​(or individual local fiber feature values) are uniquely applicable to a specific point in a valuable document image, or uniquely applicable to that point and its (spatial) environment within the valuable document image. The environment of the corresponding point can include that point. The environments of different points are not identical, but can overlap. Generally, the local fiber feature values ​​of different points in a valuable document are different from each other. A single local fiber feature value can be determined for each point and its environment, or different local fiber feature values ​​can be determined for each point and its environment, and these local fiber feature values ​​can be used to classify the valuable document.

[0013] Determine the corresponding point for at least one local fiber feature value and, if necessary, the environment of the corresponding point, such as a sub-image or raster point corresponding to a valuable document image, where the raster point has an environment region belonging to it (as described below).

[0014] For example, local fiber characteristic values ​​characterize the local density of fibers in the environment at the corresponding location and / or characterize the local distribution of fibers in the environment at the corresponding location and / or characterize the location of one or more fibers in the environment at the corresponding location and / or characterize the local properties of fibers in the environment at the corresponding location. This property may particularly relate to the appearance of the fibers at the corresponding location and in the environment at the corresponding location (e.g., fiber shape, size, color, or brightness), or, in the case of magnetic fibers, the level of the magnetic signal of the fibers at the corresponding location and in the environment at the corresponding location. For example, representative values, such as average values, of the local properties of the fibers at the corresponding location and, if necessary, in the environment at the corresponding location can be used as local fiber characteristic values. Alternatively, the distance of the corresponding location relative to the nearest fiber can be used as a local fiber characteristic value, which characterizes the local location of one or more fibers in the environment at the corresponding location. The dependence of the corresponding location on an area of ​​influence defined around the corresponding fiber (defined below) can also be used as a local fiber characteristic value.

[0015] In a preferred embodiment, a measure of the local density of fibers is used as a local fiber feature value. For this purpose, the number of fibers or fiber density (the number of fibers divided by the area of ​​the sub-image / environment region) can be used as a local fiber feature value in the corresponding sub-image or environment region. Alternatively, a (e.g., statistical) distance measure representing the distance between fibers closest to the corresponding point or in the corresponding sub-image / environment region, such as the average distance between these fibers in the sub-image / environment region and their nearest neighbors (or, for each fiber, the nearest neighbor), can be used as a local fiber feature value.

[0016] In other preferred embodiments, a measure of the local properties of the fiber at a point and, if necessary, in the environment of that point is used as a local fiber characteristic value, which in particular relates to the shape (e.g., aspect ratio, orientation, curvature) or size (e.g., length, width, area) or color or brightness of the fiber at that point and, if necessary, in the environment of that point.

[0017] Fibers are individual, elongated objects, for example, having a length ranging from several millimeters, and whose optical and / or magnetic properties differ from those of the substrate of the valuable document. In this application, the term "fiber" does not refer to the paper fibers of the paper substrate of the valuable document. Fibers are manufactured separately from the substrate material of the valuable document and are introduced into or applied to the substrate material after its manufacture. Fibers are, for example, made of paper, or of a material different from the substrate material of the valuable document, such as plastic or metal, and the substrate material of the valuable document is made of paper. These fibers are, for example,

[0018] - Light-emitting fibers distributed in or on a non-luminescent valuable document substrate, or

[0019] - Magnetic fibers distributed in or on a non-magnetic valuable document substrate, or

[0020] - Black, gray, or monochrome or multicolored fibers distributed in or on a light-colored or white substrate of valuable documents.

[0021] In some types of valuable documents, the fibers are distributed substantially uniformly across the document. However, statistical variations during document manufacturing—such as when applying the fibers to the document or introducing them into the material used to create the substrate (e.g., paper or plastic)—can lead to deviations from an absolutely uniform distribution. In other types of valuable documents, the document has one or more specific fiber-containing regions, the positions of which are fixedly predetermined.

[0022] The analysis apparatus is specifically configured to perform image analysis on images of valuable documents, in which fibers that may be contained within the images are located (based on the signal distribution of optical or magnetic signals). To identify fibers, adjacent image points in the image of the valuable document whose signal strength exceeds a threshold can be searched, and it can be checked whether the objects generated by these adjacent image points have the expected shape, particularly thickness and length, and, if necessary, shape of the fibers. The located fibers may have the same or different properties, such as the same or different shapes or colors.

[0023] The valuable document can be statically introduced into the sensor's detection area for inspection. In this case, a two-dimensional image sensor is preferred, which detects optical or magnetic signals emitted from the valuable document. However, it is preferable to transport the valuable document alongside the sensor along the transport direction so that different sections of the valuable document can be scanned sequentially by the sensor. In this case, an optical or magnetic image sensor designed as a sensor row (or linear array sensor) is preferred.

[0024] In the first variant, the sensor is a magnetic sensor, and the image detection device is configured to detect the magnetic signal of the valuable document, particularly the fiber, in a resolvable point manner. In this case, the image of the valuable document is a magnetic signal image of the valuable document. The fiber is a magnetic fiber whose magnetic properties differ from those of the substrate of the valuable document. In this case, the characteristic magnetic signal of the fiber is a magnetic signal that is significantly different from the typically low or negligible magnetic signal of the low-magnetic or non-magnetic substrate of the valuable document. The image detection device is a magnetic signal image detection device configured to detect the magnetic signal of the fiber in a resolvable point manner. This image detection device particularly has multiple magnetic detector elements and may be, for example, a high-resolution magnetic detector row (for inspecting valuable documents transported past the sensor) or magnetic detector elements arranged in a two-dimensional manner.

[0025] In the second variant, the sensor is an optical sensor, and the image detection device is configured to detect the optical signal of the valuable document, particularly the fiber, in a resolvable point manner. The fiber possesses optical properties different from those of the substrate of the valuable document. In this case, the image detection device is an optical image detection device (e.g., an optical image sensor) configured to detect the optical signal of the fiber, particularly its emission or reflection, in a resolvable point manner. This image detection device has multiple photosensitive elements and is, for example, a CCD or CMOS-based image sensor or a two-dimensional image sensor. The image of the valuable document can be a reflective image or an emitting image of the valuable document.

[0026] In the case of an optical sensor, an illumination device, which may be part of the sensor, can be used to illuminate a valuable document introduced into the detection area of ​​the optical sensor. This illumination device can be adapted to illuminate or optically excite the valuable document. The valuable document to be examined is introduced into the detection area of ​​the optical sensor in such a way that it can be illuminated or optically excited by the light from the illumination device, and that the light emitted from the valuable document can be detected by an optical image detection device. The light emitted from the valuable document can be reflected light, i.e., illumination light from the illumination device that is (directionally or non-directionally) reflected by the valuable document. Alternatively, the light emitted from the valuable document can be luminescent light, emitted by the valuable document as a result of light from the illumination device suitable for optical excitation. The term luminescence is understood to refer to the general concept of radiation, such as fluorescence or phosphorescence, emitted back from the valuable document after (optical or electrical) excitation.

[0027] In the first embodiment of the second variant, the optical sensor is a reflective sensor, and the image of the valuable document is a reflective image of the valuable document. An optical image detection device is configured to detect this image. In this case, the characteristic optical signal of the fiber is the reflective signal of the fiber. In the reflective image, the fiber appears black, gray, or colored against a background composed of a substrate of the valuable document or other anti-counterfeiting features of the valuable document. This background can be dark, light, or white, but it can also be structured or multicolored, as long as the reflective signal of the fiber is significantly different from it. In this case, the fiber can be luminous or non-luminous, magnetic or non-magnetic. The reflective sensor may optionally have an illumination device configured to illuminate the valuable document with illumination light.

[0028] In a second embodiment of the second variant, the fiber is a light-emitting fiber, and the optical sensor is a light-emitting sensor. Accordingly, the image of the valuable document is thus a light-emitting image of the valuable document, an optical image detection device is configured to detect this image, and the characteristic optical signal of the fiber is the light-emitting signal of the fiber. The light-emitting signal can be a fluorescent or phosphorescent signal. In this case, the background of the non-light-emitting valuable document substrate is typically dark. The light-emitting sensor may have an illumination device configured to illuminate the valuable document with excitation light emitted by the illumination device, so as to optically excite the light-emitting fibers of the valuable document. Alternatively, if the light emission is electroluminescence, the light emission of the fiber can also be electrically excited. The image detection device of the light-emitting sensor is an optical image detection device configured to detect the light-emitting image of the valuable document.

[0029] All the following descriptions refer to the first and second variants (in this case, two implementations of the second variant).

[0030] The analysis apparatus can be configured to identify one or more sub-images of a valuable document image, each sub-image containing a fragment of the valuable document image. These sub-images can be, for example, squares, rectangles, circles, or any other shape. These sub-images may contain unequal fragments of the valuable document image, which may partially overlap or not overlap each other. Overlapping multiple sub-images increases the probability that one of the used sub-images lies precisely within a potentially forged segment of the valuable document (whose location is not known beforehand). In this case, the local fiber characteristic values ​​of the forged segment deviate more from the expected values ​​compared to when the sub-image is only partially located within the forged segment. Therefore, in the case of overlapping sub-images, the associated forgeries are more easily detected.

[0031] Preferably, the corresponding sub-images are determined such that characteristic optical or magnetic signals of multiple fibers of the valuable document are expected to be present in the corresponding sub-images. The analysis apparatus is configured, for example, to determine at least one local fiber feature value for one or more sub-images, and to classify the valuable document as suspected of being counterfeit (or not suspected of being counterfeit) based on at least one local fiber feature value of one or more sub-images. The position or size of the sub-image within the valuable document image can be predetermined or dynamically selected, for example, based on the position of the fibers (previously located) or the transport position of the valuable document. Local fiber feature values, for example, characterize the local density of fibers in the corresponding sub-image and / or characterize the local distribution of fibers in the corresponding sub-image and / or characterize the position of one or more fibers in the corresponding sub-image and / or characterize the local properties of fibers in the corresponding sub-image.

[0032] In an extended design of the invention, the analysis apparatus, for analyzing valuable document images, is configured to systematically scan the valuable document images point-by-point using grid points, wherein a local fiber feature value is determined for each grid point of the valuable document image, and the valuable document is classified as potentially counterfeit based on the determined local fiber feature value. A grid point can be a single pixel of the valuable document image or multiple pixels of the valuable document image combined into a single grid point. Grid points have, for example, predefined positions on the corresponding valuable document image and are preferably uniformly distributed across the valuable document. The local fiber feature value determined for a corresponding grid point can be applied point-by-point to the corresponding grid point or to an environmental region of the corresponding grid point. The environmental region may also contain the corresponding grid point. If necessary, the corresponding grid point with its associated environmental region can correspond to a sub-image of the valuable document image.

[0033] For each grid point in a valuable document image, the determined local fiber feature value represents, for example, the density / location / property / distribution / quantity / distance of one or more fibers closest to the corresponding grid point. For instance, an average local fiber feature value can be determined over multiple grid points, and the dispersion of these local fiber feature values ​​near this average value, considering multiple grid points, is used to classify the valuable document as potentially forged. If the dispersion is too large, the valuable document is classified as potentially forged.

[0034] In an extended design of the invention, the analysis device, for analyzing a valuable document image, is configured to identify one or more prominent grid points that are prominent in terms of local fiber feature values ​​determined at the respective grid points, and particularly to identify one or more prominent regions of the valuable document image in which prominent grid points exist, based on the position of the prominent grid points on the valuable document image. Subsequently, based on one or more characteristics of the respective prominent regions and / or based on the local fiber feature values ​​determined for the grid points of the respective prominent regions (for this purpose, the local fiber feature values ​​of the prominent grid points can be used, and if necessary, the local fiber feature values ​​of non-prominent grid points located within the prominent regions can also be used), the analysis device classifies the valuable document as potentially counterfeit.

[0035] For example, grid points are classified (or graded) as conspicuous by comparing the local fiber characteristic values ​​determined for a given grid point with expected values ​​applicable to that corresponding local fiber characteristic value. If the deviation exceeds an acceptable fluctuation range, the corresponding grid point is classified as conspicuous. For example, if at least a minimum proportion of grid points within a conspicuous range are conspicuous, a valuable document can be classified as potentially forged. Alternatively, a valuable document can be classified as potentially forged based on the average local fiber characteristic value obtained by averaging across all grid points within the corresponding conspicuous range. Alternatively, the dispersion of these local fiber characteristic values ​​around the average value, considering multiple environmental regions within the conspicuous range, can also be used to classify a valuable document as potentially forged.

[0036] For example, a corresponding sub-image / corresponding environmental region can be identified as a prominent area. Alternatively, the prominent area of ​​the valuable document image can be determined dynamically based on the distribution of prominent grid points. In particular, the analysis device can be configured to determine the distribution of prominent grid points and, based on this distribution, determine one or more prominent areas for analyzing the valuable document image. Prominent sub-images can also be selected from a set of preset sub-images. Alternatively, based on this distribution, a prominent area with variable shape and size can be defined, for example, through image analysis, in which the proportion of prominent grid points in each row / column / segment / column section of the valuable document image is determined. If this proportion is greater than a reference value, the corresponding row / column / segment / column section is classified as prominent. Subsequently, prominent rows / columns / segments / column sections or grid points can constitute a prominent area, and the valuable document can be classified as potentially counterfeit based on the characteristics of this prominent area.

[0037] The characteristics of the prominent areas on which a valuable document is classified as suspected forgery are, for example, the difference between the local fiber feature values ​​at grid points of the prominent area of ​​the valuable document image and the expected values ​​applicable to the corresponding local fiber feature values, and / or the shape of the prominent area or the distribution of multiple prominent areas on the valuable document image, and / or the number of prominent areas found in the valuable document image. The expected values ​​may be predetermined or determined based on local fiber feature values ​​of other areas of the valuable document image or determined using the average local fiber feature value of the entire valuable document image.

[0038] In some embodiments, the analysis apparatus, for analyzing a valuable document image, is configured to determine at least one of a plurality of local fiber feature values ​​for a (area-like) environmental region belonging to a corresponding grid point, the local fiber feature value being determined for the corresponding grid point of the valuable document image. The corresponding grid point having the environmental region can be one of the aforementioned sub-images of the valuable document image. The environmental region can constitute a sub-image that moves during scanning. The environmental region belonging to the corresponding grid point is preferably located at least in proximity to the corresponding grid point, for example, directly next to or around the corresponding grid point, so that the corresponding grid point is located within the corresponding environmental region. The corresponding grid point is, for example, located at the edge or inside of its environmental region. The size of the environmental region can be fixedly preset or selected based on a predicted fiber density. The corresponding local fiber feature value is determined within the environmental region of the corresponding grid point. The environmental regions of adjacent grid points may or may not overlap.

[0039] In other embodiments, the analysis apparatus, for analyzing a valuable document image, is configured to determine one or more local fiber feature values ​​point-by-point for each corresponding grid point in the valuable document image. In this case, the local fiber feature value is determined only at the corresponding grid point itself. Accordingly, to determine the corresponding local fiber feature value, it is not necessary or necessary to define an environmental region of the grid point. For example, the local distance of the corresponding grid point relative to the nearest fiber or the subordination of the corresponding grid point to one or more interaction regions of the fiber (see below) can be used as the local fiber feature value.

[0040] Preferably, the analysis device is configured to determine each sub-image of a valuable document image such that a minimum number of characteristic optical or magnetic signals of fibers are expected to be present in the corresponding sub-image. In particular, the sensor can provide expected values ​​of fiber areal density for one or more valuable document types, respectively. This is especially preferred for valuable document types in which fibers are distributed substantially uniformly (but randomly) across the valuable document, either substantially uniformly across the entire document or at least substantially uniformly across the segment of the valuable document in which the sub-image is determined. To determine at least one sub-image of a valuable document image, the analysis device can be configured to select the area of ​​each sub-image of the valuable document image at least so largely as to account for the expected fiber density values ​​of the valuable document type being examined, such that at least the minimum number of characteristic optical or magnetic signals of fibers are expected to be present in the corresponding sub-image, taking into account the random statistical distribution of fibers on the valuable document.

[0041] For example, the analysis device can be configured to determine, when identifying at least one sub-image of a valuable document image, such that their areas F are each at least equal to the minimum quantity M divided by the expected density value DE: F ≥ M / DE. Therefore, when the expected fiber density value DE is small, the sensor determines the sub-image with the larger area. This ensures that the expected fiber quantity in the corresponding sub-image is sufficiently large to enable clear differentiation during inspection. Because for a genuine valuable document equipped with fibers, a fiber count of zero is highly unlikely. As the expected fiber count increases, the probability of misclassifying a genuine valuable document as a structural counterfeit decreases.

[0042] With a large expected value, the probability of a genuine valuable document with uniformly distributed fibers having very few (e.g., zero) fibers in the sub-image to be examined becomes lower. Otherwise, if only a small number of fibers are expected in a sub-image, it is possible that even for a genuine valuable document, fibers may be accidentally not found in the examined sub-image. Therefore, to avoid a genuine valuable document being unintentionally classified as suspected of being counterfeit due to the small number of fibers found in a sub-image, it is advantageous to define a sub-image with a relatively large expected value for the number of fibers and / or to determine two or more sub-images that cover the largest possible proportion of the valuable document.

[0043] Preferably, the analysis device is configured to determine at least two sub-images of a valuable document image (these sub-images are arranged, for example, side-by-side on the valuable document image), particularly at least two such sub-images of a valuable document image, which together cover preferably at least 50% of the area of ​​one facet of the valuable document, particularly at least 75% of the area of ​​one facet of the valuable document. These sub-images may be unevenly distributed on the valuable document, or they may be uniformly arranged, for example, row-by-row and / or column-by-column. In particular, these sub-images can be arranged on the valuable document in a grid consisting of at least two rows and / or at least two columns. The analysis device may also be configured to, when determining sub-images of a valuable document image, substantially comprehensively divide the valuable document image into multiple sub-images, for example, regardless of the type of valuable document.

[0044] In an extended design of the invention, a sensor is configured to inspect one or more specific types of valuable documents, each having at least one interfering (e.g., optical or magnetic) anti-counterfeiting element that interferes with or hinders the detection of characteristic optical or magnetic signals of the fibers. To inspect such valuable documents, the sensor's analysis apparatus is configured to classify a specific type of valuable document as potentially counterfeit based on at least one local fiber characteristic value at only one or more points in the image of the valuable document where the detection of the optical or magnetic signals of the fibers is not hindered by the specific (e.g., optical or magnetic) anti-counterfeiting element. "Not hindered" is understood to mean that the optical or magnetic signals of the fibers are unaffected by the specific anti-counterfeiting element or are affected at most only to such an extent that at least one local fiber characteristic value at the corresponding point can still be determined despite the possible presence of the specific anti-counterfeiting element at that location. Therefore, the following locations in the image of the valuable document are excluded or avoided: locations where the optical or magnetic signals of the fibers are physically shielded, for example, by absorptive or reflective anti-counterfeiting elements; or locations where there is interference with the optical or magnetic signals of the anti-counterfeiting elements. Preferably, only the following locations in the image of the valuable document are selected, or only the local fiber feature values ​​of these locations are analyzed, and these locations are outside of interfering anti-counterfeiting elements.

[0045] For example, during the preparation phase of valuable document inspection (e.g., during sensor adaptation), the sensor analysis device may receive information regarding the location of interfering (optical or magnetic) anti-counterfeiting elements on multiple valuable document types, if necessary. The analysis device can be configured to, based on the information regarding the location of one or more interfering anti-counterfeiting elements, determine the points or sub-images of the corresponding valuable document image to be analyzed during the analysis process such that the optical or magnetic signals of the fibers at the corresponding points or in the corresponding sub-images are not interfered with by the interfering (e.g., optical or magnetic) anti-counterfeiting elements. In particular, the position of sub-images on the valuable document (and, if necessary, the shape or size of the sub-images) can be determined such that these sub-images do not overlap with the interfering anti-counterfeiting elements.

[0046] In the case of optical sensors, when determining the points or sub-images to be analyzed, areas of the valuable document image that contain fluorescent or absorptive optical anti-counterfeiting elements, such as patches, fluorescent elements, or dark-printed areas of the valuable document, are often skipped or ignored. For example, the points or sub-images to be analyzed are preferably located in white and light-colored printed areas. In the case of magnetic sensors, when determining the points or sub-images to be analyzed, areas of the valuable document image that contain magnetic anti-counterfeiting elements with very large magnetic signals, such as magnetic security threads or magnetic printing areas, are often skipped or ignored. The magnetic signals of the magnetic printing areas are on the order of the magnetic signals of magnetic fibers. Here, these interfering magnetic anti-counterfeiting elements may be located on the valuable document surface facing the sensor or on the valuable document surface facing away from the sensor (but the magnetic signals of these magnetic anti-counterfeiting elements will still interfere).

[0047] The sensor's analysis device may also have information about the transport location (or transport posture) of the respective valuable documents to be examined and / or information about the type of valuable document (e.g., currency or denomination) of the respective valuable documents to be examined. The transport location can be one of four possible transport locations: the front of the valuable document facing up / down, or the prominently printed edge of the valuable document facing forward / backward. The analysis device can be configured to determine only points or sub-images of the valuable document image for analysis based on the information about the transport location and / or based on the information about the type of valuable document to be examined, in which the detection of optical or magnetic signals of the fibers is not hindered by any interfering anti-counterfeiting elements. In this case, the analysis device classifies the respective valuable documents to be examined as suspected counterfeit only based on the local fiber characteristic values ​​of the points or sub-images determined for analysis, i.e., only based on the local fiber characteristic values ​​of those points or sub-images where the detection of optical or magnetic signals of the fibers is not hindered by interfering anti-counterfeiting elements.

[0048] For example, during the preparation phase of valuable document inspection (e.g., during sensor adaptation), the sensor analysis device receives information regarding the location of one or more interfering (e.g., optical or magnetic) anti-counterfeiting elements for one or more valuable document types, based on the valuable document type and / or the transport location of the valuable document. The analysis device may, for example, be configured to, based on information about the valuable document type of each valuable document to be inspected and using information about the location of interfering (optical or magnetic) anti-counterfeiting elements of the corresponding valuable document type, determine only such points or sub-images of each valuable document to be inspected for analysis, in which the detection of optical or magnetic signals of the fibers is not hindered by the interfering anti-counterfeiting elements. Alternatively or additionally, the analysis device may be configured to, based on information about the transport location of each valuable document to be inspected and using information about the location of interfering (optical or magnetic) anti-counterfeiting elements of the corresponding type of valuable document, determine only such points or sub-images of each valuable document to be inspected for analysis, in which the detection of optical or magnetic signals of the fibers is not hindered by the interfering anti-counterfeiting elements.

[0049] Information regarding the transport location or document type of each valuable document to be inspected can be provided within the sensor itself (e.g., when the transport location or document type remains unchanged), for example, stored in the sensor's data memory and transmitted to the analysis device. Alternatively, this information can be input to the analysis device from outside the sensor, for example, from another sensor or device within the valuable document processing equipment that has identified the transport location or document type of each valuable document to be inspected, and stores the currently set transport location or document type on the valuable document processing equipment. Information regarding the location of a specific anti-counterfeiting element of the valuable document, related to the transport location and / or the document type, can also be provided within the sensor itself, for example, stored in the sensor's data memory, which the analysis device can access. Alternatively, this information can be input to the analysis device from outside the sensor, for example, from another sensor or device within the valuable document processing equipment.

[0050] In some embodiments, the analysis apparatus, for analyzing a valuable document image, is configured to assign (planar) action regions to each located fiber in the valuable document image (all or part of the fibers in the valuable document image, or each fiber in the corresponding environmental region of a corresponding sub-image / corresponding grid point). The corresponding fiber is preferably located inside (e.g., at the center) of its assigned action region. For analysis, the analysis apparatus locally analyzes the action regions of fibers located at and, if necessary, in the environment of the corresponding point in the valuable document image (e.g., for multiple different sub-images or grid point environmental regions), for each corresponding point in the valuable document image. This includes the action regions of fibers present in the corresponding sub-image or in the corresponding environmental region of the corresponding grid point. For example, this involves local analysis of the action regions with respect to surfaces covered or uncovered by the action regions or with respect to the overlap of adjacent action regions. Based on the local analysis of the (locally present) action regions located at and, if necessary, in the environment of the corresponding point, the analysis apparatus determines the corresponding (or individual) local fiber characteristic values ​​for each point (e.g., sub-image or environmental region).

[0051] For example, local fiber characteristic values ​​can be determined based on the overlap of the action area, the distribution of the action area, the location of the action area, and the area of ​​the action area.

[0052] The size and / or shape of the area of ​​effect can be the same for each located fiber in the valued document image. However, the size and / or shape can also be different in relation to the extent on the valued document. For example, these areas of effect can each have a circular, elliptical, or rectangular shape, with the corresponding fiber located at the center. Alternatively, the shape and / or size can also be chosen based on the fiber shape, for example, in the form of a tubular structure surrounding the corresponding fiber. The area of ​​effect is preferably defined as having an area smaller than that of a sub-image or environment region. In this case, specific local fiber feature values ​​can be determined more precisely. However, the sub-image or environment region can also be smaller than the area of ​​effect, for example, in cases where local fiber feature values ​​are determined only at the corresponding grid point, such as the local distance of the corresponding grid point relative to the nearest fiber, or where the corresponding grid point (yes / no) belongs to at least one area of ​​effect of the fiber.

[0053] In particular, the analysis apparatus is configured to select the size of the effective area based on a density expectation value, specifically an expected fiber areal density (if localized) for analyzing valuable document images. If the fibers are uniformly distributed across the valuable document, the density expectation value can be the same at every point in the valuable document image, or it can be applied only to the corresponding sub-image or environmental region. Therefore, a larger density expectation value defines a smaller effective area than a smaller one. Consequently, even for different types of valuable documents (with different density expectation values), it is easier to identify conspicuous areas of too few or too many fibers in the valuable document image by analyzing the effective area.

[0054] In one embodiment, the analysis apparatus is configured to analyze a valuable document image by separately analyzing the entire valuable document image, one or more sub-images, or at least one or more environmental regions of a raster point.

[0055] - Identify the covered area of ​​the functional region of at least one fiber belonging to the valuable document image, its corresponding sub-image, or its corresponding environmental region, and / or identify the free area of ​​the functional region of at least one fiber not belonging to the valuable document image, its corresponding sub-image, or its corresponding environmental region, and / or identify overlapping regions where the functional regions of at least two (adjacent) fibers of the valuable document image, its corresponding sub-image, or its corresponding environmental region overlap.

[0056] - Identify one or more prominent areas (e.g., sub-images, environmental regions) of a valuable document image that are prominent in terms of their covered / free / overlapping areas (e.g., in terms of their area / area ratio), and

[0057] - Valuable documents are classified as suspected forgeries based on the characteristics of the covered and / or free and / or overlapping areas in one or more conspicuous areas of the image.

[0058] The characteristics used to classify a valuable document as potentially forged include, for example, the area and / or area ratio of covered and / or free and / or overlapping regions within one or more prominent areas of the valuable document image, and / or the shape of the prominent areas or the distribution of multiple prominent areas on the valuable document image, and / or the number of prominent areas found in the valuable document image. A region / sub-image / environment area can also be classified as prominent based on the area of ​​the largest free and / or overlapping region found within it. A region is prominent, for example, if the area or area ratio of the covered / free / overlapping region within the corresponding prominent region is larger than expected. This can be determined by comparison with a predetermined reference value for area / area ratio, or by comparison with other regions (sub-images, environment areas) or with the average area ratio of covered / free / overlapping regions across the entire valuable document image.

[0059] In an extended design of the invention, the valuable document image detected by the image detection device is a first valuable document image, which is detected from a first valuable document face (front or back) of the valuable document to be inspected, and the analysis device additionally includes a second valuable document image, which is detected from a second face (back or front) of the same valuable document to be inspected opposite to the first face. In this case, the analysis device is configured to determine at least one local fiber feature value for at least one first sub-image of the first valuable document image and for at least one second sub-image of the second valuable document image, wherein the local fiber feature values ​​for the first and second sub-images preferably correspond to each other and are comparable to each other. For example, the local fiber feature values ​​of the first and second sub-images respectively characterize the local density of fibers in the respective first or second sub-image, or characterize the local distribution of fibers in the corresponding first or second sub-image, or characterize the position of one or more fibers in the corresponding first or second sub-image, or characterize the local properties of fibers in the corresponding first or second sub-image. The valuable document is then classified as potentially forged, particularly in terms of the presence of constructive forgeries, based on the local fiber feature values ​​of at least one first sub-image and at least one second sub-image. The first and second sub-images may be located in the same or different sections of the respective valuable document.

[0060] The local fiber feature values ​​of the first and second sub-images can be examined separately (e.g., compared with expected values ​​respectively), or the local fiber feature values ​​of the first and second sub-images can be combined or compared with each other. In the case of combination, for example, the average fiber density of the first sub-image and (if necessary, the corresponding) second sub-image can be determined, or the fiber counts of the first and second sub-images can be added together, and the valuable document can be classified as suspected of being counterfeit based on the average fiber density or based on the total number of fibers in the first and (if necessary, the corresponding) second sub-images. The analysis device can also compare the local fiber feature values ​​(e.g., density or average fiber distance) of the first sub-image and (if necessary, the corresponding) second sub-image with each other, and classify the valuable document as suspected of being counterfeit based on the comparison results, for example, if the deviation between them is greater than an acceptable fluctuation range. This allows for the detection of counterfeit documents with fibers only on one side (or one side of the image).

[0061] The selection of the sub-image, or the area of ​​a possible sub-image on one side of a valuable document, may be limited by other anti-counterfeiting elements or dark printing areas on the valuable document. By analyzing at least one sub-image on both the front and back sides, a larger area of ​​the valuable document is examined, and the examination becomes more reliable compared to examining only one side. Therefore, by analyzing sub-images on both the front and back sides, it is also possible to reliably examine valuable documents for which only small sub-images are feasible or for which the valuable document has relatively few fibers.

[0062] In a particularly advantageous variant of double-sided analysis, the analysis apparatus is configured to determine a corresponding second sub-image of a second valuable document image for at least one first sub-image of a first valuable document image. This second sub-image is a fragment of the second valuable document image, wherein the first sub-image and the corresponding second sub-image contain optical or magnetic signals of the front and back sides of the same valuable document segment of the same valuable document. The analysis apparatus, for example, is configured to determine corresponding image points of the second valuable document image for image points of the first valuable document image. To discover a second sub-image (e.g., the back side) corresponding to the first sub-image (e.g., the front side), it may be necessary to mirror the first or second sub-image (front or back side image) and, if necessary, to slightly rotate and / or move these images relative to each other.

[0063] The analysis apparatus is further configured to determine at least one local fiber feature value for each of one or more sub-image pairs consisting of a first sub-image and a corresponding second sub-image, wherein, preferably, local fiber feature values ​​that correspond to each other are determined for the first and second sub-images, these local fiber feature values ​​are comparable to each other, and valuable documents are classified as, for example, suspected of being forgeries in terms of the presence of structural forgeries, based on the local fiber feature values ​​of the first sub-image and the corresponding second sub-image. The local fiber feature values ​​of the first sub-image and the corresponding second sub-image can be examined separately from each other, or the first sub-image and the second sub-image can be combined or compared to each other (see examples above).

[0064] More precise examination of valuable documents is achieved by analyzing corresponding sub-images of the same valuable document segment on both the front and back sides. Since these two corresponding sub-images depict the same valuable document segment (e.g., the front and back of the same sheet of paper), this two-sided analysis not only increases the area examined but is also (due to the spatial consistency of these two sub-images on the valuable document) more convincing than analyzing any two random sub-images of the same valuable document. This is because, for any two sub-images of the same valuable document, in the case of a constructive forgery, it can be expected that these two sub-images are not uniformly located in the forged segment or uniformly located in the genuine segment; rather, in many cases, one sub-image is located in the forged segment of the constructive forgery and the other in the genuine segment. Therefore, the combined local fiber characteristic value (e.g., average value) of such two sub-images will be closer to the expected value of a genuine valuable document than in the case of corresponding two sub-images.

[0065] However, when analyzing two corresponding sub-images of the front and back sides containing identical valuable document segments, it is certain that if one sub-image (e.g., the front sub-image) is located within the forged segment of the constructive forgery, then the corresponding other sub-image (the back sub-image) is also located within the forged segment of the constructive forgery. Therefore, it is more likely that both corresponding sub-images are located within the forged portion of the constructive forgery. In this case, the combined local fiber feature value of these two corresponding sub-images will deviate significantly from the expected value. As a result, a greatly improved discriminative power is achieved in forgery identification through bi-directional analysis of the same valuable document segment.

[0066] A pair of corresponding sub-images consisting of a first sub-image and a corresponding second sub-image can be classified as suspicious based on two local fiber feature values. Furthermore, a valuable document can be classified as suspicious based on the number of such pairs of sub-images consisting of a first sub-image and a corresponding second sub-image of the corresponding valuable document, particularly based on the number of such pairs. For example, if the number of such sub-image pairs of the same valuable document is at least 1 or 2, then the valuable document is classified as suspicious.

[0067] The second valuable document image can be detected using the same image detection device, wherein the first and second valuable document images are detected sequentially, for example, when the valuable document is transported again next to the optical sensor in a flipped transport position. Alternatively, the second valuable document image can also be detected using a separate image detection device (if necessary, simultaneously with the first valuable document image), said separate image detection device being arranged, for example, on a side opposite to the aforementioned image detection device.

[0068] In an extended design of the invention, the analysis device is configured to compare local fiber feature values ​​with a (fiber-related) expected value E for one or more points (sub-images, grid points, or environmental regions of grid points) of a valuable document image. The analysis device then classifies the valuable document as potentially forged based on the comparison results obtained for the one or more points of the valuable document image. For example, the analysis device is configured to classify the valuable document as potentially forged based on the difference between the corresponding local fiber feature value at the corresponding point and the expected value determined for the corresponding point during the comparison. The expected value may be different for different points (sub-images, grid points, or environmental regions of grid points) of the valuable document image.

[0069] For example, the analysis device identifies those points (e.g., sub-images, raster points, or the surrounding area of ​​raster points) whose local fiber characteristic values ​​deviate from corresponding expected values ​​(e.g., exceeding acceptable fluctuation ranges). The acceptable fluctuation range is preferably at least 20%, particularly preferably at least 40%, of the corresponding expected value. The valuable document is classified as potentially counterfeit based on the number and / or location and / or distribution of these points (sub-images, raster points, or the surrounding area of ​​raster points) whose local fiber characteristic values ​​deviate from the expected values ​​(exceeding acceptable fluctuation ranges). In particular, the spatial arrangement of the points (sub-images, raster points, or the surrounding area of ​​raster points) deviating from the expected values ​​can be compared with a known nominal arrangement for the corresponding valuable document, and the valuable document can be classified as potentially counterfeit based on this comparison.

[0070] As an expected value, one could use, for example, an expected value for the number of fibers, which gives the expected number of fibers for a given location if the valuable document to be examined is a true valuable document of the corresponding valuable document type. Alternatively, a density expected value could be used, which gives the expected fiber density in the corresponding sub-image. The expected value can be a natural number, a positive real number, or zero. Different expected values ​​can be used for fibers with different properties, such as fibers with different colors.

[0071] If the expected value for fiber quantity or fiber density is greater than zero, then (as an alternative to quantitatively determining the difference from the expected value) it is also possible to simply check whether the fiber quantity or fiber density in the corresponding sub-image is equal to zero (or greater than zero). This is because even though the expected value for fiber quantity or fiber density is greater than zero, a fiber quantity or fiber density equal to zero is sufficient to determine forgery suspicion. In this case, the magnitude of the difference from the corresponding expected value is not important. For example, to examine a valuable document whose fibers are distributed (non-uniformly or substantially uniformly) across the entire document (i.e., the expected value for fiber quantity or fiber density is greater than zero everywhere), the analysis apparatus can be configured to classify the corresponding valuable document as suspected forgery if at least a minimum number of sub-images have a fiber quantity of zero, where the minimum number is >0.

[0072] The expected value can be predefined for the corresponding valuable document or for a specific point, for example, determined before the valuable document is examined, especially depending on the type of valuable document. Different expected values ​​are typically predefined or determined for different valuable document types. Alternatively, the expected value can be determined during the valuable document examination, for example, based on local fiber feature values ​​determined for multiple points in the corresponding valuable document image. In the case of a predefined expected value, it can be stored in the sensor's analysis device, for example, in the analysis device's data memory. However, the predefined expected value can also be stored externally, for example, in a valuable document processing device, and transmitted from there to the analysis device.

[0073] As an alternative or supplement to checking based on predefined expected values, the analysis apparatus can also be configured to determine an expected value for comparison with the local fiber feature values ​​of at least one point (sub-image, raster point, or the surrounding area of ​​a raster point) based on local fiber feature values ​​found for one or more other points (sub-image, raster point, or the surrounding area of ​​a raster point) of the same valuable document image. Thus, relative checks can be performed on the local fiber feature values ​​of at least two different points (sub-image, raster point, or the surrounding area of ​​a raster point) of the same valuable document image. This is particularly advantageous for checking valuable documents where the fibers are substantially uniformly distributed across the entire document. This is because, in this case, it is not necessary to store, for example, predefined individual expected values ​​for newly published valuable documents. For example, the average local fiber feature value of multiple other points of the same valuable document image can be used as the expected value.

[0074] In particular, the analysis device can be configured to perform relative analysis on multiple sub-images of the same valuable document with respect to local fiber feature values ​​determined in these sub-images. Here, the local fiber feature values ​​of a sub-image can be compared with the local fiber feature values ​​of one or more other sub-images of the same valuable document, and the valuable document can be classified as potentially counterfeit based on the comparison results, for example, if there is a strong discrepancy between the local fiber feature values ​​of different sub-images of the same valuable document. For example, the analysis device can be configured to calculate an expected value for comparison with the local fiber feature values ​​of at least one sub-image, based on local fiber feature values ​​found at one or more other locations / sub-images of the same valuable document. For example, the analysis device can determine the average areal density of the fibers of the valuable document based on one or more sub-images of the same valuable document and use this average areal density as an expected density value. For example, a ratio of local fiber feature values ​​determined for different locations / sub-images can also be formed, and the banknote can be classified as potentially counterfeit based on this ratio.

[0075] The analysis device can progressively classify individual points (sub-images, raster points, or environmental regions of raster points) of a valuable document under inspection as either suspected of forgery or not. For example, the analysis device can be configured to classify corresponding points as suspected based on fiber characteristic values ​​applicable to those points, such as if the fiber characteristic value is significantly different from expected values, for example, if the difference exceeds a threshold. Otherwise, the corresponding point is classified as not suspected. Subsequently, the analysis device classifies the corresponding valuable document as suspected of forgery (or not suspected of forgery) based on the number of points classified as suspected in the corresponding valuable document. In particular, the analysis device can be configured to classify a valuable document as suspected of forgery if the number V of points (sub-images, grid points, or environmental areas of grid points) classified as suspicious in the corresponding valuable document reaches or exceeds a minimum number P, i.e., if V≥P, where the minimum number P>0 and preferably 1 or 2, then the corresponding valuable document is classified as suspected of forgery.

[0076] The acceptable fluctuation range S can be determined from the expected value E, for example, by pre-setting it proportionally to the expected value. However, to determine the acceptable fluctuation range S, the local fiber characteristic value of at least one fixed pre-set point of the corresponding valuable document can also be determined before the valuable document inspection (for multiple (e.g., 100) genuine valuable documents of the corresponding valuable document type). By averaging these local fiber characteristic values ​​of the multiple valuable documents, an average local fiber characteristic value for the corresponding valuable document type and, if necessary, for the corresponding point can be determined, and this average local fiber characteristic value is then used as the expected value when inspecting valuable documents of that valuable document type. The genuine valuable documents included in the multiple valuable documents do not necessarily belong to the same denomination, but only nominally have the same fiber distribution, for example, are made of the same fiber-containing substrate. Additionally or alternatively, the standard deviation s of the local fiber characteristic values ​​for the corresponding valuable document type and, if necessary, the corresponding point can also be determined from the local fiber characteristic values ​​determined separately for the multiple genuine valuable documents of the corresponding valuable document type. The acceptable fluctuation range S for the corresponding valuable document type can then be selected based on the standard deviation s of the local fiber characteristic values. For acceptable fluctuation ranges, such a value can be determined as at least twice, preferably at least four times, the standard deviation s.

[0077] In an extended design of the invention, the sensor provides expected values ​​at corresponding points (e.g., sub-images, grid points, or environmental regions of grid points) of valuable document images for one or more valuable document types. The analysis device has information about the valuable document type of each valuable document to be examined. The analysis device can be configured to determine, (based on the valuable document type of each valuable document to be examined and using the expected values ​​of local fiber feature values ​​at the corresponding points provided for one or more valuable document types) the expected values ​​of local fiber feature values ​​at the corresponding points of each valuable document to be examined, which are expected for the valuable document type of the valuable document to be examined at the corresponding points. The information about the valuable document type of the valuable document to be examined can be provided within the sensor itself, for example, stored in the sensor's data memory, which the analysis device can access, or this information can be input to the analysis device from outside the sensor, for example, from another sensor or from a valuable document processing device. For example, if a stack of valuable documents with the same valuable document type is examined, the information about the valuable document type of the valuable document to be examined can be the same for multiple valuable documents to be examined sequentially. However, information about the type of valuable document can also vary from one valuable document to the next, and this information can be transmitted individually, i.e. dynamically, from other sensors that determine the type of valuable document to the sensor according to the invention for each valuable document.

[0078] When the analysis device has information about the type of the valuable document to be examined, it can determine the points (sub-images, grid points, or environmental areas of grid points) of the valuable document image based on the type of the valuable document to be examined, and corresponding local fiber feature values ​​are determined for these points. For example, these points can be specifically determined in unprinted or lightly printed areas of the valuable document based on the type of the valuable document. In particular, a number of points are fixedly predefined in the analysis device, and the analysis device determines specific points from these fixedly predefined points for analysis based on the type of the valuable document to be examined. These points can be selected, for example, to utilize as large an area as possible of the unprinted / lightly printed area of ​​the corresponding valuable document. Subsequently, the valuable document is classified as suspected of being counterfeit only based on the local fiber feature values ​​of these points, while other points not determined for analysis are still not considered. In particular, the position and / or area and / or shape of the sub-image can be selected based on the type of the valuable document to be examined. The position and / or shape and / or area of ​​a sub-image can also be the same for multiple valuable file types, such as multiple denominations of the same currency.

[0079] To determine the location of a point to be analyzed on a document image, dynamic information about the document's transport location can be used, for example, to specifically identify unprinted or poorly printed areas. However, if the document's transport location is fixed and preset (as is the case with some document processing equipment), the location of unprinted or poorly printed areas is known in advance, and dynamic information about the document's transport location is not needed to determine the point's location.

[0080] Specifically, for inspecting valuable documents where the fibers are substantially uniformly distributed across the entire document, the sensor can possess information regarding the average fiber count or average fiber areal density for one or more valuable document types. This information can vary for different valuable document types. The analysis device can be configured to determine or identify the average fiber count or average fiber areal density of each valuable document to be inspected based on the information about the valuable document type, and to determine the area of ​​a sub-image of the valuable document to be inspected based on the average fiber count or average fiber areal density. For example, the area can be determined so large based on the average fiber count or average fiber areal density that a minimum number of fibers is expected in the corresponding sub-image. In the case of a larger areal density, a sub-image with a relatively small area is sufficient, while in the case of a smaller areal density, a sub-image with a relatively large area is required. The information regarding the average fiber count or average fiber areal density on the valuable document, possessed by the analysis device, can be provided in the sensor or in the analysis device itself, or can be transmitted from the valuable document processing equipment.

[0081] In an extended design of the invention, the sensor is configured to record, for example, unique fingerprint data of the corresponding valuable document, which can be used to track the valuable document, and the sensor itself is also used to track the valuable document when necessary. For example, tracking of valuable documents, especially for counterfeit or suspected counterfeit valuable documents, is performed to determine the processing information of the corresponding valuable document, which can indicate the source or custodian of the corresponding (e.g., counterfeit or suspected counterfeit) valuable document.

[0082] Sensors, particularly their analysis devices, can be configured to determine unique fingerprint data for a given valuable document based on one or more characteristics of fibers located in an image of the valuable document (e.g., relating to fiber position, distance, shape, size, color, or brightness) and / or based on one or more local fiber feature values ​​already determined by the sensor or its analysis device for one or more locations in the image of the valuable document (and, if necessary, the respective environments of these locations). These unique fingerprint data are unique to the given valuable document and are generally different for different valuable documents of the same type. Sensors, such as their analysis devices or other means of the sensor, can be configured to assign the unique fingerprint data of the given valuable document to the processing data of the given valuable document (e.g., relating to the source and / or face value of the valuable document and / or the time point and, if necessary, the location of the valuable document check performed by the sensor), and to store and, if necessary, these unique fingerprint data together with the processing data of the given valuable document, for example, in a list of multiple valuable documents, corresponding to each other respectively. The processing data of the given valuable document can be determined at least in part by the sensor itself or has been provided to the sensor externally.

[0083] The unique fingerprint data of a corresponding valuable document may involve, for example, the overall image of the valuable document or a corresponding point in the image of the valuable document and, if necessary, the respective environment of the corresponding point, such as the number and / or location and / or chromatic coordinates and / or size and / or local density and / or fluorescence or phosphorescence properties of fibers in one or more sub-images of the valuable document image or one or more grid points with environmental regions. The unique fingerprint data may, for example, include a vector whose coordinates are composed of the positions of all fibers in the valuable document image. Alternatively, the unique fingerprint data may also include a vector whose different coordinates are given by local fiber feature values, such as local fiber density, of different points and, if necessary, the environment of the different points in the valuable document image (e.g., sub-images or grid points with environmental regions). However, the different coordinates of the vector may also be composed of different types of local fiber features of one or more points. Alternatively, as a vector, only a single local fiber feature value or a single numerical value for the corresponding property of a fiber may be used as the unique fingerprint data, for example, using the average of the respective properties (e.g., chromatic coordinates, size, distance) of all fibers located in the valuable document image.

[0084] In order to be used for tracking, the sensor may have an operating mode (different from the normal inspection mode) for tracking valuable documents (genuine, counterfeit, or suspected of counterfeiting). In this operating mode, the analysis device is configured to determine unique fingerprint data of the valuable document specified for tracking. The sensor records an image of the valuable document from the document, and this determination is based, for example, on the characteristics of fibers located in the image of the valuable document specified for tracking and / or on one or more local fiber feature values ​​of the valuable document specified for tracking. A list is available for the sensor to use for tracking, in which multiple unique fingerprint data of (counterfeit / suspected) valuable documents are configured (or assigned) with processing information such as the source of the corresponding valuable document and / or the face value and / or the time and, if necessary, the location of the valuable document inspection. This list may be generated at least in part by the sensor itself, or the list may have been input into the sensor from an external source. Based on the unique fingerprint data of the valuable document specified for tracking, the sensor can, in particular, determine the processing data of the valuable document specified for tracking, especially regarding the source / denomination / time / location of the valuable document inspection, by consulting a list, and output this processing data as the result of the tracking.

[0085] The present invention also relates to a valuable document processing apparatus having the described sensor and a transport device that introduces the corresponding valuable document into the detection area of ​​the sensor. Alternatively, the valuable document may also be manually introduced into the detection area of ​​the sensor.

[0086] The present invention also relates to a method for examining valuable documents, particularly by means of the described sensor, consistent with the foregoing description, wherein the valuable document has a plurality of fibers distributed on the valuable document. In this method, the valuable document is introduced into the detection area of ​​the sensor, and a valuable document image of the valuable document is detected by means of an image detection device of the sensor, wherein the valuable document image contains characteristic optical or magnetic signals of the fibers. For analyzing the valuable document image, the analysis device of the sensor is configured to locate the fibers contained in the respective valuable document image and determine at least one local fiber feature value for one or more different points in the valuable document image, the local fiber feature value being applicable to the corresponding point and, if necessary, to the environment of the corresponding point in the valuable document image. The analysis device is further configured to classify the valuable document as potentially counterfeit based on the local fiber feature value determined for said one point / say different points in the valuable document image.

[0087] To inspect a valuable document having luminescent fibers distributed on it, the valuable document is introduced into the detection area of ​​an optical sensor, and the valuable document is illuminated there with excitation light emitted by the sensor's illumination device to optically excite the luminescent fibers, or correspondingly electrically excite the valuable document in the detection area. The luminescent image of the valuable document, containing the emission of light from the luminescent fibers, is then detected by an image detection device of the sensor. Subsequently, the optical sensor's analysis device performs the aforementioned analysis of the luminescent image.

[0088] The invention is described below by way of example with reference to the accompanying drawings. In the drawings:

[0089] Figure 1 This diagram illustrates the structure of a valuable document processing device with an optical sensor for inspecting valuable documents.

[0090] Figures 2a-2c Show real banknotes ( Figure 2a , Figure 2b ) and related counterfeit goods ( Figure 2c ) fiber distribution and possible sub-images,

[0091] Figures 3a-3c Show real banknotes ( Figure 3a , Figure 3d ) and related counterfeit goods ( Figure 3b , Figure 3c , Figure 3e , Figure 3f The fiber distribution and predefined raster of sub-images in the image.

[0092] Figures 4a-4c Show real banknotes ( Figure 4a , Figure 4c ) and related counterfeit goods ( Figure 4b , Figure 4d The fiber distribution and predefined raster of sub-images in the image.

[0093] Figures 5a-5b Showing the fiber distribution on opposite sides of a genuine banknote and corresponding sub-images.

[0094] Figures 6a-6c This shows a luminescent image systematically scanned within a grid of grid points. Figure 6a ), where each positioned fiber is assigned an effective area ( Figure 6b ), where an environment region is defined around each grid point ( Figure 6c ).

[0095] In one embodiment, banknotes with luminescent (e.g., fluorescent or phosphorescent) fibers are considered as valuable documents to be inspected, and an optical sensor records and analyzes the luminescent image of the banknote. However, the invention also relates to the inspection of banknotes with reflective or magnetic fibers, or to the inspection of other valuable documents with luminescent, reflective, or magnetic fibers, wherein, for inspection of magnetic fibers, a magnetic sensor is used, which accordingly records and analyzes the magnetic signal image of the valuable document.

[0096] Figure 1 An exemplary schematic structure of a valuable document processing device 1 is shown, which has an input cell 2 in which a stack of banknotes 3 to be processed is provided. The valuable document processing device also has a sorter 8, which sequentially grabs one banknote from the input stack and transfers it to a transport device 10 (conveyor belt and / or transport roller) (shown only schematically), which transports the banknotes along the transport direction x past the optical sensor 25 for inspection.

[0097] The optical sensor 25 includes an optical image sensor 20 that converts the intensity of light emitted by a banknote being transported alongside it into a corresponding sensor signal. The optical excitation of the banknote's light emission is achieved, for example, by excitation light sources 27 and 28 arranged on either side of the image sensor 20. However, only one light source may be used. The image sensor 20 has, for example, a sensor row or multiple sensor rows for different spectral components of the emitted light. These one or more sensor rows are arranged transversely to the banknote's transport direction x. The image sensor 20 is operated by a control device (not shown) such that it detects the banknote's light emission at multiple detection time points to perform an optical scan of the banknote being transported alongside it. During this process, adjacent detection areas of the banknote arranged along the transport direction are detected sequentially in time. The sequentially detected detection areas of the banknote correspond to image points of the emitted image.

[0098] Viewed along the banknote's transport direction x, optical sensor 25 is arranged on the left side of the transport path. Opposite to optical sensor 25 on the right side of the transport path, another optical sensor 29 may be arranged. This other optical sensor also has an optical image sensor (not shown), and if necessary, an illumination device for optical excitation of the banknote's light emission, and if necessary, its own analysis device. If necessary, the luminescent image recorded by the optical image sensor of the opposite optical sensor 29 is transmitted to the analysis device 19 of optical sensor 25 so that the two luminescent images of the same banknote can be analyzed together.

[0099] Image sensor 20 transmits the recorded luminescent image to analysis device 19 of optical sensor 25. Analysis device 19 may be contained within the housing of optical sensor 25, or it may be external to the housing, for example, contained within valuable document processing device 1. Analysis device 19 determines corresponding local fiber feature values ​​for one or more points, such as one or more sub-images, of the luminescent image recorded by image sensor 20. The authenticity of the banknote is checked based on these local fiber feature values, and, if necessary, the banknote is classified as suspected of being counterfeit.

[0100] The data storage 26 of the analysis device 19 can store information about raster points or sub-images and, if necessary, expected values ​​for one or more valuable file types. The control device 50 of the device 1 can transmit information to the analysis device 19 about the respective valuable file types to be examined and, if necessary, the transport location of the valuable file 3.

[0101] Based on the authenticity of the corresponding banknotes determined by the analysis device 19, the transport device 10 and the switches 4 and 5 along the transport route are controlled by the control device 50 such that the banknotes are input into one of a plurality of output slots 30 and 31 and stored there. For example, banknotes identified as genuine are stored in the first output slot 30, while banknotes suspected of being counterfeit are stored in the second output slot 31. At the end of the transport route shown (reference numeral 6), additional output slots and / or other devices for storing or destroying banknotes and / or rejection slots may be provided, where the banknotes are stored for individual handling, for example, by an operator.

[0102] In the example shown, the valuable document processing device 1 also includes an input / output device 40 for an operator to input data and / or control instructions, for example via a keyboard or touch screen, and for outputting or displaying data and / or information about the processing, particularly about the banknotes being processed.

[0103] Example 1

[0104] Figure 3a The illuminating fiber distribution of a genuine 10-yuan banknote (denomination A) is shown as an example. Figure 3b An example of the fiber distribution of a structural counterfeit 66 for this banknote 80 is shown. This structural counterfeit 66 is composed of fibers with an average fiber density of 1 / cm². 2 The composition consists of part 82 of a genuine banknote and part 62 of a counterfeit banknote that is completely free of fibers.

[0105] For analysis, fibers are first located in the luminescent image recorded by optical sensor 25. The luminescent image is then systematically scanned within a grid of, for example, 2×4 grid points. In a first embodiment, fiber density is used as a local fiber characteristic value. For each grid point, a local fiber density is determined in the surrounding environment corresponding to a sub-image of the banknote image. Figure 3d , Figure 3e , Figure 3f In the example, the environmental regions / sub-images 11-14 and 21-24 are distributed on the banknote in a 2x4 grid. These eight grid points are located, for example, at the center of the respective environmental region or sub-image. In the analysis device 19 of the optical sensor 25, the grid of sub-images 11-14 and 21-24 is fixedly preset for multiple banknote types. The area of ​​the sub-image is, for example, F = 8 cm². 2 Suppose that optical sensor 25 is set up to examine banknotes of currencies A and D, wherein all denominations have fibers that are uniformly but randomly distributed on the banknote, and these fibers have a specific average surface density (density expectation value).

[0106] Optical sensor 25 receives or obtains information about the type of banknote to be inspected from the valuable document processing device 1 or from other sensors of the valuable document processing device 1. A table stored in the data memory 26 of the analysis device 19 specifies which preset sub-images 11-14, 21-24 should actually be inspected for a given banknote type (see Table 1). According to Table 1, for currency A, denomination 10, all sub-images in columns 1 and 2 should be inspected; for currency A, denomination 50, only 7 out of the possible 8 sub-images should be inspected (sub-image 24 is omitted). Using the information about the type of banknote to be inspected provided to optical sensor 25, analysis device 19 can select one or more of the possible sub-images for analysis based on the banknote type.

[0107] Table 1, which is stored in the database, also provides the expected density value DE of the fiber for the corresponding banknote type.

[0108] Density Expected Value DE Sub-image to be inspected Currency A, denomination 10 <![CDATA[1 / cm 2 ]]> 11-14、21-24 Currency A, denomination 50 <![CDATA[1 / cm 2 ]]> 11-14、21-23 Currency D, denomination 10 <![CDATA[5 / cm 2 ]]> 11-14、21-24 Currency D, denomination 50 <![CDATA[2 / cm 2 ]]> 11-14、21-24

[0109] Table 1 shows the face values ​​of currencies A and D, 10 and 50 respectively.

[0110] The acceptable fluctuation range S of the fiber density can be calculated based on the expected density value and area, or it can be fixedly preset, for example, S = 0.5 / cm². 2 .

[0111] Based on the recorded images of valuable documents, the analysis unit 19 of the optical sensor 25 determines the number of fibers N in each sub-image to be examined and calculates the corresponding fiber density D = N / F using the area F. Subsequently, the analysis unit compares this fiber density with the corresponding expected density value DE for each sub-image to be examined; for currency A, DE = 1 / cm². 2 Here, check whether the fiber density D is within the range of DE+ / -S. For currency A, denominations of 10 and 50, check, for example, according to Table 1, whether the fiber density in the sub-image is within E+ / -S = 1+ / -0.5 / cm. 2 Within the range, i.e., whether it is within 0.5 / cm 2 Up to 1.5 / cm 2 If the fiber density D is greater than or less than DE+ / -S, the corresponding sub-image or banknote is classified as not suspected.

[0112] Figure 3b , Figure 3e The structural counterfeit 66 in the table is a structural counterfeit targeting currency A, denomination 10. For currency A, denomination 10, all sub-images 11-14 and 21-24 are examined according to Table 1. No fibers were found in sub-images 11 and 21 (fiber density D = 0), and therefore these sub-images are classified as suspicious. For sub-images 12 and 22, the fiber density D is still within the acceptable range DE+ / -S, and therefore these sub-images are not classified as suspicious. Furthermore, for sub-images 13, 14, 23, and 24, fiber densities D within the range DE+ / -S were also found, and these sub-images are classified as not suspicious. Since the two sub-images 11 and 21 are classified as suspicious, Figure 3b , Figure 3e The structural counterfeit item 66 in the list is classified as suspected of being counterfeited.

[0113] Figure 3c , Figure 3f The counterfeit 67 in the image is a counterfeit of currency A, denomination 50. Both the genuine banknote and the counterfeit 67 have luminescent security elements in region 68. In the luminescent image, this security element outperforms the luminescence of the luminescent fibers. The analysis device 19 of the optical sensor 25, from Table 1, shows that for currency A, denomination 50 (different from currency A, denomination 10), the fiber density is only checked in sub-images 11-14 and 21-23, but not in sub-image 24 (located in the region of luminescent element 68). The inspection results of sub-images 11-14 and 21-23 are consistent with... Figure 3b , Figure 3e The results of the inspection were consistent with those of counterfeit item 66.

[0114] If the transport location of the banknote is variable, the sensor can receive information about the transport location and determine which sub-images should be inspected based on that location. For this purpose, information about the sub-images to be inspected can be stored in Table 1 for up to four transport locations for each banknote type. During sensor adaptation, to establish such an expanded Table 1, it can be determined which of the sub-images 11-14, 21-24 the luminescent element 68 is located in at the corresponding transport location and therefore must be omitted during inspection. For example, for the different transport locations of the structural counterfeit 67, where the luminescent element 68 is located in the lower left of the luminescent image, sub-image 21 is omitted instead of sub-image 24.

[0115] Example 2

[0116] Figure 2a The illuminating color-mixed fibers of a genuine banknote 70 of currency B are shown exemplarily. The banknote 70 has color-mixed fibers only in the left half; more precisely, these color-mixed fibers have a density of 1 / cm². 2 The average surface density. The right half of banknote 70 has no mixed-color fibers. In this example, the sensor does not receive information about the type of banknote to be inspected, but is instead fixed to currency B, which has this fiber distribution across all denominations.

[0117] Regardless of the banknote type, in the analysis device 19, a unique sub-image 11 is always determined in the left half of the banknote for all denominations of currency B, see [reference]. Figure 2b The area of ​​this sub-image can be fixed and preset, for example, preset to 10cm. 2 However, it is also feasible to select the area or size based on a known density expectation value for currency B (given by a known average areal density), which is stored in the analysis device 19. Assuming the expected value E should be 10, then with a density expectation value of 1 / cm²... 2 In this case, a 10cm diameter should be selected. 2 Sub-image of area 11.

[0118] For analysis, fibers are first located in the recorded emission images. In the second embodiment, the number of fibers located in sub-image 11 is used as a local fiber feature value. Figure 2c The image shows a counterfeit banknote 64 for currency B, which has absolutely no mixed-color fibers. Therefore, the fiber count N = 0 is determined in sub-image 11. However, since a fiber count exceeding zero (E > 0) is expected for currency B in sub-image 11, [the following is omitted as it is not explicitly stated in the original text]. Figure 2c 64 of the counterfeit items were classified as suspected of being counterfeit.

[0119] To also detect counterfeit banknotes with incorrect fiber counts, a more detailed examination can be performed, in which the fiber count N of the corresponding sub-image is compared with an expected value E. In this case, for the banknote to be examined, the number N of mixed-color fibers in sub-image 11 is determined and compared with a predetermined expected value, for example, E = 10. Preferably, this comparison does not require exact consistency with the expected value, but allows an acceptable fluctuation range S around the expected value E, for example, S = 5. In this comparison, it is checked whether the fiber count N is within the range of E + / - S. If not, the banknote is classified as suspected of being counterfeit; if so, it is classified as not suspected of being counterfeit.

[0120] In a specific example, with E=10 and S=5, the banknotes are classified according to the number of fibers N as follows:

[0121] - If it does not contain or contains at most 4 mixed-color fibers (such as in Figure 2c (As in case 64 of the counterfeit items): suspected of being counterfeit.

[0122] -If the mixed fiber count is 5 to 15 (as in...) Figure 2a , Figure 2b In cases where there are 70% genuine banknotes: there is no suspicion of counterfeiting.

[0123] - If the count of mixed-color fibers exceeds 15: there is suspicion of forgery.

[0124] When the transport position of the banknote in the valuable document processing device is variable, information about the transport position can be transmitted to the optical sensor 25. Based on this information, the optical sensor determines the position of the sub-image 11 such that the sub-image 11 is located within the banknote segment containing fibers in the case of a genuine banknote 70. Therefore, in Figure 2a , Figure 2b In the transport position of the banknote 70, sub-image 11 is placed in the left half of the valuable document image. In different transport positions of the banknote 70 where the fiber is on the right side of the valuable document image, sub-image 11 is placed in the right half of the valuable document image.

[0125] 3rd Embodiment

[0126] exist Figure 4a The image shows a genuine banknote of currency C, denomination 50, which has fibers that are roughly evenly distributed on the banknote, and these fibers have a density of 0.66 / cm². 2 The average face density (expected density value) applies to all denominations of currency C.

[0127] and Figure 4bAn example of the fiber distribution of a structurally counterfeit 65 for banknote 90 is shown, wherein a portion 92 of the genuine banknote is combined with a counterfeit portion 63, in which a lower fiber density exists than in the genuine banknote. The genuine portion 92, for example, has a fiber density of 0.67 / cm². 2 The average fiber density is 0.1 / cm², while the counterfeit part 63 has 0.1 / cm². 2 The average fiber density.

[0128] Due to the well-known low average areal density of fibers, the analysis device 19 determines a sub-image with a larger area in the case of currency C than in the case of currency A.

[0129] During sensor adaptation, Table 2 is determined for each banknote type to be inspected by the sensor. This table contains sub-image definitions (position and size), expected values ​​E for each sub-image, and acceptable fluctuation range S, as shown in Table 2. The respective area is obtained by multiplying the sub-image width dx by the sub-image height dy.

[0130] Sub-images 11, 12, 21, and 22 in Table 2 all have the same shape and size. See [link / reference] Figure 4c .

[0131]

[0132] Table 2 lists all denominations used for Currency C.

[0133] However, due to the explicit sub-image definition in Table 2, different sub-image sizes are also possible on the same banknote. Here, the expected value E and fluctuation range S are the same for all sub-images 11, 12, 21, and 22 because they have the same area. But the expected value and fluctuation range can also be different for these sub-images, for example, when the sub-images have different areas. A larger fluctuation range S is also allowed when the area is larger. The optical sensor 25 also stores corresponding tables for other denominations or currencies for which other values ​​are generally applicable. In the case of uneven fiber distribution, as for example for Figure 2a Like banknote 70 of currency B in the image, the corresponding table will contain different expected values ​​E for different sub-images.

[0134] The acceptable fluctuation range S can be given in the table. The acceptable fluctuation range can also be calculated based on the expected value E, for example, 40% of the expected value E, or it can be adjusted for the operator of the valuable document processing equipment.

[0135] Optical sensor 25 receives information about the type of banknote to be inspected, i.e., currency C, and, if necessary, denomination, from valuable document processing device 1 or from other sensors. Using the information about the type of banknote to be inspected, analysis device 19 obtains information from Table 2 about a pre-determined sub-image for the type of banknote to be inspected.

[0136] For analysis, the fibers are first located in the luminescent image recorded by optical sensor 25. Based on the luminescent image recorded from the banknote to be inspected, for example, the number of fibers N in each of sub-images 11, 12, 21, and 22 can be determined and compared with a corresponding expected value E+ / -S. In the example of currency C, denomination 10, the expected number of fibers in each of sub-images 11, 12, 21, and 22 is within the range of E+ / -S, i.e., 11 to 21. Figure 4b , Figure 4d In the structural forgery 65, 18 mixed-color fibers were actually found in sub-images 11 and 21, therefore these sub-images are not classified as suspicious. Conversely, fewer mixed-color fibers, 8 or 10 respectively, were found in sub-images 12 and 22, and therefore these sub-images are classified as suspicious. Since at least one sub-image is classified as suspicious (12 and 22), therefore... Figure 4b , Figure 4d The 65th category of structural counterfeit items is those suspected of being counterfeit in terms of structural counterfeit items.

[0137] As a supplement or alternative to the number of fibers in a corresponding sub-image, the properties of the fibers in that sub-image, such as fiber length, fiber aspect ratio (width-to-length ratio), fiber shape or color, or the corresponding average value of that property in the corresponding sub-image, or the dispersion of that property in the corresponding sub-image near that average value, can also be used as local fiber feature values. This allows for the identification of conspicuous sub-images whose properties, average values, or dispersion deviate from reference values ​​(exceeding an acceptable fluctuation range). Furthermore, when one or more sub-images are found to be conspicuous in this respect, the valuable document is classified as potentially forged.

[0138] Example 4

[0139] In the fourth embodiment, a banknote having fibers evenly distributed on the banknote is considered, such as... Figure 4a The currency C in the banknote is 90.

[0140] For analysis, the fibers were first located in the luminescent images recorded by the optical sensor 25, and the number N of fibers in each sub-image 11, 12, 21, 22 was determined.

[0141] Unlike the third embodiment, the expected value E of the fiber count is not stored in the sensor; that is, the expected value is not predefined but determined when the corresponding banknote is examined. Here, a relative check of the fiber count is performed on at least two different sub-images, wherein the fiber count of at least one particular sub-image is defined for the expected value E of one or more other sub-images.

[0142] For example in Figure 4c In the case of banknote 90 of currency C, the fiber count of sub-image 11 (e.g., N = 18) can be preset to an expected value E for other sub-images 12, 21, and 22, i.e., E = 18. Here, the fiber count N of sub-images 12, 21, and 22 is compared with E = 18. If the difference exceeds an acceptable fluctuation range S (in the case of S = 4, for example, the expected fiber count is in the range of 14-22), the banknote to be inspected is classified as suspected of being counterfeit. For Figure 4b , Figure 4d The following results were obtained from counterfeit item 65:

[0143] Sub-image 11: N=18 is defined as the expected value E.

[0144] Sub-image 12: 10 fibers, suspected of being forged.

[0145] Sub-image 21: 17 fibers, no suspicion of forgery.

[0146] Sub-image 22: 9 fibers, suspected of being forged.

[0147] When the fibers are uniformly distributed and the sub-image areas are equal, such as in Figure 4c , Figure 4d In the case of sub-images 12, 12, 21, and 22, the average fiber count of the other sub-images 12, 21, and 22 can also be used as the expected value for one of these sub-images: thus, for example, the average fiber count of sub-images 12, 21, and 22 can be used as the expected value for sub-image 11, which in the above example is E = 12. With a fluctuation range of S = 4, a fiber count of 8-16 is expected. In this case, the deviation in the fiber count (N = 18) in sub-image 11 is classified as suspicious.

[0148] In an additional examination, the average fiber count of sub-images 11, 12, and 21 can be used as the expected value for sub-image 22, which is E = 15 in the example above. With a fluctuation range of S = 4, the fiber count of 11-19 is expected, and the deviation of sub-image 22 (N = 9) is classified as suspicious.

[0149] Example 5

[0150] In the fifth embodiment, a banknote with luminescent fibers is considered. Two luminescent images are recorded from opposite sides of the banknote in the valuable document processing device 1, and the fibers depicted or imaged are positioned there.

[0151] consider Figure 2a Taking banknote 70 of currency B as an example, this banknote only has luminescent fibers in its left half. When examined in the valuable document processing device 1, the optical sensor 25 records a luminescent image of the front of banknote 70, see [reference needed]. Figure 5a Furthermore, the opposing optical sensor 29 simultaneously records the luminescent image of the back of the banknote 70, see [reference]. Figure 5b .

[0152] Because the emitted light is typically strongly attenuated by the substrate of the valuable document, only the light emitted from fibers located on or immediately below the surface of the valuable document can usually be detected on the corresponding surface. However, the light emitted from fibers deeper within the substrate or on the other side of the valuable document cannot be detected. Therefore, the luminous fibers on the front of the banknote can only be identified in the luminous image of the front of the banknote, but not in the luminous image of the back of the banknote, and vice versa.

[0153] Examining two opposing luminescent images enables a particularly accurate double-sided inspection of banknotes, in which local fiber feature values ​​are determined for two corresponding sub-images of opposite sides of the same banknote segment.

[0154] In the front image of banknote 70, the number N1 of luminescent fibers is determined, for example, in sub-image 11. See [reference needed]. Figure 5a In the reverse image of banknote 70, a sub-image 11' is defined, which is identical to the sub-image 11 in the front image and covers the same banknote segment as the sub-image 11 in the front image. Therefore, sub-images 11 and 11' show the front and back of the same banknote segment. For example, these two sub-images 11 and 11' are located within the white area of ​​banknote 70. The number N2 of luminescent fibers is determined for the sub-image 11' on the reverse side.

[0155] The total number of luminescent fibers formed in the two sub-images 11 and 11', in the corresponding first sub-image 11 and the corresponding second sub-image 11', is the total number of fibers (N) contained in these two sub-images. ges =N1+N2). For example, in the case of 70 genuine banknotes, the expected total number of luminescent fibers is 20 (i.e., the expected value E = 20). Then, based on the total number N... ges The banknotes examined were classified as either suspected of being counterfeit or not.

[0156] For example, in the case of a structural counterfeit, if, for instance, in the left-hand portion of banknote 70, a genuine banknote has been replaced by paper without luminescent fibers, then N is correspondingly determined in both sub-image 11 and sub-image 11'. ges The total number of luminescent fibers is 0, and thus the structural counterfeit is sorted out. The total number of fibers N can be... ges The image is compared to the expected value E = 20, and if the deviation from E is too large (exceeding the acceptable fluctuation range S), it can be classified as suspected of being forged; if the deviation is only small, it can be classified as not suspected of being forged. However, other local fiber features can also be used to examine corresponding sub-images 11, 11'.

[0157] Example 6

[0158] In the sixth embodiment, consider Figure 4b The constitutive forgery 65 in the text is... Figure 4a The banknote 90 is a counterfeit, wherein a portion 92 of the genuine banknote is combined with a counterfeit portion 63 of lower fiber density.

[0159] For analysis, the fibers were first located in the emission image recorded by the optical sensor 25. The emission image was then systematically scanned within the grid of grid point R. Figure 6a In the example, the grid points R are distributed on the banknote in a 3x9 grid.

[0160] The distance 'a' of the corresponding grid point R relative to the nearest fiber is used as the local fiber feature value of the corresponding grid point. See [link to relevant documentation]. Figure 6a The distance is compared to a reference distance, chosen based on the expected fiber density value DE. Grid points with a distance a greater than the reference distance are classified as conspicuous. The document is then classified as potentially counterfeit, for example, based on whether one or more conspicuous grid points are found in the banknote image, or, for example, based on whether one or more conspicuous areas are found, where a conspicuous area has multiple (e.g., a minimum proportion) conspicuous grid points. The document can also be classified as potentially counterfeit based on the location of conspicuous grid points or areas on the document.

[0161] exist Figure 6a In the example of the structural counterfeit 65, the two grid point columns on the right each have multiple grid points that are prominent in this respect and constitute a prominent area. Therefore, structural counterfeit 65 is classified as suspected of being counterfeit.

[0162] Example 7

[0163] The seventh embodiment is also based on the grid point R described in the sixth embodiment, see [link to embodiment]. Figure 6b .

[0164] To analyze the emission images, each located fiber is assigned a (planar) active region W, see [reference]. Figure 6b The gray-marked areas in the diagram. The area of ​​effect W is a region defined only for analytical purposes, surrounding the corresponding fiber, but it does not actually exist on the valuable document. In the example shown, a circular area of ​​effect W is defined around the corresponding fiber, with the fiber located inside (e.g., at the center) of its assigned area of ​​effect W. The size or area of ​​the area of ​​effect W is preferably determined based on the expected density value DE of the fiber. For example, the area F of the area of ​​effect W is chosen using the formula F = K / DE, where K is a numerical factor. For example, with a numerical factor K = 1, the area of ​​...

[0165] For example, the affinity (yes / no) of a corresponding grid point R with one or more influence regions W surrounding the fiber is used as a local fiber feature value. Therefore, for each grid point, it is checked whether it lies within one or more influence regions W. If yes, the corresponding grid point is classified as belonging to an influence region; otherwise, it is classified as not belonging to an influence region.

[0166] Then, one or more prominent areas of the valuable document image are searched, within which the number or proportion of raster points classified as not belonging to the area of ​​effect is greater than expected. Figure 6a In the example, in the two rightmost columns of the counterfeit portion 63 of the constitutive counterfeit 65, the proportion of grid points classified as not belonging to the effective area is 3 out of 4, equivalent to 75%. This proportion is compared, for example, to a reference proportion (e.g., 50%), and the corresponding column of the grid is classified as conspicuous if it exceeds the reference proportion. The reference proportion can be predetermined, adjustable, or determined based on the proportion of grid points not belonging to the effective area in other banknote areas. For example, the proportion of grid points not belonging to the effective area in each right column can be compared to the average proportion of grid points not belonging to the effective area determined for multiple left columns of the grid.

[0167] The valuable document can then be classified as suspected forgery based on whether one or more prominent columns are found. Therefore, constitutive forgery 65 is classified as suspected forgery due to two prominent right columns.

[0168] Example 8

[0169] The eighth embodiment is also based on the grid point R described in the sixth embodiment and the active area W used in the seventh embodiment.

[0170] Each grid point R is assigned a (area-shaped) ambient region 111-119, 121-129, 131-139, 141-149 around the corresponding grid point. This ambient region corresponds to a sub-image of the banknote image. See [link / reference]. Figure 6c Environmental regions 111-119, 121-129, 131-139, and 141-149 overlap. Grid point R is located, for example, at the center of the corresponding environmental region.

[0171] The proportion, or percentage, of the area (gray or dark gray) covered by at least one active region W or not covered by at least one active region W within the corresponding ambient area around the corresponding grid point is used as a local fiber characteristic value. For this purpose, the coverage area can be identified within the corresponding ambient area (in... Figure 6c These covered areas (appearing gray or dark gray) belong to the functional area of ​​at least one fiber (which may be fibers inside or outside the corresponding environmental area, with the functional area of ​​these fibers partially located within the environmental area). Alternatively or additionally, free areas (white) may be identified within the corresponding environmental area, which do not belong to the functional area of ​​at least one fiber. Alternatively or additionally, overlapping areas may also be identified within the corresponding environmental area, in which the functional areas of (adjacent) fibers overlap (dark gray).

[0172] The local fiber characteristic values ​​of the corresponding environmental region are determined based on the dimensions of the corresponding covered area and / or free area and / or overlapping area. For example, the following parameters can be used as local fiber characteristic values.

[0173] - The proportion (e.g., percentage) of the area covered by the (gray) coverage area in the corresponding environmental area, or

[0174] - The proportion (e.g., percentage) of the area occupied by the (white) free area within the corresponding environmental area, or

[0175] - The percentage (e.g., area) of the (dark gray) overlapping area within the corresponding environmental area, or

[0176] - The area of ​​the covered area and / or the area of ​​the free area and / or the area of ​​the overlapping area.

[0177] Subsequently, the valuable documents are classified as suspected of being forgery based on one or more of these proportions or areas, which have been determined for the environmental areas of the valuable documents being examined.

[0178] For example, the proportions or areas of corresponding environmental regions can be compared with reference values, and if they exceed the reference values, the corresponding environmental regions can be classified as conspicuous. For example, for the structural forgery 65, environmental regions 118, 119, 128, 129, 138, 139, 147, 148, and 149 are conspicuous in this respect. Subsequently, the valuable document can be classified as suspected of forgery based on whether one or more conspicuous environmental regions are found.

[0179] Alternatively, the average proportion or area of ​​all environmental regions in a valuable document image can be determined, and the dispersion of these proportions or areas around this average can be analyzed. If the dispersion is too large, the valuable document is classified or sorted as potentially forged.

Claims

1. A sensor (25) for inspecting a valuable document, the valuable document having multiple fibers distributed on the document, and the valuable document being placed in the detection area of ​​the sensor for inspection, wherein, The sensor has: Image detection device, configured to detect a valuable document image of a valuable document (3), wherein the valuable document image shows characteristic optical or magnetic signals of fibers, and Analysis device (19), which is configured for analyzing valuable document images, Locate the fibers contained in the corresponding valuable file image, and At least one local fiber feature value is determined for each of multiple different points in the valuable document image. This local fiber feature value is applicable to the corresponding point or to the corresponding point and its environment in the valuable document image. The analysis device (19) has information about the valuable document type of the valuable document to be examined, which is the currency or denomination of the valuable document to be examined. The analysis device is configured to determine the points in the valuable document image according to the valuable document type of the valuable document to be examined, and to determine the corresponding local fiber feature values ​​for these points. Based on the at least one local fiber characteristic value, the valuable document (3) is classified as suspected of being forged.

2. The sensor (25) according to claim 1, wherein, The analysis device (19) is configured to analyze the valuable document image by scanning the valuable document image point by point using grid points (R), wherein a local fiber feature value is determined for each grid point of the valuable document image, and the valuable document (3) is classified as suspected of being counterfeit based on the local fiber feature value determined for the grid point.

3. The sensor (25) according to claim 2, wherein, The analysis device (19) is configured for analyzing valuable document images. Identify one or more prominent grid points that are prominent in terms of the local fiber feature values ​​determined at the respective grid points. Identify one or more prominent areas of a valuable document image, within which prominent raster points exist, and Valuable documents are classified as suspected forgeries based on one or more characteristics of the corresponding prominent area and / or based on the local fiber feature values ​​determined for the grid points of the corresponding prominent area.

4. The sensor (25) according to claim 2 or 3, wherein, Analytical apparatus (19) To analyze valuable document images, local fiber feature values ​​are determined for the environmental regions (111-119, 141-149) belonging to the corresponding raster points. These local fiber feature values ​​are determined for the corresponding raster points in the valuable document image. To analyze valuable document images, settings are used to determine corresponding local fiber feature values ​​point by point for each grid point in the valuable document image.

5. The sensor according to claim 1, wherein, The analysis device (19) is configured for analyzing valuable document images. Identify one or more sub-images (11-14, 21-24, 111-119, 141-149) of the valuable file image, where each sub-image contains a fragment of the valuable file image. Determine local fiber feature values ​​for the corresponding sub-image. Based on the local fiber feature values ​​of one or more sub-images, the valuable document (3) is classified as suspected of being forged.

6. The sensor (25) according to claim 1, wherein, A measure of the local density of a fiber is used as a local fiber characteristic value, wherein the measure of the local density of the fiber is The number (N) of fibers in the environment at the corresponding location, or Fiber density in the environment at the corresponding location, or The distance between fibers in the environment at the corresponding location is a characteristic distance quantity.

7. The sensor according to claim 5, characterized in that, The analysis device (19) is configured to identify at least two sub-images of a valuable document image.

8. The sensor according to claim 7, characterized in that, The at least two sub-images together cover at least 50% of the area of ​​one face of the valuable document.

9. The sensor according to claim 7, characterized in that, The at least two sub-images are arranged row by row and / or column by column on the valuable document.

10. The sensor (25) according to claim 1, wherein, A measure of the local property of a fiber at a given point or of a fiber at that point and in its environment is used as a local fiber characteristic value. This local property relates to the shape, size, color, or brightness of the fiber at the given point or of a fiber at that point and in its environment. A representative value of the local property of the fiber at the given point or of a fiber at that point and in its environment is used as a local fiber characteristic value.

11. The sensor (25) according to claim 1, wherein the sensor is configured to determine unique fingerprint data of a corresponding valuable document based on one or more characteristics of fibers located in the image of the valuable document and / or based on one or more local fiber feature values ​​of the corresponding valuable document determined for one or more locations of the image of the valuable document, assign such unique fingerprint data to the processing data of the corresponding valuable document, and store such unique fingerprint data and the processing data of the corresponding valuable document in a mutually corresponding and associated manner.

12. The sensor according to claim 5, characterized in that, The analysis device (19) has information about the value file type of the value file to be examined and is configured to determine a sub-image based on the value file type of the value file to be examined. A plurality of sub-images are fixedly predefined in the analysis device, and the analysis device is configured to determine a specific sub-image from these fixedly predefined sub-images for analysis based on the value file type of the value file to be examined.

13. The sensor according to claim 1, characterized in that, The sensor is configured to inspect one or more specific types of valuable documents, each having at least one interfering anti-counterfeiting element (68) that hinders the detection of characteristic optical or magnetic signals of the fibers. The analysis device (19) of the sensor is configured to classify a particular type of valuable document as suspected of being counterfeit based on at least one local fiber feature value at one or more of the following points of the valuable document image, wherein the detection of optical or magnetic signals of the fibers at the points is not hindered by the at least one interfering anti-counterfeiting element (68).

14. The sensor (25) according to claim 1, wherein, The analysis device (19) is configured for analyzing valuable document images. Assign an effective region (W) to each of the positioned fibers. For multiple different points in a valuable file image, local analysis is performed on the fiber regions present at each point or the fiber regions present at and in the environment. Analysis is conducted regarding the areas covered or uncovered by the affected regions (W) or the overlap of adjacent affected regions. Based on the local analysis of the action area (W) located at the corresponding point or in its environment, the local fiber characteristic values ​​of each point are determined. The analysis device (19) is configured to select the size of the action area (W) according to the density expectation value (DE), which gives the fiber surface density expected for the valuable document image or for the valuable document image and its corresponding points.

15. The sensor (25) according to claim 12, wherein, The analysis device (19) is configured for analyzing valuable document images. Assign a target area to the fibers located in the valuable file image, corresponding sub-image, or corresponding environmental region, and... Identify the coverage area belonging to the functional area of ​​at least one of these fibers, and / or Identify free regions that do not belong to the functional region of at least one of these fibers, and / or Identify overlapping regions where the functional areas of at least two adjacent fibers overlap, and Identify one or more prominent areas of a valuable document image that are prominent in terms of their covered area and / or free area and / or overlapping area, and Valuable documents are classified as suspected forgeries based on one or more characteristics of the covered and / or free and / or overlapping areas in one or more conspicuous regions of the image.

16. The sensor (25) according to claim 1, wherein, The analysis device (19) is configured for analyzing valuable document images. For one or more points in a valuable file image, the local fiber feature value of the corresponding point is compared with the expected value, where the value is used for the corresponding valuable file and the expected value is used for the corresponding point in the corresponding valuable file. It is predetermined, or The analysis device (19) is configured to determine, based on local fiber feature values, a predicted value to be compared with the local fiber feature values ​​of a corresponding point, wherein the local fiber feature values ​​are determined for one or more other points of the same valuable document image, and Based on the comparison results obtained from one or more points of the image of the valuable document, the valuable document (3) is classified as suspected of being forged.

17. The sensor according to claim 1, characterized in that, The valuable document image is a first valuable document image, which is detected from a first valuable document face of the valuable document (3) to be inspected, and the analysis device has a second valuable document image, which is detected or has been detected from a second face of the same valuable document opposite to the first face, and the analysis device (19) is configured to... At least one local fiber feature value is determined for at least one first sub-image (11) of the first valuable file image and for at least one second sub-image (11') of the second valuable file image, and The local fiber feature values ​​of the first sub-image are combined or compared with the local fiber feature values ​​of the second sub-image, and Based on the local fiber feature values ​​of the first sub-image and the local fiber feature values ​​of the second sub-image, that is, based on the result of the combination or the result of the comparison, the valuable document (3) is classified as suspected of being forged. The second sub-image is a second sub-image corresponding to the first sub-image. The second sub-image is a fragment of the second valuable document image. The first sub-image and the corresponding second sub-image contain optical or magnetic signals of the front and back sides of the same valuable document segment of the valuable document (3).

18. A valuable document processing device (1), having The sensor (25) according to any one of the preceding claims, and Transport device (10) for introducing valuable documents into the detection area of ​​sensor (25).

19. A method for inspecting valuable documents (3) using sensors, wherein, The valuable document has multiple fibers distributed on it, and the following steps are performed in the method: The valuable document (3) is introduced into the detection area of ​​the sensor (25), and the valuable document image is detected by means of the sensor's image detection device, wherein the valuable document image contains characteristic optical or magnetic signals of the fibers. The valuable document image is analyzed by means of a sensor analysis device (19), which is configured for analyzing the valuable document image. Locate the fibers contained in the corresponding valuable file image, and At least one local fiber feature value is determined for each of multiple different points in the image of the valuable document. This local fiber feature value is applicable to the corresponding point or to the corresponding point and its environment. The analysis device (19) has information about the type of the valuable document to be examined, which is the currency or denomination of the valuable document. The analysis device is configured to determine the points in the image of the valuable document based on the type of the valuable document to be examined, and to determine the corresponding local fiber feature values ​​for these points. Based on the at least one local fiber characteristic value, the valuable document (3) is classified as suspected of being forged.

Citation Information

Patent Citations

  • Method and device for identifying paper money

    CN102890840A

  • Method and device for examining a security feature of a value document

    CN104094321A

  • Banknote management system and method and banknote recognition method

    CN107204068A

  • Device and method for classifying value documents, in particular bank notes, and value document processing system

    CN109906473A

  • Object to be detected for checking its genuineness processor for checking its genuineness

    JP1999265421A