JP2004153405A - Method and apparatus for confirming document - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method and apparatus for confirming a document capable of enhancing the falsification preventing effect of documents and being easily realized. <P>SOLUTION: A read section 14 reads a print image of a regular printed matter (original) and a feature quantity extract section 84 extracts a feature of an unreproducible random pattern uncontrollably formed when a printer prints out the printed matter, obtains a feature vector and registers it in a memory 86 in advance. Similarly the read section 14 reads a print image even from a printed matter 12 being a confirmation object which is confirmed as to whether or not the printed matter 12 is an original or not, and the feature quantity extract section 84 extracts the feature of an unreproducible random pattern and calculates a feature vector. Then a comparison section 88 compares a calculated calculation feature quantity vector with a registered feature quantity vector registered in the memory 86 and discriminates whether or not the printed matter of the confirmation object is the original depending on the similarity between the both. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a document confirmation method and apparatus, and more particularly to a document confirmation method and apparatus for confirming the originality of a document on which an image is recorded.
[0002]
[Prior art]
It is necessary to take special measures to prevent counterfeiting of paper documents such as securities, various rights certificates, insurance policies, resident cards, birth certificates, guarantees, passports, banknotes, confidential documents, etc. is there. As a technology for preventing forgery for this purpose, conventionally, a method using an advanced printing technique or a special ink which is difficult to obtain in order to make forgery difficult for printing the identification code is generally used together with the printing of the identification code. Alternatively, a method of attaching a forgery prevention sheet by a special technique such as holography is often used.
[0003]
In addition, as a technique for identifying a document without using an identification code, a reactant that exhibits a color developing reaction on a medium such as paper, a strip that emits a fluorescent color by irradiating ultraviolet rays, an infrared absorbing fiber or a strip, and the like in a paper making process. There is a method of forming a random foreign matter pattern (hereinafter referred to as a random pattern) on a sheet of paper by, for example, fabricating the paper (for example, see Patent Documents 1 to 4). In addition, there are a method of applying a magnetic material to a random pattern and a method of inserting a thread into a sheet (for example, see Patent Document 5). As described above, forgery can be prevented by using a special sheet on which a random pattern is formed by a technique for mixing foreign materials such as functional materials.
[0004]
In addition, a method of preventing forgery by providing a functional material that forms a random pattern in a sheet (paper) together with an identification code (see Patent Literature 6), or forgery by printing an invisible pattern with infrared absorbing ink There is also a method of preventing the occurrence (see Patent Literatures 7 and 8).
[0005]
[Patent Document 1]
JP-A-6-287985
[Patent Document 2]
JP-A-7-166498
[Patent Document 3]
JP-A-8-120598
[Patent Document 4]
JP-A-10-269333
[Patent Document 5]
JP-A-10-219597
[Patent Document 6]
JP-A-2002-83274
[Patent Document 7]
JP-A-6-210987
[Patent Document 8]
JP-A-2002-146254
[0006]
[Problems to be solved by the invention]
However, the first purpose of the above-described technology for mixing foreign materials such as special printing and functional materials is to distinguish it from ordinary paper, and it is expected to be used in a large volume of issuance such as banknotes. In other words, since the printed contents are basically the same, such as banknotes and securities, and the same securities and the same securities, the counterfeiting criminals only need to learn the technology of fabricating such functional materials once. There was a problem that mass counterfeiting became possible. Identification and collation of individual papers can be performed with visible information called a printed identification code, but in this case, the identification code is only a means for specifying a counterfeit ticket, and there is a limit in preventing forgery.
[0007]
In addition, even if the printed contents are different, such as various certificates and confidential documents, in order to secure the originality, special printing of the identification code and special paper with functional material, etc. Alternatively, it is best to use both. However, unlike large-volume printing at a specific printing station, such as banknotes or securities, in the case of small-volume printing, such as printing at a limited amount by a company or local government, the use of special printing or special paper is highly effective. There was also a problem that it became a cost.
[0008]
Furthermore, with the development of personal computers and network technology, there is a growing demand for easily outputting documents such as certificates on demand, whose originality must be guaranteed. The availability of consumables such as ink and paper must be improved. However, it is originally difficult to obtain these consumables, and there is a problem that such measures are inconsistent with these technologies which are expected to have a deterrent effect of counterfeiting. In addition, since these consumables are expensive and are not generally distributed, there is a problem that output may become impossible due to exhaustion of consumables.
[0009]
As described above, it has been difficult with conventional techniques to simultaneously achieve the security of securing originality or preventing forgery, the convenience of being easy at any time, and the low cost.
[0010]
SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a document checking method and apparatus which can improve the effect of preventing forgery of a document and can be easily realized.
[0011]
[Means for Solving the Problems]
In order to achieve the above object, an invention according to claim 1 is a document confirmation method for confirming the originality of a document on which an image is recorded. Is read, and based on the result of reading the image from the original document, the feature of the non-reproducible disorder portion included in the image is extracted, and the first feature information indicating the extracted feature is stored. Reading an image recorded on the document to be confirmed, extracting a feature of a non-reproducible disorder portion included in the image based on a result of reading the image from the document to be confirmed, and extracting the extracted feature. Comparing the second characteristic information shown with the first characteristic information stored in advance, and determining whether or not the document to be confirmed is the original based on the comparison result. .
[0012]
According to the first aspect of the present invention, an image recorded in a document serving as an original document is read in advance, and a feature of a non-reproducible disorder portion included in the read image is extracted. The non-reproducible disordered portion is, for example, a state of scattering of the image forming material or a state of uneven permeation of the document which occurs when an image is recorded on the document, imperfect shape reproduction, a low coverage of the image forming material. This is a microscopic image disturbance that cannot be controlled at the time of recording the image included in an image recorded on a recording medium such as paper, such as microscopic unevenness at the time. The document referred to here includes a printed matter on which an image is recorded by printing and a handwritten document, and may be a paper medium or a resin medium. The term “image” refers to not only a photographic image and a graphic image but also all kinds of image information, including character information, recorded on a recording medium so as to be visually recognizable.
[0013]
The feature extracted in this way, that is, the feature of the unreproducible disordered portion of the regular document (original) is registered in advance as first feature information. The first feature information may be stored in a storage device such as a memory, or may be stored in the original document itself by printing.
[0014]
When confirming whether or not the document is the original, the characteristics of the unreproducible disorder portion are extracted from the document to be confirmed in the same manner as the original, and the unreproducible disorder of the extracted document to be confirmed is extracted. The second feature information indicating the feature of the unit is compared with the first feature information. That is, by comparing the feature of the unreproducible disordered portion of the document to be confirmed with the feature of the unreproducible disordered portion of the original document registered in advance, for example, the similarity of the two allows the document to be confirmed to be updated. It is determined whether or not it is the original.
[0015]
As described above, by utilizing the characteristic of the disordered portion that cannot be reproduced, each document can be individually identified, and it is easy to confirm whether or not the document is the original. Therefore, the forgery prevention effect can be improved. Further, special printing, special paper, and the like are unnecessary as in the related art, so that the present invention can be easily realized.
[0016]
In the above-described document confirmation method, it is important to read an image from the same portion of a regular document (original document) and the document to be confirmed, that is, to align the documents. For this purpose, a reference point for aligning the read result and the document is set on the document, an observation area on the document is specified based on the reference point, and an observation area within the observation area is specified. It is good to read an image. In order to confirm the originality, the observation area does not need to be the entire image recorded in the document, but may be a part of the recorded image.
[0017]
In this case, the reference point may be, for example, at least one of an edge of the document, a corner of the document, a predetermined mark recorded on the document, a hole penetrating the document, and irregularities formed on the document. Can be used.
[0018]
When a predetermined mark is recorded on a document as a reference point by printing or the like, the image and the mark absorb light of different predetermined wavelengths from each other to facilitate identification of the observation region. If the document is recorded with the image forming material, for example, the document is irradiated with the first light having a wavelength absorbed by the image forming material on which the mark is recorded, and the transmitted or reflected light of the first light is transmitted through the document. Receiving light, identifying the observation area on the document, irradiating the document with a second light having a wavelength that is absorbed by the image forming material on which the image is recorded, and applying the second light to the document. By receiving the transmitted light or the reflected light, the image in the specified observation region can be read. Alternatively, the document is irradiated with light containing the predetermined wavelengths different from each other, and among the transmitted light or reflected light of the document, the first light having a wavelength absorbed by the image forming material on which the mark is recorded is received. Identifying the observation region on the document, receiving the second light having a wavelength absorbed by the image forming material on which the image is recorded, and reading the image in the identified observation region. . In such a case, for example, one of the image forming material of the image and the mark may include an infrared absorbing material or an ultraviolet absorbing material.
[0019]
Further, in order to improve robustness and to make it possible to identify more documents, it is preferable to provide a plurality of the observation areas in the documents.
[0020]
Further, the above-described document checking method is particularly effective when the document is a printed matter printed by an electrophotographic method.
[0021]
Note that the above document confirmation method can be easily realized by a document confirmation device configured as follows. That is, as set forth in claim 9, there is provided a document checking device for checking the originality of a document on which an image is recorded, wherein the reading means reads an image recorded on the document, A feature extracting unit that extracts a feature of a non-reproducible disorder portion included in the image based on a reading result of the image by the unit; and an image of an original document that is read by the reading unit. A storage unit for storing first feature information indicating the feature extracted by the feature extraction unit, and an image of a document to be confirmed read by the reading unit, and extracted by the feature extraction unit based on the read result The second feature information indicating the feature is compared with the first feature information stored in the storage unit, and based on the comparison result, the document to be confirmed is To characterized in that it has a discriminating means for discriminating whether the present, the.
[0022]
The document referred to here includes a printed matter on which an image is recorded by printing and a handwritten document, and may be a paper medium or a resin medium. The image refers to not only a photographic image and a graphic image but also all image information recorded on a recording medium so as to be visually recognizable, including character information.
[0023]
Further, the first feature information may be stored in a storage device such as a memory by a storage unit, or may be stored in the original document itself by printing or the like. Further, the reading means may be of a type that irradiates a document with light and observes reflected light to read an image, or may be a type that reads an image by observing transmitted light.
[0024]
In the above-described document confirmation device, the reading unit includes a positioning unit for positioning the document, and the document is positioned by the positioning unit, and the document is positioned on the document. It is preferable that an observation area is specified and an image in the observation area is read.
[0025]
Alternatively, the reading means detects a predetermined reference point on the document for performing alignment, specifies an observation area on the document based on the detected reference point, and May be read. In this case, the reference point may be, for example, at least one of an edge of the document, a corner of the document, a predetermined mark recorded on the document, a hole penetrating the document, and irregularities formed on the document. Can be used.
[0026]
Further, when the image and the mark are recorded by printing or the like with an image forming material that absorbs light of different wavelengths, the reading unit may have a wavelength that is absorbed by the image forming material on which the mark is recorded. Irradiating means for irradiating the document by switching between first light and second light having a wavelength component of the first light of substantially 0, and transmitted light of the document irradiated by the irradiating means Alternatively, a light receiving means for receiving the reflected light may be provided. Alternatively, the reading unit may include: an irradiating unit that irradiates the document with light; and a wavelength at which the transmitted or reflected light of the document irradiated by the irradiating unit is absorbed by the image forming material on which the mark is recorded. And a light receiving means for selectively receiving the first light and the second light having the wavelength component of the first light of substantially zero.
[0027]
Further, it is preferable that the reading means has a light shielding member for shielding the document from external light.
[0028]
In the above-described document checking apparatus, the document may be a printed matter printed by an electrophotographic method.
[0029]
BEST MODE FOR CARRYING OUT THE INVENTION
Next, prior to the description of the embodiment according to the present invention, a non-reproducible disordered portion (hereinafter, random pattern) included in an image recorded on a document will be described. Hereinafter, a non-reproducible random pattern of an image forming material generated in a printed matter on which an image is printed by a printing apparatus will be mainly described.
[0030]
Non-reproducible random patterns are, for example, in the case of a toner image, irregularities (jaggies) in the tangential direction of the line edges (in the case of characters, TEP-tangential edge profile), irregularities in micro density, and voids in the image area (voids) This is known as toner scattering near the image portion (blur, image dependent noise), toner scattering in the background portion (image independent noise), and the like. Note that the image forming material is a substance that imparts a color property such as a thin film on an object, and is, for example, a printing ink, a paint, a colored particle (toner) for electrophotography or ionography, a liquid toner, or an ink jet printer. Includes ink, thermal transfer printer ink (ribbon), sublimation printer ink, and the like.
[0031]
Next, the reason for the occurrence of a non-reproducible random pattern and the advantages obtained by using its features will be described.
[0032]
Ideally, the image forming material of the printed matter is given an appropriate amount to give a predetermined density to a specific position on the medium based on data or original image information. However, in reality, it is difficult to strictly control the position and amount during printing by the printing apparatus, and uncertainty such as the position and amount of the image forming material attached to the print material does not hinder the purpose of the printed material. Is tolerated. In other words, this allowable range is determined by the visual spatial frequency characteristics and the like in the general printed matter, and allows the existence of a random pattern that cannot be reproduced due to no microscopic control. Alternatively, even if the density is macroscopically uniform, such as gray reproduction, the uneven distribution of the image forming material may be microscopically permitted.
[0033]
For example, in electrophotographic technology, uncontrollable scattering and poor reproduction of toner due to electric field disturbance, repulsion between toners of the same polarity, mechanical jumping, and the like are observed in processes such as development, transfer, and fixing. There is no problem in the use of. In the inkjet technology, so-called bleeding due to uneven ink penetration is observed depending on the quality of the print medium. Hereinafter, an example of such a non-reproducible random pattern will be described in detail.
[0034]
FIGS. 1A to 1C are partially enlarged views of three copy results obtained by copying (printing) the same gray image by an electrophotographic copying machine. In digital electrophotography, halftone reproduction is performed by binarization using a dither method, a density pattern method, or the like. As can be seen from FIG. 1, even when the same gray image is printed, The shape of the pixels that form the halftone dots and the distribution of the individual toners that form each pixel cannot be controlled, and each copy result is a microscopically distinct image. Further, even in the case of a plurality of copies in analog electrophotography, even if the original image is the same, that is, even if the latent image potential is the same halftone, the individual toner positions of the development amounts that cancel the potential cannot be controlled. In other words, even if the toner coverage on the paper surface observed as the optical density is the same, it is extremely unlikely that the microscopic toner adhesion unevenness will be the same.
[0035]
2A and 2B show an example in which a halftone dot portion of an offset printed image is copied by an electrophotographic copying machine. FIG. 2A is a partially enlarged view of an original image (offset printed image), and FIGS. An enlarged view of the same portion as (A) of the result of three copies taken from the original image of (A) is shown. As can be seen from FIG. 2, it can be seen that each copy result is a microscopically different image, such as dot shape and toner scattering, even though the same original image was copied.
[0036]
3A and 3B show examples of printing using an electrophotographic printer (such as a laser printer). FIG. 3A shows a pattern of a printed image, and FIGS. It is an enlarged view of the part circled. From FIG. 3, it can be seen that although the print result is based on the same data, the shape of each square constituting the pattern and the scattering of toner are not the same, and each print result is still a microscopically different image. It turns out that it becomes.
[0037]
FIG. 4 shows an example in which the same pattern as that shown in FIG. Although the three images in FIGS. 4A to 4C are the same image, the degree of penetration of the ink into the paper is different, and the images are microscopically different. Such deviation of the ink from the original pattern also occurs in offset printing and letterpress printing. For example, in halftone printing by letterpress printing, the ink protrudes from the halftone dot diameter of the original pattern and becomes thick (dot gain), and its shape is uncertain.
[0038]
As described above, the adhesion position of the image forming material cannot be completely controlled microscopically by any printing apparatus. If this random pattern information is used, a plurality of prints or copies output from the same image data can be obtained. Each can be separated and identified. Further, in order to reproduce a microscopic random pattern of the image forming material adhering position, it is necessary to control the position of the toner particle in the order of toner particle size (several to 10 μm), and in the case of liquid ink, the ink drop size (number). Size and position control on the order of picoliters and position control on the order of the porous structure (several to 10 μm) on the medium side are required. That is, it is almost impossible to intentionally obtain the same printed matter microscopically.
[0039]
As described above, since the random pattern is a disturbance of an image formed without being controlled by the printing apparatus, it cannot be reproduced by the printing apparatus. The present invention takes this fact on the contrary and uses this non-reproducible random pattern as information (identification collation information) for determining whether or not the printed matter is the original, and confirms the originality of the printed matter. Things.
[0040]
Next, an example of an embodiment according to the present invention will be described in detail with reference to the drawings.
[0041]
[overall structure]
FIG. 5 shows a schematic configuration diagram of a printed matter confirmation device to which the present invention is applied. The printed matter confirmation device 10 shown in FIG. 5 includes, as reading means, a reading unit 14 for reading a printed image including a non-reproducible random pattern from a printed material 12 (see FIG. 6) as a document. The reading unit 14 includes a switch 16 for starting reading, and a determination unit 18 that determines whether the printed matter 12 is authentic (original / non-original) based on a result of reading by the reading unit 14. It is configured to be connected to the determination unit 18. Note that the reading unit 14, the switch 16, and the determining unit 18 may be integrated, or may be configured as physically different devices, and used by being connected via a connecting means such as a cable. Is also good.
[0042]
The switch 16 is turned on so as to start reading after the reading unit 14 inserts or brings the printed material 12 to be read into the reading unit 14 into a reading standby state. The switch 16 may be a switch operated by a fingertip by an operator, or a switch that is turned on when the printed material 12 to be read contacts the reading unit 14 or approaches a predetermined distance by various contact or non-contact sensors. . In the present embodiment, a case where an operator operates with a fingertip will be described as an example.
[0043]
Further, the printed matter 12 is obtained by printing an image on paper using an image forming material such as toner or ink, and an identification code for identifying the printed matter 12 is previously recorded at a predetermined position on a printing surface (or a back surface). Have been. With respect to the method of recording the identification code, the identification code may be printed, a bar code may be printed, or the identification code may be embedded in a print image using a digital watermark technique. Hereinafter, a case where the identification number is printed at a predetermined position of the printed matter 12 will be described as an example.
[0044]
[Detailed configuration of reading unit]
Next, the reading unit 14 will be described in detail.
[0045]
The reading unit 14 includes an illuminating unit 20 that irradiates light to the printed matter 12 to be read, and a light receiving unit 22 that receives reflected light or transmitted light of the light emitted by the illuminating unit 20 from the printed matter 12. An image printed on the printed matter 12 is read by irradiating the printed matter 12 with light by the illumination unit 20 and receiving the reflected light or transmitted light by the light receiving unit 22. Hereinafter, a specific configuration example of the reading unit 14 will be described.
[0046]
6 and 7 show an example of the reading unit 14. 6 and 7 are an external view and a sectional view of the reading unit 14, respectively.
[0047]
In the reading unit 14 shown in FIGS. 6 and 7, a housing 30 housing the illumination unit 20 and a housing 32 housing the light receiving unit 22 are supported by a support member 34 so as to be arranged vertically. The lower surface of the housing 30 is formed of a transparent material such as glass, and the other portions are formed of a light-shielding material. The housing 32 has a cylindrical shape and is formed of a material having a light-shielding property. The upper surface 32A of the housing 32 is used as a stage on which the printed matter 12 is placed, and a circular opening 32B for outputting light from the illumination unit 20 is formed in the center of the upper surface 32A.
[0048]
That is, in the reading unit 14, the printed matter 12 to be read is inserted into the gap 36 provided between the housings 30 and 32, is mounted on the upper surface 32 </ b> A (stage) of the housing 32, and is mounted. Light is emitted from the back surface of the printed matter 12 by the illumination unit 20, and the light receiving unit 22 observes the transmitted light after being scattered and absorbed by the printed matter 12 and the image forming material on the printed matter 12.
[0049]
In the gap 36, a pair of plate-shaped abutting members 38 are provided upright on the upper surface 32 </ b> A of the housing 32 so as to form an angle of approximately 90 degrees as positioning means for positioning the printed matter 12. The two sides L1 and L2 constituting one corner P of the printed matter 12 inserted into the gap 36 are abutted by the abutting member 38, so that the position of the printed matter 12 with respect to the illumination unit 20 and the light receiving unit 22 is always the same. The position is determined. The positioning of the printed matter 12 by the abutting member 38 makes it possible to easily specify an observation target area on the surface of the printed matter 12.
[0050]
In the present embodiment, as shown in FIG. 6, the printed material 12 in which the image is printed on the paper cut into a square is premised, so that the abutting member 38 is arranged as described above, and the corner of the printed material 12 is formed. The printed matter 12 can be positioned by determining the L1 and L2 positions of the two sides constituting the part P and the corner P, but the shape, the number, the arrangement position, and the like of the abutting members 38 are appropriately selected according to the shape of the printed matter 12. Needless to say.
[0051]
Further, in the present embodiment, an example is described in which the insertion of the printed material 12 into the reading unit 14 and the positioning by abutting against the abutting member 38 are manually performed by an operator. Although not shown, a transport device for transporting the printed material 12 is provided, and the transport of the printed material 12 automatically inserts the printed material 12 into the reading unit 14 and positions the printed material 12 by abutting against the abutting member 38. It may be made to be performed in a specific manner.
[0052]
As shown in FIG. 7, the illumination unit 20 includes a light source 40 that outputs light, and a light diffusion plate 42 that scatters the light output from the light source 40. As the light source 40, for example, an LED, a halogen lamp, a fluorescent lamp, a xenon discharge tube, or the like can be used. The light source 40 outputs light toward the light diffusion plate 42.
[0053]
The light diffusion plate 42 is fitted above the light source 40, more specifically, in the opening 32 </ b> B, and the output light from the light source 40 passes through the light diffusion plate 42 on the upper surface 32 </ b> A (stage) of the housing 32. The printed matter 12 is irradiated. By passing through the light diffusing plate 42 in this manner, the output light of the light source 40 can be applied to the observation area S of the printed matter 12 substantially uniformly, and there is an effect of reducing irradiation unevenness (shading). Note that, instead of the light diffusion plate 42, a condensing lens that condenses light on a predetermined observation target area of the printed matter 12 may be used.
[0054]
The light receiving unit 22 includes an image sensor 44 and a lens unit 46 that forms an image on the light receiving surface of the image sensor 44 from light transmitted through the printed matter 12. CMOS or CCD can be used for the image sensor 44.
[0055]
Here, it is considered that the individual difference of the uncontrollable random pattern such as the scattering of the image forming material to be observed is generally clearer as the observation area S is expanded because more information is obtained. However, for example, unnecessarily widening a halftone area suitable for random pattern observation as an observation target, or extracting a random pattern portion from the entire normal print information area, requires calculation load and redundancy in the determination unit 18 in the subsequent stage. Information is undesirably increased. Further, when the observation area S is widened, the measurement units such as the illumination unit 20 and the light receiving unit 22 become large, and the installation area of the reading unit 14 and the cost are disadvantageous. That is, based on the feature of interest and the cost and size of the measurement unit such as the illumination unit 20 and the light receiving unit 22, the observation target and the area of the observation area that make the individual differences of the printed matter as large as possible should be determined.
[0056]
As shown in FIG. 1 to FIG. 4, the observation target is a halftone or a fine pattern (a high spatial frequency region where the density change is steep) in which the medium (paper surface) is not uniformly covered with the image forming material. The printing section is appropriate. The area of the printing surface to be observed (for example, the area of the observation area S as indicated by a dotted line in FIG. 6) is empirically 0.1 to 3000 mm in order to identify a random pattern. ² It is desirable to be within the range. When observing an area of 6.3 mm × 5.0 mm within this range, for example, the area of the light diffusion plate 42 of the illumination unit 20 is 50 mm. ² A degree is sufficient.
[0057]
Further, for the light receiving unit 22, for example, when an observation area S of 6.3 mm × 5.0 mm is imaged over the entire effective pixel area using a CCD having a black and white, an effective pixel number of 1300 × 1030 (about 1.3 million pixels) and a square lattice CCD, In this case, the observation area per pixel is about 4.9 × 4.9 μm (6.3 mm / 1300 = 4.9 μm, 5.0 mm / 1030 = 4.9 μm). When the image forming material is a toner, its size is about 10 μm, so that the random pattern can be sufficiently read. Even in the case of ink, the pores of the paper are about the size of vegetable fibers, that is, several to several tens of micrometers, so that the permeation state can be sufficiently observed.
[0058]
At this time, if the CCD is 2/3 type and the pixel size is 6.7 μm × 6.7 μm (square lattice) (CCD effective screen size 8.7 mm × 6.9 mm), the magnification by the lens unit 46 is 1.38 ×. (Lateral magnification, 8.7 / 6.3 = 1.38, 6.9 / 5.0 = 1.38). Further, for example, when a black-and-white CCD having an effective pixel number of 640 × 480 (approximately 300,000 pixels) is used and an observation area S of 7.7 mm × 5.7 mm is imaged over the entire effective pixel area, the observation area per pixel in this case is , About 12 μm × 12 μm (7.7 mm / 640 = 12 μm, 5.7 mm / 480 = 12 μm). That is, even in this case, the approximate state of the image forming material can be observed. When the CCD is a 1/3 type and has a pixel size of 8.4 μm × 8.3 μm (square lattice) (CCD effective screen size of 5.4 mm × 4.0 mm), the magnification of the optical system is 0.7 times (horizontal magnification, (5.4 / 7.7 = 0.7, 4.0 / 5.7 = 0.7).
[0059]
In order to eliminate the influence of disturbance light, it is preferable that the entire reading unit 14 is optically closed. For this purpose, for example, as shown in FIG. What is necessary is just to comprise so that movement is possible. FIG. 8 shows a state before the printed matter is inserted, and the housing 30 is set at the upper position so that a gap 36 for inserting the printed matter 12 is formed between the housings 30 and 32. . After the printed material 12 is inserted into the gap 36 and positioned by the butting member 38, the housing 30 is lowered so that the gap 36 is eliminated, and the printed material 12 is sandwiched between the housings 30 and 32. Thus, the entry of light from the outside of the reading unit 14 can be prevented, and only the light from the illuminating unit 20 is irradiated on the printed matter 12, so that a stable image can always be obtained. In this case, the switch 16 may be turned on by detecting that the housing 30 has completely descended.
[0060]
Further, in order to avoid the influence of ambient light, the reading unit 14 may be provided with a light shielding plate 50 as shown in FIG. 9 so that external light does not enter from around the stage 32A on which the printed matter is placed. Preferably, 50 is an elastic body so as not to damage the printed matter. In order to prevent the printed matter 12 in the observation area S from waving, a pressing plate 52 may be provided, and the pressing plate 52 is desirably supported by an elastic body such as a spring.
[0061]
Specifically, in the example of the reading unit 14 illustrated in FIG. 9, the holding plate 52 is a circular acrylic plate having a diameter of 20 mm and a thickness of 1.2 mm, for example, and springs 54 are provided at three points around the holding plate 52. With the structure, the printed matter 12 can be pressed down uniformly. Further, in this example, a ring-shaped light shielding plate 50 is provided around the pressing plate 52 to shield the printed matter 12 (more specifically, the observation area S) from external light. The light shielding plate 50 is made of, for example, black silicon rubber having an outer diameter of 25 mm, an inner diameter of 22 mm, and a thickness of 1.8 mm. That is, the printed matter 12 is placed on the stage 32A, the casing 30 is lowered, the printed matter 12 is sandwiched between the stage 32A and the pressing plate 52, and the observation area S of the printed matter 12 is optically moved by the light shielding plate 50 by further lowering the casing 30. Can be closed from the surroundings. With such a structure, a random pattern that cannot always be reproduced stably can be observed. In this case, the switch 16 may be turned on by detecting that the housing 30 has completely descended.
[0062]
In the example of the reading unit 14 shown in FIGS. 6 to 9, a so-called transmitted light type reading unit 14 that irradiates light from behind the paper by the illumination unit 20 is shown, but the present invention is not limited to this. is not. For example, as shown in FIG. 10, the light source 40 may be provided on the housing 30 side, and the optical waveguide optical system 60 may be provided around the lens unit 46 to perform dark field illumination (oblique illumination) or the like.
[0063]
In addition, in the example of the reading unit 14 illustrated in FIGS. 6 to 10, an example in which a two-dimensional image sensor is used has been described, but a scanner type using a line image sensor may be used.
[0064]
For example, as illustrated in FIG. 11, the reading unit 14 includes the illuminating unit 20 and the light receiving unit 22 sandwiching a glass plate 70 on which the printed matter 12 is placed. Alternatively, it may be configured to move at a predetermined speed in a predetermined direction along the surface of the glass plate 70 by an optical system feeding mechanism (not shown). By moving the illumination unit 20 and the light receiving unit 22, a printed image can be read from the printed material 12 while scanning. In addition, instead of the movement of the illuminating unit 20 and the light receiving unit 22 by the optical system feeding mechanism, the printed material 12 may be conveyed on the glass plate 70 at a predetermined speed by the printed material feeding function.
[0065]
FIG. 11 shows a lighting unit 20 in which a fluorescent lamp is used as the light source 40 and a reflecting plate 72 is arranged in the direction opposite to the glass plate 70 with respect to the light source (fluorescent lamp) 40 instead of the light diffusing plate 42. This is an example of the configuration. With this configuration, the light output from the light source (fluorescent lamp) 40 toward the glass plate 70 is directly radiated to the printed matter 12 on the glass plate 70, and the light output from the light source (fluorescent lamp) 40 toward the glass plate 70. The light output other than the above is reflected by the reflection plate 72, is converted into light traveling in the direction of the glass plate 70, and is then applied to the printed matter 12 on the glass plate 70. Therefore, the output light of the light source (fluorescent lamp) 40 can be used efficiently, and there is an effect of reducing irradiation unevenness. In FIG. 11, the light receiving unit 22 is configured by a contact sensor 74 having a CCD as the image sensor 44 and a selfoc lens 76.
[0066]
Although FIG. 11 shows an example of the transmitted light type reading unit, a reflected light type may be used. In addition, although the example in which the optical system is a close contact type is shown, a reduction optical system may be used.
[0067]
[Detailed configuration of judgment unit]
Next, the determination unit 18 will be described in detail.
[0068]
As illustrated in FIG. 5, the determination unit 18 includes a signal processing circuit 80 that performs predetermined processing on a signal indicating a result of reading the print image by the reading unit 14, and a control circuit 82 that controls driving of the signal processing circuit 80. A feature extracting unit 84 that extracts a feature of a non-reproducible random pattern from an output signal from the signal processing circuit 80 as a feature extracting unit, and a memory 86 that stores the feature extracted by the feature extracting unit 84. And comparing the feature quantity extracted by the feature quantity extracting unit 84 with the feature quantity registered in the memory 86 as a discriminating means, and determining the authenticity (original / non-original) of the printed matter 12 based on the comparison result. A comparison unit 88 for determination and a determination result signal output unit 90 for outputting a signal indicating the result of the true / false determination are provided.
[0069]
The signal processing circuit 80 is connected to the control circuit 82, the feature amount extracting unit 84, and the light receiving unit 22 of the reading unit 14. The control circuit 82 is connected to the switch 16 and the reading unit 14. A signal indicating ON / OFF from the switch 16 is input to the determination unit 18. When the signal indicating ON is input from the switch 16, the determination unit 18 prints to the reading unit 14. A signal for instructing image reading is transmitted, and a signal for instructing measurement start is transmitted to the signal processing circuit 80.
[0070]
The signal processing circuit 80 receives the measurement start instruction from the control circuit 82 and receives a signal indicating the result of light reception by the light receiving unit 22, that is, the signal indicating the result of reading the print image including the non-reproducible random pattern by the reading unit 14. It has become. The signal processing circuit 80 performs predetermined signal processing such as amplification on the received signal, and outputs the signal to the feature amount extracting unit 84. That is, image data representing a print image read from the printed material 12 by the reading unit 14 is input to the feature amount extraction unit.
[0071]
The feature amount extracting unit 84 performs feature extraction of the print image from the input image data. A conventionally known technique can be used for feature extraction, and an example is shown below.
[0072]
The reading result of the print image by the reading unit 14 is divided (quantized) into meshes of an appropriate size (the number of meshes d = vertical M × horizontal N), and each mesh is represented by a certain density value (density level q) (sample). ) And convert it to a mosaic image. After such quantization and sampling, the density of the j-th mesh is expressed as x _j Then, this pattern becomes x = (x ₁ , X ₂ , ... x _d ) ^t (T represents transposition). This vector is called a feature vector. Each element of the vector gives the density of the corresponding image area. The obtained pattern is represented as one point on the feature space spanned by the feature vectors.
[0073]
12 to 15 show examples in which the print images (copy results and print results) shown in FIGS. 1 to 4 are quantized and sampled. FIGS. 12A to 12C are examples in which the read results of the print images of FIGS. 1A to 1C are quantized every 20 × 20 pixels and sampled into two gradations. FIGS. 13A to 13C are examples in which the read results of the print images of FIGS. 2B to 2D are quantized every 20 × 20 pixels and sampled into 255 gradations. 2D to 2F are examples in which the read results of the print images of FIGS. 2B to 2D are quantized every 20 × 20 pixels and sampled into two gradations. FIGS. 14A and 14B are examples in which the read results of the print images of FIGS. 3B and 3C are quantized every 20 × 20 pixels and sampled into two gradations. FIGS. 15A to 15C are examples in which the read results of the print images in FIGS. 4A to 4C are quantized every 20 × 20 pixels and sampled into two gradations.
[0074]
As can be seen from FIGS. 12 to 15, microscopically, different patterns are obtained for each print image, so that the feature vectors also represent unique features. That is, the feature of the random pattern that cannot be reproduced can be represented by this feature vector.
[0075]
In addition, as shown in FIGS. 1 to 4, since the image printed on the printed matter contains a non-reproducible random pattern, the image data as the read result is used as a feature of the non-reproducible random pattern. May be used. However, as in the present embodiment, the image data that is the reading result of the image printed on the printed matter is quantized and sampled as shown in FIGS. The characteristics of the non-reproducible random patterns become clearer and each can be easily identified.
[0076]
Further, the feature amount extraction unit 84 is connected to the memory 86 and the comparison unit 88. At the time of registration, the feature amount extraction unit 84 stores information representing the obtained feature vector as first feature information in the memory 86 together with the identification code of the printed matter 12. That is, the feature amount extraction unit 84 has a function of a storage unit. Also, at the time of collation or identification (hereinafter, referred to as “identification collation” when collation and identification are not distinguished), the feature amount extraction unit 84 sets information representing the obtained feature vector as second feature information. , To the comparing section 88. The present invention is not particularly limited as to a method of associating information representing a feature vector with an identification code, but may be performed, for example, using a table indicating the correspondence between feature vectors and identification codes, An identification code may be used for at least a part of the data name of the information representing the feature vector.
[0077]
The comparison unit 88 is connected to the memory 86, and can arbitrarily read the registration information in the memory 86. The comparing unit 88 compares the information representing the feature vector (referred to as “calculated feature vector”) input from the feature amount extracting unit 84 and the feature vector registered in the memory 86 (referred to as “registered feature vector”). Then, based on the similarity, it is determined whether or not the printed matter 12 is the original, that is, whether the printed matter 12 is authentic. The degree of similarity between the calculated feature vector and the registered feature vector used for the true / false determination can be obtained by calculating the distance (Euclidean distance, Mahalanobis distance, etc.) between the calculated feature vector and the registered feature vector. The closer they are, the more similar they are. The comparing unit 88 may compare the calculated feature vector with a registered feature vector having the same identification code among registered feature vectors registered in the memory 86 (collation), or may compare the registered feature vector with all registered vectors. (Identification).
[0078]
In the present embodiment, the case where the authenticity of the printed matter 12 is determined based on the distance between the calculated feature vector and the registered feature vector will be described as an example. However, the determination may be made based on the angle formed by the vectors. Further, without performing the mosaic processing as described above, it is also possible to directly collate the images obtained by the imaging device and evaluate the similarity based on a correlation value, a cumulative square error, or the like.
[0079]
In addition to performing identification and collation in the real space from the image obtained by the light receiving unit 22, for example, the obtained image is transformed into a frequency domain by two-dimensional Fourier transform, and is identified and collated in the Fourier space. May be. In this case, the image registered in advance and the image of the printed material to be inspected are synthesized in Fourier space, and a correlation intensity image is obtained by performing an inverse Fourier transform, and the similarity between the two images is evaluated from the peak value. be able to. For example, when the magnitude of the amplitude peak is equal to or larger than a predetermined threshold, it is determined that the images are the same, that is, the same printed matter.
[0080]
Further, instead of the above-described identification / collation at the image data level, identification / collation at the level of the extracted feature may be performed. For example, there is a method of calculating each center of gravity of minute points (ink) scattered in an island shape, and featuring the distance and position between these centers of gravity. In this method, a feature can be described with data smaller than the data amount generally handled at the image data level.
[0081]
Further, the comparison unit 88 is also connected to the determination result signal output unit 90. The comparing unit 88 outputs a signal indicating the result of the authenticity determination of the printed matter 12 determined by comparing the calculated feature vector and the registered feature vector to the determination result signal output unit 90. The determination result signal output unit 90 is connected to a subsequent device, and outputs a signal indicating a true / false determination result to the subsequent device in order to control the operation of the subsequent device. For example, a determination result may be displayed on a display unit such as a liquid crystal display in response to an output signal from the determination result signal output unit 90, or a predetermined process start or prohibition of a subsequent device may be controlled. You may make it.
[0082]
[Action]
Next, the operation of the present embodiment will be described. In order to determine whether the printed matter 12 is true or false, the printed matter confirmation device 10 needs to register the feature vector of the original printed matter 12 in advance. FIG. 16 shows a feature vector registration process executed in the printed matter confirmation device 10 for this purpose.
[0083]
When registering the feature vector, the operator inserts the original of the printed matter 12 to be registered into the reading unit 14, abuts against the abutting member 38, positions the original of the printed matter 12, and turns on the switch 16.
[0084]
As shown in FIG. 16, when the switch 16 is turned on, the printed matter confirmation device 10 proceeds from step 100 to step 102, acquires the identification code of the printed matter 12, and in step 104, reads the observation area S by the reading unit 14. Read the print image inside.
[0085]
Specifically, for example, the identification code may be input by an operator from input means such as a keyboard (not shown), or the reading unit 14 reads an image of an area including the identification code from the printed material 12 from the printed material 12, and reads the reading result in the OCR ( Optical Character Recognition) may be obtained. In the case where the reading unit 14 reads the image of the area including the identification code, the printed material 12 is positioned by the abutting member 38, so that it can be easily specified by the distance from the edge of the printed material 12. it can.
[0086]
To read the print image, the control circuit 82 sends a print image read instruction to the reading unit 14 to cause the reading unit 14 to read the print image in the observation area S, and to output a signal indicating the read result. The data is received by the processing circuit 80 and subjected to predetermined signal processing to obtain image data representing a print image in the observation area S. As described above, this image data includes a random pattern that cannot be reproduced during printing. Since the printed material 12 is positioned by the abutting member 38, the observation region S can be easily specified by the distance from the edge of the printed material 12.
[0087]
In the next step 106, the image data representing the print image in the observation region S is quantized and sampled in a predetermined step by the feature amount extraction unit 84, converted into a mosaic image, and the process proceeds to step 108. , A feature vector is calculated from the quantized and sampled image data.
[0088]
Finally, in step 110, data representing the calculated feature vector is stored in the memory 86 together with the identification code acquired in step 102, and the registration processing in FIG. 16 is completed.
[0089]
As a result, the feature vector (registered feature vector) of the print image in the observation region S is registered in the memory 86 as the feature amount of the printed matter 12 loaded in the reading unit 14 in association with the printed matter 12. .
[0090]
The data representing the registered feature vector is recorded on the printed material 12 itself, that is, the data representing the registered feature vector is encoded and printed on the front surface (or the back surface) of the printed material, thereby printing the printed material 12 and the registered feature vector. Can also be associated. In this case, the memory 86 can be omitted.
[0091]
Next, a case where the authenticity of the printed matter 12 is determined by collation will be described. FIG. 17 shows a collation process of the printed matter 12 performed by the printed matter confirmation device 10 for this purpose.
[0092]
When collating the printed matter 12, the operator inserts the collated printed matter 12 into the reading unit 14, abuts against the abutting member 38, positions the original printed matter 12, and turns on the switch 16.
[0093]
As shown in FIG. 17, when the switch 16 is turned on, the printed matter confirmation device 10 proceeds from step 120 to step 122 to acquire the identification code of the printed matter 12 and, at step 124, to read the inside of the observation area S by the reading unit 14. Read the printed image. Then, the image data representing the print image in the observation region S obtained as a result is quantized and sampled in a predetermined step in the next step 126, converted into a mosaic image, and in step 128 Calculate the vector. The processing in steps 120 to 128 is the same as that in the registration processing (steps 100 to 108 in FIG. 16), and thus detailed description is omitted.
[0094]
In the next step 130, the comparison unit 88 selects and reads out the registered feature vector corresponding to the identification code acquired in step 122 from among all the registered feature vectors registered in the memory 86, and Then, the calculated feature vector calculated in step 128 is compared with the read registered feature vector. As a result of this comparison, if the similarity between the two is equal to or greater than a predetermined threshold value, the process proceeds from the next step 134 to step 136, where the printed matter 12 to be compared is determined to be the “original” (real). Otherwise, the process proceeds from step 134 to step 138, where it is determined that the data is "non-original" (fake).
[0095]
More specifically, the distance between the calculated feature vector and the registered feature vector is determined, and if this distance is shorter than a predetermined threshold value, it is determined as “original”, and if the distance is longer than the threshold value, it is determined as “non-original”. become.
[0096]
The threshold value used at this time is preferably set to a predetermined allowable range in consideration of errors in the registered feature vector and the calculated feature vector (read error by the reading unit 14, quantization sampling error, and the like). In other words, the magnitude of the threshold may be appropriately selected according to a request to determine whether the authenticity is strict or weak. In addition, since the allowable range may differ depending on the type of the printed matter 12, that is, the threshold value may be different for each printed matter 12, the threshold value may be stored in the memory 86 together with the identification code when the feature vector is registered.
[0097]
In addition, since an accident such as an operation error or misalignment may occur at the time of collation, a final judgment is made based on a plurality of judgment results, or when the comparison result is not judged to be the original printed matter. The retry may be permitted up to a predetermined number of times.
[0098]
If the encoded data representing the registered feature vector is printed on the front surface (or the back surface) of the printed material 12, the reading unit 14 (or the dedicated reading unit) reads the front surface (or the back surface) of the printed material 12. It is needless to say that the data may be read and used for collation.
[0099]
Finally, at step 140, a signal indicating the determination result of “original” or “non-original” is output from the determination result signal output unit 90, and the collation processing in FIG. 17 is terminated.
[0100]
In the above matching process, the case where the registered feature vector corresponding to the identification code is selected and the calculated feature vector and the registered feature vector are compared on a one-to-one basis has been described. Alternatively, the calculated feature vector may be compared with all registered feature vectors without using the identification code. Hereinafter, this case will be described. FIG. 18 shows a process of identifying the printed material 12 performed by the printed material confirmation device 10. Since it is not necessary to associate each printed material 12 with the registered feature vector, in the registration process, the process of step 102 in FIG. 18 (obtaining the identification code of the printed material 12) can be omitted. It is needless to say that it is unnecessary to register the information indicating the feature vector only.
[0101]
When the operator identifies the printed matter 12, the operator inserts the printed matter 12 to be checked into the reading unit 14, strikes the striking member 38, positions the original of the printed matter 12, and turns on the switch 16.
[0102]
As shown in FIG. 18, in the identification process, the identification code is unnecessary, so that the processing step of acquiring the identification code of the printed matter 12 can be omitted, and when the switch 16 is turned on, the printed matter confirmation device 10 starts from step 150. Proceeding to step 152, the reading unit 14 reads the print image in the observation area S, performs quantization and sampling in the next step 154, and calculates a feature vector in step 156. The processing in steps 150 to 156 is the same as the matching processing (steps 120 and 124 to 128 in FIG. 17).
[0103]
In the next step 158, the comparing unit 88 compares the calculated feature vector with all the registered feature vectors registered in the memory 86. As a result of the comparison with all the registered feature vectors, if the value of the highest similarity having the highest similarity is equal to or more than a predetermined threshold value, the process proceeds from the next step 160 to step 162, and the printed matter 12 to be verified is “ It is determined that the original is (genuine). Otherwise, the process proceeds from step 160 to step 164, where it is determined that there is no corresponding (fake).
[0104]
That is, the distance between the calculated feature vector and each of the registered feature vectors is obtained, and basically, it is determined that the registered printed material of the registered feature vector at the closest distance to the calculated feature vector is the “original”. Even if the distance is short, if the distance is longer than the preset threshold value, it is determined that the registered printed matter is “not applicable”.
[0105]
The threshold value used at this time is preferably set with a predetermined allowable range, as in the case of the matching processing. In addition, since an accident such as an operation error or misalignment may occur at the time of collation, a final judgment is made based on a plurality of judgment results, or when the comparison result is not judged to be the original printed matter. The retry may be permitted up to a predetermined number of times.
[0106]
Finally, in step 166, a signal indicating the determination result of “original” or “not applicable” is output from the determination result signal output unit 90, and the identification processing in FIG. 18 is terminated.
[0107]
As described above, in the present embodiment, the random pattern of the non-reproducible image of the printed matter 12 is used as identification collation information, and the features of the random pattern of the non-reproducible image of the regular printed matter 12 (original) are registered in advance. The authenticity (original / non-original) of the printed matter 12 to be checked is determined by comparing the printed matter 12 to be checked with the feature of the non-reproducible random pattern in the printed matter 12 to be checked. Thereby, the characteristics of each printed matter 12 can be captured and identified / collated.
[0108]
In general, forgery or falsification of a printed matter is performed by illegally obtaining electronic data as print information and falsifying the data. The forged or falsified data is printed on a medium such as paper by a printer. By registering the characteristic of the non-reproducible random pattern of the regular printed matter 12 (original) as identification collation information as described above. Such spoofing can be easily detected. In addition, only important information such as an ID number, a name, and an amount may be physically scraped from a print medium and rewritten. By registering the feature of the pattern, it is possible to cope with such physical tampering.
[0109]
To summarize the above, the authenticity (original / non-original) of the printed matter 12 is determined by using a non-reproducible random pattern, and contaminants such as a functional material are mixed into a print medium such as paper as in the related art. No need to do.
[0110]
In addition, it is difficult to counterfeit a non-reproducible random pattern because it is out of control, and it is possible to prove the originality that the printed matter 12 or the printing information printed on the printed matter 12 is unique and authentic. It is possible. Also, since special consumables are not required, it can be realized at very low cost.
[0111]
In addition, since it is not necessary to perform special processing on the printed matter 12, if it is necessary to determine the authenticity, if an unreproducible random pattern can be observed on the printed matter 12, such a random pattern is printed later. That way, you can always adapt later. In other words, the present invention can be applied to printed materials that have already become widespread.
[0112]
By the way, when a non-reproducible random pattern is used, it is particularly important to specify which part of the printed surface to be observed at the time of identification matching, that is, which part of the printed matter to be checked, corresponds to the observation area S. It is rare to use feature data reflecting an unreproducible random pattern of the entire printed matter, and as described above, it is common to use information on a limited portion of the printed surface. Therefore, it is necessary to always specify the same location for the same printed matter at the time of registration and identification and collation. This alignment can be performed by scanning the entire printing surface and comparing it with registered information, but is not efficient. The simplest method of specifying the observation area S is a method of measuring a distance from a physical boundary such as an edge of a printed matter as described above. In the above, the abutting member 38 is provided in the reading unit 14 for this purpose. Although the positioning is performed by abutting the abutting member 12 against the abutting member 38, this method is easily affected by the deformation of the printed matter. Hereinafter, a case where the observation region S is specified by a method other than the method using the abutting member 38 will be described as the second and third embodiments.
[0113]
[Second embodiment]
In the second embodiment, a case will be described in which the position of the observation region S on the printing surface of the printed matter 12 is specified by a method of attaching a positioning mark to the printed matter 12. Note that the configuration of the printed matter confirmation device 10 may be basically the same as that of the first embodiment shown in FIG. 1, and a detailed description thereof will be omitted.
[0114]
As shown in FIG. 19, a solid black circular mark having a diameter much larger than a random pattern that cannot be reproduced, for example, a diameter of 4 mm, is printed as a positioning mark 200 on the surface (printing surface) of the printed matter 12. The position of the observation area S can be specified using the mark 200 as a clue. For example, the center of gravity of the circular mark 200 may be set as the starting point P, and the observation area S may be determined within a circle having a radius r with a point Q (point P = point Q) at a distance 0 from the center.
[0115]
Also in this case, it is desirable that the approximate position of the observation region S on the printing surface is determined by the butting member 38 and the like. In this range of the approximate position, a lens system having a wide viewing angle (angle of view) is used (prepared separately from the lens unit 46 of the light receiving unit 22), and the image sensor 44 of the light receiving unit 22 has a size of M × M. An image having a size LxL (L> M) including the circular mark 200 is read, and an MxM circular mark image corresponding to the LxL image corresponding to the MxM circular mark image is obtained in the range of the LxL image using a correlation method, a residual sequential test method, or the like. All you have to do is find the position on the image.
[0116]
Here, an example of a method of cutting out a circular mark for position detection existing in a fine pattern such as a halftone dot including many random patterns will be described. First, the average density of the area equal to or smaller than the area of the circular mark and equal to or larger than the area of the fine pattern (dot size) is determined, and the threshold density for separating the two is determined by an optimum threshold determination method or the like. Alternatively, when the fine pattern average density is known, a threshold value may be set between the fine pattern and the solid density. An image having a density equal to or higher than this threshold value is a circular mark for position detection.
[0117]
Specifically, when the halftone dot is 200 spi, the pixel size diameter is about 127 μm. For example, if the diameter of the circular mark is 4 mm, the average density may be obtained in a 300 to 500 μm square. The center of gravity = starting point P of the circular mark thus cut out (which can no longer be said to be circular at this time) may be obtained. For example, if three marks are prepared, the center of gravity (the starting point P ₁ , P ₂ , P ₃ 3) By obtaining three points, the point Q on the printing surface can be uniquely determined. In this example, the position of the starting point P may be shifted by about 150 to 250 μm every reading, but it is sufficient to specify the approximate position. The process here is also a collation process, but the amount of information is much smaller than the information of a random pattern that cannot be reproduced, so that the observation region S can be easily specified.
[0118]
Although FIG. 19 shows an example in which the positioning mark 200 is attached to the printed matter 12 by printing, the method of attaching the mark 200 to the printed matter 12 is not limited to printing. For example, the printed material 12 may be provided with a hole penetrating the printed material 12 or unevenness formed on the printing surface as the positioning mark 200. Even the mark 200 formed by such holes or irregularities can be read by the reading unit 14 (light receiving unit 22) in the same manner as the mark formed by the image forming material. As shown in FIG. 20, a protrusion 202 such as a needle may be provided as a positioning member on a stage 32A on which the print 12 is placed, and the position of the mark 200 may be mechanically detected by the protrusion 202. . FIG. 20 shows an example in which the positioning mark 200 is a hole.
[0119]
By attaching the positioning mark 200 to the printed matter 12 in this way, even if the printed matter 12 is deformed and the butting member 38 for making a rough registration is meaningless, the position of the observation area S can be easily adjusted. Can be identified.
[0120]
Note that the reading unit 14 can read the corners and edges of the printed matter 12, and perform the positioning of the printed matter 12 by the corners and edges of the printed matter 12 instead of the positioning marks 200 by printing, holes, or irregularities. Is also possible.
[0121]
[Third Embodiment]
Next, as a third embodiment, a case where the positioning mark 200 is printed using an image forming material having a different light absorption wavelength from the image forming material for printing a print image will be described. In the following, since the configuration of the printed matter confirmation device 10 may be basically the same as that of the first embodiment shown in FIG. 1, detailed description will be omitted.
[0122]
Examples of an image forming material for printing the positioning mark 200 (hereinafter, referred to as a “marking material” to distinguish it from an image forming material for printing a printed image) include, for example, a near-infrared region (wavelength A material (a toner, an ink, or the like) containing a dye or a pigment that absorbs only at a wavelength of 700 nm to 1.5 μm can be used.
[0123]
More specifically, as the marking material, for example, a transparent toner containing a near-infrared absorbing material is used, and the entire observation region S is painted out as shown in FIG. First, as the first light, infrared light is irradiated and observed with a light receiving unit 22 (a lens system having a wide viewing angle) including a light receiving element sensitive to the infrared region, as shown in FIG. It is possible to detect a portion filled with a transparent toner containing a near-infrared absorbing material, that is, an observation region S. Next, irradiation of infrared light is stopped, visible light (not including infrared light) is irradiated as the second light, and a region detected by infrared light is observed by the light receiving unit 22. As shown in C), a random pattern in the observation area S can be detected.
[0124]
In this case, it is required that the printed matter 12 can be irradiated by switching light (first light) having a wavelength in the near-infrared region separately from light (second light) for normal image reading. Is done. Aside from the light source 40, a light source that outputs light in the near-infrared region for detecting the position of the observation region S may be provided in the reading unit 14, but the light source 40 is, for example, a visible or infrared light source such as a halogen lamp. If the light source including the wavelength components is used, and the same method is used for both wavelength components, the light source 40 can be commonly used for image reading and for detecting the position of the observation region S. is there. For this purpose, a mechanism for selectively inserting the infrared transmission filter 210 and the infrared absorption filter 212 between the light source 40 such as a halogen lamp and the imaging device 44 may be provided.
[0125]
FIG. 22A shows an example in which a mechanism for selectively inserting the infrared transmission filter 210 and the infrared absorption filter 212 is provided in the illumination unit 20 (this illumination unit 20 is a twelfth embodiment). Corresponding to the irradiation means). As shown in FIG. 22A, first, if the infrared transmission filter 210 is inserted, the observation region S can be detected. Then, the infrared absorption filter 212 is inserted instead of the infrared transmission filter 210. Then, a random pattern that cannot be reproduced can be read.
[0126]
As shown by a dotted line in FIG. 22A, a mechanism for selectively inserting the infrared transmission filter 210 and the infrared absorption filter 212 may be provided on the light receiving unit 22 side (this light receiving unit 22 is (Corresponding to the light receiving means of claim 13). Further, as shown in FIG. 22B, two image sensors 44A and 44B are provided in the light receiving section 22, and the transmitted light from the printed matter 12 is separated by the separating member 220 such as a dichroic mirror into infrared light to be transmitted to the image sensors 44B and 44B. The light of other wavelengths may be guided to the image sensor 44A (the light receiving section 22 also corresponds to the light receiving means of claim 13).
[0127]
Here, the case where the observation region S is covered with the marking material has been described as an example, but a positioning mark may be printed in a region other than the observation region S with the marking material. However, covering the observation region S with the marking material also has the effect of protecting the non-reproducible random pattern of the observation region S.
[0128]
Further, here, the infrared light detects the reflected light of the printed material 12 and the visible light detects the transmitted light of the printed material 12 by the light receiving unit 22, but the illumination method is not limited to this. Also, here, the case where a marking material containing a near infrared absorbing material is used has been described, but an ultraviolet absorbing material may be used instead of the near infrared absorbing material. In this case, it is needless to say that the light source for detecting the position of the observation region S needs to include an ultraviolet wavelength.
[0129]
It is sufficient that the marking material and the image forming material have different light absorption wavelengths, and the printed image may be printed with an image forming material containing a near infrared or ultraviolet absorbing material.
[0130]
In the first to third embodiments, the description has been made on the assumption that the observation area S is one place for each printed matter 12, but the present invention is not limited to this. , A plurality of observation areas S may be provided (that is, a plurality of points Q are provided). The printed matter 12 may be damaged or deformed due to aging or storage / use conditions, and the registered observation area S may not be able to be read. In such a case, it is not possible to use a plurality of observation areas S. It is valid.
[0131]
For example, as shown in FIG. ₁ ~ Q ₅ In a specific range (in this case, a rectangular area as shown). Robustness can be improved by assigning priorities to the respective observation regions S and performing collation in order as necessary. The advantage of providing a plurality of observation areas S is not only that such robustness is improved, but also that information that can separate many printed materials can be obtained without unnecessarily increasing redundancy.
[0132]
Further, in addition to the feature amount of the non-reproducible random pattern for each observation region S, the position of each point Q on the printing surface is determined randomly for each printed material 12, and the positional relationship between the points Q is used as feature information. Is also good. In this case, for example, if the point Q ₁ Range of information including point and point Q ₂ Is included in a specific range of printed matter, ₁ Range of information including point and point Q ₂ , Even if the information matches a specific range of information including the point Q on the printed surface of both printed materials. ₁ , Q ₂ Is unlikely to match up to the relative positional relationship between the two, and the reliability can be further improved. On the contrary, when the information of a plurality of observation areas S is used, the information of each observation area S, that is, each point Q _n It can be said that the amount of information in a specific range including can be kept low.
[0133]
FIG. 24 shows another example in which information on a plurality of observation regions S is used. In FIG. 23, it is intended to mainly provide redundancy at the information measurement location so that it can cope with breakage or deformation of the paper. _n In contrast to the example shown in FIG. 24, the example shown in FIG. 24 mainly intends to use the positional relationship of the observation region S as feature information. _n These points are placed in a relatively narrow range so that the relative positional relationship of the light-receiving unit 22 can be easily observed by the light receiving unit 22. For example, as shown in FIG. 24, the inside of a halftone square region is an observation target. The entire area is observed by the image sensor 44 via the lens unit 46. After determining the approximate position of the square area by abutting the printed matter, as shown in FIG. ₁ , P ₂ , P ₃ Is determined by detecting.
[0134]
For example, the image sensor 44 is a CCD having a black-and-white 1 / 1.8 type, an effective pixel number of 2452 × 1634 (approximately 4.1 million pixels), a square lattice of 3.1 μm pixels (effective pixel area of 7.6 × 5.1 mm), and an optical system. Assuming that the magnification (lateral magnification) is 0.15, the area of the printing surface formed on the imaging surface of the CCD is about 51 mm × 34 mm. In this case, for example, P ₁ P ₂ = 45mm, P ₁ P ₃ If = 28 mm, a square can be sufficiently accommodated in this imaging plane.
[0135]
Then, vertex P ₁ , P ₂ , P ₃ The point Q is set so that it fits within this square based on ₁ , Q ₂ , Q ₃ Determine the position of These points Q _n Is determined using a random number function or the like so as to be different for each printed material. Point Q _n Vertex P _n Is stored in the memory 86 together with the identification code of the printed matter. Then, each point Q _n As described above, non-reproducible random pattern information is acquired based on the above. As shown, P on the printing surface or the imaging surface ₁ Is the origin (0,0) and the point Q ₁ (X ₁ , Y ₁ ) At the top left vertex, point Q ₁₀ (X _{1 + m} , Y _{1 + n} ) Is the first observation area S ₁ And Here, x and y may be distances or pixel numbers. Similarly, point Q ₂ (X ₂ , Y ₂ ) At the top left vertex, point Q ₂₀ (X _{2 + m '} , Y _{2 + n '} ) Is the lower right corner of the square observation area S ₂ , Point Q ₃ (X ₃ , Y ₃ ) At the top left vertex, point Q ₃₀ (X _{3 + m "} , Y _{3 + n "} ) Is the lower right corner of the square observation area S ₃ Ask for. Point Q ₁ , Q ₂ , Q ₃ These observation areas S ₁ , S ₂ , S ₃ Are stored in the memory 86 together with the preceding identification code. Thus, the point Q randomly determined for each printed material 12 ₁ , Q ₂ , Q ₃ Is located in the observation area S ₁ , S ₂ , S ₃ Can be used as the identification and collation information of the printed matter 12 together with the information of the random pattern that cannot be reproduced.
[0136]
Here, for example, Q ₁ Q ₁₀ Assuming that the area on the printing surface of the square having a diagonal of 3 mm × 3 mm, the area on the imaging surface of the CCD is 0.45 mm × 0.45 mm (3 × 0.15 = 0.45). That is, the observation area per CCD pixel is about 21 μm × 21 μm (3000 / (450 / 3.1) ≒ 20.67), and a random pattern that cannot be reproduced can be sufficiently observed.
[0137]
Thus, the observation area S ₁ , S ₂ , S ₃ Is the vertex P ₁ , P ₂ , P ₃ , A part of the entire image information obtained by the image pickup device 44 (here, CCD) is used, and the calculation load can be reduced. On the other hand, for those who attempt to forge the printed matter 12, even if they have the technology to forge a non-reproducible random pattern, the starting point P _n Is very large compared to the size of the image forming material scattered at random, and its execution is extremely difficult. Needless to say, the point P for position detection is used. _n And number of observations, observation area S _n Are not limited to those shown here.
[0138]
In the above description, as shown in FIGS. 1 to 4, a description has been made mainly of a non-reproducible random pattern made of an image forming material such as toner or ink. Rather, microscopically non-reproducible random patterns are observed even with light-modulating materials such as thermal printer paper and electrostatic recording paper. In this case, the present invention is applicable. However, it is known that a non-reproducible random pattern appears more conspicuously on a printed image when an image is printed using toner as an image forming material. As described above, in printed matter recorded using charged particle powder such as electrophotography, even when macroscopic image quality is homogeneous, when a large number of charged particles are collected in the image forming process Since uncontrollable scattering of toner occurs due to the disturbance of the electrostatic and mechanical particles, microscopically, various toner adhesion patterns unique to each printed matter are formed. Further, since the toner particles have a larger particle size than the pigment particles in the liquid ink, features can be easily extracted even with a low resolution such as a particle distribution and a toner shape. Therefore, the present invention is particularly effective for electrophotographic printed matter.
[0139]
Further, in the above description, the non-reproducible random pattern observed in the printed matter 12 in which an image is printed on paper has been described. However, even in a medium other than paper, if a similar non-reproducible random pattern can be confirmed, The present invention is applicable.
[0140]
For example, in recent years, with the spread of ID cards and high-performance smart cards (IC cards), interest in the security level has increased, and even when printing on a resin medium used for such a card, reproduction has been difficult. Since an impossible random pattern can be observed, the present invention is applicable. FIG. 25 shows an example of application to the ID card 230. Here, a non-reproducible random pattern of a picture portion such as a character or a face printed on a card base material (resin) by, for example, electrophotography, mainly a so-called toner blur (toner splatter) is used. . For example, as shown in FIG. 25, a toner blur of a part of a character or a face photograph (here, an eye part) can be used as identification collation information. The identification and collation processing may be performed as described above. As described above, the present invention can detect a non-reproducible random pattern to be registered at the time of printing information, even if the same material as the genuine material is used for the card base material and the printed information actually exists. If not, it can be seen that it is forged or falsified.
[0141]
Also, in the above description, a printed matter printed by a printing device such as a copying machine or a printer was described. In such a case, the characteristics of the random pattern are different between the regular handwritten document and the copy result, so the present invention can be applied to the handwritten document.
[0142]
【The invention's effect】
As described above, the present invention can improve the anti-counterfeiting effect by performing identification and collation of a document by using a non-reproducible random pattern included in an image of the document, and can perform special printing and special printing. Since paper is not required, it has an excellent effect that it can be easily realized.
[Brief description of the drawings]
FIGS. 1A to 1C are partially enlarged views showing the result of copying the same gray image by an electrophotographic copying machine. FIG.
2A is a partially enlarged view of an original image subjected to offset printing, and FIGS. 2B to 2D are partially enlarged views of a result obtained by copying the original image of FIG.
FIG. 3A is a partially enlarged view showing a result of printing the printed image pattern by using an electrophotographic printer, and FIGS. 3B and 3C are images showing the image pattern of FIG.
FIGS. 4A to 4C are partially enlarged views of the result of printing the image pattern of FIG. 3A by an ink jet printer.
FIG. 5 is a block diagram illustrating a schematic configuration of a printed matter checking device according to the first embodiment.
FIG. 6 is a perspective view showing an example of a reading device according to the first embodiment.
FIG. 7 is a sectional view of the reading device of FIG. 6;
FIG. 8 is a perspective view showing another example of the reading device according to the first embodiment.
FIG. 9 is a cross-sectional view illustrating another example of the reading device according to the first embodiment.
FIG. 10 is a cross-sectional view showing another example of the reading device according to the first embodiment.
FIG. 11 is a cross-sectional view illustrating another example of the reading device according to the first embodiment.
FIGS. 12A to 12C are examples of quantized samples of the read results of the print images of FIGS. 1A to 1C.
13 (A) to 13 (C) and (D) to (F) are examples of quantized samples of the read results of the printed images of FIGS. 2 (B) to 2 (D).
FIGS. 14A and 14B are examples of quantized samples of the read results of the print images of FIGS. 3B and 3C.
FIGS. 15A to 15C are examples of quantized samples of the read results of the print images of FIGS. 4A to 4C.
FIG. 16 is a flowchart illustrating a registration process executed by the printed matter confirmation device.
FIG. 17 is a flowchart illustrating a collation process performed by the printed matter confirmation device.
FIG. 18 is a flowchart illustrating an identification process performed by the printed matter confirmation device.
FIG. 19 is a diagram illustrating an example of a printed material provided with a positioning mark according to the second embodiment.
FIG. 20 is a perspective view illustrating an example of a reading device according to a second embodiment.
FIGS. 21A to 21C are diagrams for explaining an example of a printed matter provided with a positioning mark and a method of specifying an observation area according to the third embodiment.
FIGS. 22A and 22B are cross-sectional views illustrating an example of a reading device according to a third embodiment.
FIG. 23 is an example in which a plurality of observation areas are provided in a printed matter.
FIG. 24 is another example in which a plurality of observation areas are provided in a printed matter.
FIG. 25 is an example of an ID card.
[Explanation of symbols]
10. Printed matter confirmation device
12 Printed matter
14 Reading unit
16 switches
18 Judgment unit
20 Lighting unit
22 Receiver
38 Butt member
40 light source
44 Image sensor
46 Lens unit
50 Shade plate
80 signal processing circuit
82 Control circuit
84 Feature Extraction Unit
86 memory
88 Comparison section
90 Judgment result signal output section
200 mark
202 protrusion
210 Infrared transmission filter
212 infrared absorption filter

Claims (17)

A document confirmation method for confirming the originality of a document on which an image is recorded,
Read the image recorded in the original document in advance,
Based on the result of reading the image from the original document, the feature of the non-reproducible disturbance included in the image is extracted,
First feature information indicating the extracted features is stored,
Read the image recorded on the document to be checked,
Based on the result of reading the image from the document to be checked, extract the characteristic of the non-reproducible disorder included in the image,
Comparing the second feature information indicating the extracted feature with the first feature information stored in advance,
Based on the comparison result, determine whether the document to be confirmed is the original document,
Document confirmation method characterized by the following. In the document, a reference point for positioning the document is determined in advance,
Based on the reference point, specify the observation area on the document, read the image in the observation area,
2. The document confirmation method according to claim 1, wherein: The reference point is at least one of an edge of the document, a corner of the document, a predetermined mark recorded on the document, a hole penetrating the document, and irregularities formed on the document.
3. The document confirmation method according to claim 2, wherein: The image and the mark are recorded by an image forming material that absorbs light having a predetermined wavelength different from each other,
Irradiating the document with a first light having a wavelength absorbed by the image forming material having the mark recorded thereon,
Receiving transmitted light or reflected light of the document of the first light and identifying the observation region on the document;
Irradiating the document with a second light having a wavelength that is absorbed by the image forming material on which the image is recorded;
Receiving the transmitted light or reflected light of the document of the second light, and reads the image in the specified observation region,
The document confirmation method according to claim 3, wherein: The image and the mark are recorded by an image forming material that absorbs light having a predetermined wavelength different from each other,
Irradiating the document with light containing the different predetermined wavelengths,
Of the transmitted light or reflected light of the document, the first light having a wavelength absorbed by the image forming material on which the mark is recorded is received, and the observation region on the document is specified.
Receiving a second light having a wavelength absorbed by the image forming material on which the image is recorded, and reading the image in the specified observation region,
The document confirmation method according to claim 3, wherein: Either the image forming material of the image and the mark includes an infrared absorbing material or an ultraviolet absorbing material,
The document confirmation method according to claim 3 or 4, wherein: The document is provided with a plurality of the observation areas,
The document confirmation method according to any one of claims 1 to 6, wherein: The document is a printed matter printed by an electrophotographic method,
The document confirmation method according to any one of claims 1 to 7, wherein: A document confirmation device for confirming the originality of a document on which an image is recorded,
Reading means for reading an image recorded on a document;
A feature extracting unit configured to extract a feature of a non-reproducible disorder portion included in the image based on a result of reading the image by the reading unit;
Storage means for reading an image of an original document by the reading means and storing first feature information indicating the features extracted by the feature extraction means based on the read result;
An image of a document to be checked is read by the reading unit, second feature information indicating the feature extracted by the feature extracting unit based on the read result, and the first feature stored by the storage unit. Determining means for comparing the information with the information and determining whether or not the document to be confirmed is the original based on the comparison result;
A document confirmation device comprising: The reading means includes a positioning means for positioning the document,
Positioning the document by the positioning means, specify an observation area on the document, read the image in the specified observation area,
The document confirmation device according to claim 9, wherein: The reading unit detects a reference point that is predetermined on the document for performing alignment, specifies the observation area on the document based on the detected reference point, and specifies the specified observation. Read the image in the area,
The document confirmation device according to claim 10, wherein: The reference point is at least one of an edge of the document, a corner of the document, a predetermined mark recorded on the document, a hole penetrating the document, and irregularities formed on the document.
The document confirmation device according to claim 11, wherein: The image and the mark are recorded by an image forming material that absorbs light of a predetermined wavelength different from each other,
The reading unit;
Irradiating means for irradiating the document by switching between a first light having a wavelength absorbed by the image forming material on which the mark is recorded and a second light having a wavelength component of the first light of substantially 0;
Light receiving means for receiving transmitted light or reflected light of the document of the light irradiated by the irradiation means,
The document confirmation device according to claim 12, comprising: The image and the mark are recorded by an image forming material that absorbs light of different wavelengths from each other,
The reading unit;
Irradiating means for irradiating the document with light,
The transmitted light or reflected light of the document irradiated by the irradiating unit is converted into a first light having a wavelength absorbed by the image forming material on which the mark is recorded and a wavelength component of the first light substantially zero. Light receiving means for selectively receiving light with the second light of
The document confirmation device according to claim 12, comprising: Either the image forming material of the image and the mark includes an infrared absorbing material or an ultraviolet absorbing material,
The document confirmation device according to claim 12, wherein The reading means has a light shielding member for shielding the document from external light,
The document confirmation device according to any one of claims 9 to 15, wherein: The document is a printed matter printed by an electrophotographic method,
The document confirmation device according to any one of claims 9 to 16, wherein:

Images