JP2003244434A - Document authenticity discriminating apparatus and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a document authenticity discriminating apparatus and a method therefor in which authenticity can be discriminated with high accuracy at the time of discriminating authenticity of a presented document by acquiring picture images making use of a wavelength characteristic with respect to transmissivity or reflectance of light, and subjecting the acquired picture images to numeric conversion to find feature quantities. <P>SOLUTION: The feature quantities subjecting the wavelength characteristic in respect to transmissivity or reflectance of light to the document are extracted on the basis of the picture images of the document picked up by light sources having a plurality of wavelengths while utilizing the wavelength characteristic, and authenticity of the document is discriminated from these feature quantities. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document authenticity discriminating apparatus for collating and confirming documents such as ID cards and licenses, and more particularly to a document authenticity discriminating apparatus which accurately reads genuineness information of a document to discriminate authenticity. The present invention relates to a document authenticity discriminating apparatus and a document authenticity discriminating method having discrimination performance.

[0002]

2. Description of the Related Art Generally, this type of document authenticity discriminating apparatus is
For example, it is incorporated into an automatic contract machine or the like, and optically checks the content of documents such as an identification card presented by the user to verify the content.

In this case, when determining the authenticity of a document, the document is photographed with visible light, and the operator looks at the photographed image data to determine whether it is a genuine document or a forged document.

However, in recent years, such documents are easily forged by using color copiers, printers, etc., which have been improved in performance with the technical development of OA equipment, and thus are classified as legitimate documents and forged documents. As they are indistinguishable, it is easy to induce fraud because they can be replicated in an elaborate manner.

In order to prevent such fraud, it is desired to develop a document authenticity discriminating apparatus having a highly accurate discrimination function.

Therefore, as an example, the characteristic data is extracted from the grayscale image of the entire screen read for each of a plurality of light sources, and the frequency distribution obtained from the characteristic data is compared with the reference frequency distribution to determine the paper A paper sheet recognition device that recognizes a type and authenticity is known (for example, refer to Patent Document 1).

However, in this case, since the paper sheets are recognized from the overall image data, the paper sheets are easily affected by the variations in the devices and the paper sheets, and the reliability of the data accuracy is low. In addition, it is necessary to provide a large amount of data storage because it requires the entire collation determination data for collation determination of the paper sheets.

Further, there is known a document collating apparatus for photographing a document by using plural kinds of light rays of plural kinds of light sources having different wavelengths and comparing the photograph with a sample image to verify the genuineness (for example, See Patent Document 2.).

In this case, it was necessary to collect a large number of counterfeit samples and analyze them. For this reason, it is necessary to provide a large amount of data storage for the sample image.

[0010]

[Patent Document 1] Japanese Patent Application Laid-Open No. 06-44437,

[Patent Document 2] Japanese Patent Laid-Open No. 2001-76203.

[0011]

SUMMARY OF THE INVENTION Therefore, according to the present invention, an image is acquired by utilizing the wavelength characteristic relating to the transmittance and the reflectance of light when judging the authenticity of a presented document, and the acquired image is digitized. By calculating the feature amount of the document,
An object of the present invention is to provide a document authenticity discriminating apparatus and a document authenticity discriminating method capable of highly accurately discriminating authenticity from a genuine article.

[0012]

SUMMARY OF THE INVENTION The present invention is a document authenticity discriminating apparatus for discriminating the authenticity of a document, wherein plural kinds of light sources having different wavelengths and light from the plural kinds of light sources are transmitted through the document. An image pickup unit that picks up an image of the document by using at least one of transmitted light or reflected light that is reflected by irradiating the document and a predetermined image of the document to be discriminated from the image picked up by the light source by the image pickup unit. A calculating means for calculating a characteristic amount of a plurality of parts; a storing means for storing a characteristic function for authenticity extraction obtained from a reference characteristic amount of a reference document calculated by the calculating means; The present invention is characterized by further comprising a genuineness degree extracting means for extracting the genuineness degree of the discriminated document from the characteristic amount of the discriminated document calculated by using the characteristic function stored in advance in the storage means.

Here, the document is an identification document, a driver's license, or a document showing other specific information.

The image pickup means may be composed of an optical image pickup device such as a CCD camera for obtaining an image.

The calculating means and the genuineness degree extracting means are
It can be obtained by calculation by a program incorporated in the RAM or the like of the control circuit unit, and can be configured by a calculation unit of a control system provided in the device.

The storage means can be composed of a hard disk provided in the apparatus or a storage device which replaces the hard disk, or a storage section such as a RAM provided in the control circuit section.

As the light source, a plurality of monochromatic light rays having different wavelengths such as visible light rays, infrared rays and ultraviolet rays are used as the light source.

The transmitted light when the light from the light source passes through the document or the reflected light when the light from the light source illuminates the document and is reflected is imaged as image data. For example, a genuine certificate printed by a dedicated printing machine and a counterfeit certificate duplicated by an OA device have different characteristics of ink and writing, and thus have different light transmittance and reflectance. Since the generated difference depends on the wavelength of light, there is a characteristic portion where the difference becomes considerably large depending on the wavelength.

Therefore, the characteristics relating to the transmittance and the reflectance can be obtained by using the image data of such a portion,
What is obtained and digitized is defined as a feature amount. With respect to the digitized feature quantity, the correspondence between the probability that the real thing can have the feature quantity and the probability can be expressed as a feature function, and the probability at this time is defined as the genuineness.

The characteristic function is data for calculating the genuineness, and data of the genuine reference characteristic amount is analyzed and stored in advance. When setting this genuineness in advance, obtain the respective authenticity from the characteristic function for extracting the authenticity obtained from the reference feature amount for each of a plurality of predesignated parts of the genuine document, and calculate the numerical value obtained by multiplying them. , Can be the final authenticity of the document.

The genuineness obtained by digitizing in this way is "1 in the real" as the multiplication is performed in the extraction process.
Since it is close to "0 in a fake", it is possible to accurately discriminate the truth from these numerical results. Especially, the more accurate the feature quantity of many parts is set, the more accurate the truth discrimination data becomes,
A highly reliable determination result can be obtained.

In particular, when the characteristic function is obtained, it is obtained only from the data of the genuine document, and it is different from the existing discrimination method that requires the collection and analysis of a large number of fake samples, so that it is easy. A high discrimination processing function can be built.

Further, a statistical method is used in which a characteristic amount is obtained from many real picked-up images, a characteristic function is determined from the distribution of the genuine characteristic amounts, and an authenticity is obtained from a plurality of thus determined characteristic functions. Since it is possible to realize the required arithmetic processing, it is possible to accurately discriminate the authenticity of the document, since the influence of the variation of the device and the document is small.

Further, in this case, the data amount of the feature function to be stored is much smaller than the data storage amount of the sample image used in the existing discrimination method, so that the storage capacity can be small.

If the number of types of light sources is increased, the discrimination performance can be further improved. In this case, it is only necessary to obtain a new feature amount from the image acquired using the added light source and set the feature function. Since the feature functions are independent, the previous feature functions can be used effectively as they are. it can. For this reason, it is easy to expand, for example, it can be used simply by multiplying the numerical values.

Therefore, it is not necessary to reset the method of obtaining the characteristic amount and the reference data. In addition to this, independent feature functions can be combined in different patterns, and various discrimination criteria can be provided.

The above-mentioned characteristic amount is captured from the image picked up by the image pickup means. For example, in the case of a license, it is limited to the photograph part or a specific part of the character, and the distribution of the average brightness value thereof. Is obtained from a function,
Since it is not obtained from the luminance distribution of the entire display surface of the license, the above-mentioned effect is produced.

[0028]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of the present invention will be described in detail below with reference to the drawings. The drawing shows a document authenticity discriminating apparatus for discriminating the authenticity of a document such as a driver's license. As shown in FIG. 1, the document authenticity discriminating apparatus 11 is provided on an upper surface of a device body 12 provided in a rectangular parallelepiped. A mounting surface 13 for documents such as a driver's license (hereinafter referred to as a license) is provided, and a single-opening-type opening / closing lid 14 is arranged on the upper surface so as to be opened and closed.

The mounting surface 13 has an image acquisition function of acquiring image data for authenticity determination of a license by optically detecting display data using transmitted light and reflected light. .

In order to acquire the image, a flat glass surface is provided, and the surface of the license is placed on the upper surface of the glass with the surface facing downward, and that portion is set at the image acquisition position.

FIG. 2 shows the internal structure of the document authenticity discrimination device 11 having the above-mentioned image acquisition function. As a light source used for the document authenticity discrimination device 11, a transmission light source 15 disposed on the lower surface side of the opening / closing lid 14 and a reflection light source 16 disposed inside the device body 12 are used.

Then, the light from both the light sources is placed on the mounting surface 13.
To partially acquire the image of the license 17 as shown in FIG. 3 which is set flat on the mounting surface 13.

As the light source 15 for transmission, two kinds of light emitting elements 15a and 15b having different wavelengths, that is, a visible light emitting element 15a and an infrared light emitting element 15b are used.

When transmitting, the transmitted light 1 is transmitted from the rear surface side of the license 17 mounted on the mounting surface 13 toward the front surface side.
5c is made to penetrate from the opening / closing lid 14 right below.

At this time, the transmitted light 15c that has passed directly under the license 17 hits the reflecting mirror 18 that is disposed at a position below it, is refracted at a substantially right angle from the vertical direction to the horizontal direction, and is substantially horizontal. CCD that turns sideways and faces sideways
It is led to the camera 19. Here, the transparent image that has passed through the license 17 is acquired.

On the other hand, the reflection light source 16 similarly uses two kinds of light emitting elements 16a and 16b having different wavelengths, that is, a visible light emitting element 16a and an infrared light emitting element 16b. Then, at the time of reflection, the reflected light 16c is emitted from obliquely downward toward the front side of the license 17 placed on the placing surface 13.

At this time, the irradiation light emitted from the license 17 from diagonally downward to diagonally upward is incident on the downward-facing display surface of the license 17 and is reflected almost directly below. The reflected light 16c impinges on a reflecting mirror 18 which is disposed at a lower position, is refracted from a vertical direction to a horizontal direction, changes its direction in a substantially horizontal direction, and is guided to a CCD camera 19 facing sideways. Here, the reflection image reflected from the license 17 is acquired.

The image signal acquired by the CCD camera 19 is A
A / D (analog / digital) conversion unit 20 converts the data into digital data, and the image data is stored in the storage unit 21. After this image data is arithmetically processed by the arithmetic unit 22,
The calculation result is displayed on the display unit 23. Further, a light source control unit 24 is connected to the arithmetic unit 22 so that the light sources 15, 1
The image acquisition control and the light amount control are performed for each of the six types.

Document authenticity determination device 1 configured as described above
The image capturing operation of No. 1 will be described with reference to the flowchart of FIG.

First, the opening / closing lid 14 of the document authenticity discriminating apparatus 11
Open and place the photo surface of the license 17 on the mounting surface 13
Set in place with the side facing down, then open and close lid 1
4 is closed and the setting by the user is completed. When the image acquisition signal is input after this setting is completed, the light source control unit 2
Reference numeral 4 indicates that the CCD camera 1 uses the reflected light 16a reflected from the license 17 by turning on only the reflection light source 16.
Capture from 9 (steps n1 to n2).

Subsequent to the capture of the reflected image, the light source control section 24 turns on only the transmissive light source 15, and the mounting surface 13 is turned on.
Make the license 17 transparent from above. The transmitted image by the transmitted light 15a is captured by the CCD camera 19 (step n).
3 to n4).

After the image is captured, the computing unit 22 determines whether the document is a genuine document or a forged document by making a true / false decision. The obtained result is displayed and output on the display unit 23 (step n
5).

In the above authenticity determination, the transmitted light when the light from the light source passes through the license 17 or the reflected light when the light from the light source is reflected from the license 17 is acquired as image data. It is possible to obtain the feature amount for authenticity discrimination.

That is, since the genuine license 17 printed by the dedicated printing machine and the forged license copied by the OA device have different ink and writing characteristics, the light transmittance and the reflectance are different. There is a difference. Further, since this difference depends on the wavelength of light, there is a characteristic portion where the difference becomes considerably large depending on the wavelength.

Therefore, the characteristics relating to the transmittance and the reflectance can be obtained by using the image data of such a portion,
What is obtained and digitized is defined as a feature amount. Further, regarding the quantified feature quantity, the correspondence between the probability that the real thing may have the feature quantity and the probability can be expressed as a feature function, and the probability at this time is defined as the genuineness.

The characteristic function is data for calculating the genuineness, and the data of the genuine reference characteristic amount is analyzed and stored in advance. When presetting the genuineness degree, the genuineness degree is individually obtained from the characteristic function obtained from the reference feature amount for each of a plurality of predesignated parts of the real license 17 (photograph portion, character portion, boundary portion, etc.). . The value obtained by multiplying them all can be the final genuineness of the license 17.

The actual degree obtained by digitizing in this way is "1 in the real article" as the numerical values are multiplied in the extraction process.
Since it is close to "0 in a fake", it is possible to accurately discriminate the truth from these numerical results. In particular, as the reference feature amount of many parts is set, the accuracy of the true / false determination data becomes higher, and a highly reliable determination result can be obtained.

In particular, when the characteristic function is obtained, it is obtained only from the data of the real license 17, and since a large number of fake samples and their analysis are not required, the authenticity discrimination function can be easily constructed.

Next, a description will be given of the arithmetic processing contents when the authenticity of the license 17 is determined. The data which is the basis of this authenticity discrimination processing is the reflection image and the transmission image acquired by the CCD camera 19, and the feature amount of the portion where the genuine article and the counterfeit article are different is extracted from these.

For example, when infrared rays are used as the light source 16 for reflection, a density difference may occur in image data as compared with visible light. Therefore, the average luminance value of the photographic portion can be selected as the feature amount. Then, the relationship between the feature amount and the genuineness is obtained in advance as a feature function.

FIG. 5 shows an example of a shape represented as a typical feature function. This feature function can be experimentally obtained by acquiring the data of the genuine reference feature amount.

For example, when the average value of the brightness of the photograph portion in the reflection image acquired by using infrared rays is used as the characteristic amount, the characteristic function is, as shown in FIG. The shape will increase.

Further, when the average luminance value of a certain area in the transmission image is used as the feature amount, the feature function can obtain a highly accurate genuineness as shown in FIG. 5B even if the feature amount is not high. It becomes the shape.

Further, when the average value of the brightness of the character portion in the reflection image by infrared rays is used as the characteristic amount, the characteristic function has a high accuracy as shown in FIG. 5C, even if the characteristic amount is low. The resulting shape. A method of calculating the genuineness degree from such a characteristic function will be described below. The authenticity determination processing operation by calculating the genuineness at that time will be described with reference to the flowchart of FIG.

The calculation unit 22 sets a1, a2, ..., Am as the feature quantity in the reflection image picked up by the reflected light of a plurality of light sources having different wavelengths such as infrared rays and visible rays, and corresponds to each feature quantity at that time. Let the characteristic functions be f1, f2, ..., fm
(Step n11).

Further, the characteristic quantities in the transmitted image taken by the transmitted light of a plurality of light sources having different wavelengths such as infrared rays and visible rays are designated as b1, b2, ..., Bn, and the characteristic function corresponding to each characteristic quantity at that time is defined. If g1, g2, ..., Gn (step n12), the realness for each feature amount of the reflection image is
f1 (a1), f2 (a2), ..., fm (am) can be represented, and the authenticity for each feature amount of the transmission image is g1 (b
1), g2 (b2), ..., gn (bn)
(Step n13).

As a result, the genuineness of the final license Gr
Can be defined as the product of all of these (step n14).

[0058]

[Equation 1] The genuineness of the license may be determined by the genuineness of monochromatic light only, or by the authenticity of a plurality of types of light sources having different wavelengths.

Since this genuineness value Gr has a value close to 1 for a genuine product and close to 0 for a fake product, a certain threshold value is set (step n).
15) Then, the authenticity is judged by the set threshold value.
At this time, when the genuineness degree is larger than the threshold value, it can be determined as a genuine article (step n16), and when not, it can be determined as a fake article (step n17).

Alternatively, the genuineness may be displayed as a result without setting such a threshold, and the operator may judge the authenticity based on this value.

In the method for obtaining the genuineness degree by the above [Formula 1], since the genuineness degree for each feature amount is completely independent of each other and each is equal, the feature amount is increased to improve the discrimination performance. In order to achieve the above, since it can be obtained by newly defining only the feature function for the feature quantity and newly multiplying it, there is an advantage of excellent expandability.

By such a method, the license is copied by a copying machine, the licensed image captured by the scanner is output and printed out, or the actual license is pasted with characters and photographs. Can be discriminated as a fake.

As described above, the feature amount is obtained from the real captured image, and the feature function is determined from the distribution of the real feature amount. Since it is possible to implement arithmetic processing using a statistical method of obtaining the genuineness from a plurality of feature functions defined in this way, the influence of variations in equipment and license is small,
The authenticity of the license can be accurately determined.

In the correspondence between the present invention and the configuration of the above-described one embodiment, the document of the present invention corresponds to the license 17 of the embodiment, and hereinafter, similarly, the light source is the transmission light source 15.
And the light source 16 for reflection, the photographing means corresponds to the CCD camera 19, the calculation means and the genuineness degree extraction means correspond to the calculation section 22, the storage means corresponds to the storage section 21,
The output unit corresponds to the display unit 23, but the present invention can be applied based on the technical idea shown in the claims, and is not limited to the configuration of the above-described one embodiment.

According to the present invention, since a characteristic function capable of highly accurately determining authenticity can be found, highly reliable and stable authenticity determination can be performed.

[Brief description of drawings]

FIG. 1 is an external perspective view of a document authenticity determination device of the present invention.

FIG. 2 is a schematic configuration diagram of a document authenticity determination device of the present invention.

FIG. 3 is a schematic plan view showing a license used in the document authenticity determination device of the present invention.

FIG. 4 is a flowchart showing an image capturing process operation of the document authenticity discriminating apparatus of the present invention.

FIG. 5 is an explanatory diagram showing an example of the shape of a feature function in infrared light according to the present invention.

FIG. 6 is a flowchart showing the authenticity determination processing operation in a plurality of types of light sources having different wavelengths according to the present invention.

[Explanation of symbols]

11 ... Document authenticity determination device 15 ... Light source for transmission 16 ... Reflection light source 17 ... License 19 ... CCD camera 21 ... Storage unit 22 ... Operation unit 23 ... Display

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Toshiya Hamasaki             Shimogyo-ku, Kyoto-shi Shioji-dori Horikawa Higashiiri Minamifudo-cho             801 OMRON Corporation (72) Inventor Toru Fujii             Shimogyo-ku, Kyoto-shi Shioji-dori Horikawa Higashiiri Minamifudo-cho             801 OMRON Corporation F term (reference) 3E041 AA10 BA11 BB02 BB03 BB05                       BB06 BC03 DB01 EA03                 5B057 AA11 BA02 BA19 DA12 DB02                       DB09 DC22 DC34 DC38                 5C077 LL14 MM03 MP04 PP46 PP55                       PQ12 PQ19                 5L096 AA06 BA03 CA02 CA17 GA51                       HA07 JA03 JA11

Claims (6)

[Claims] 1. A document authenticity discriminating apparatus for discriminating the authenticity of a document, which comprises a plurality of types of light sources having different wavelengths, and the light of the plurality of types of light sources illuminates the transmitted light or the document transmitted through the document. An image pickup unit that receives at least one of the reflected light that is reflected and picks up an image of the document, and a feature amount of a plurality of predesignated portions of the document to be discriminated from the picked-up image of each light source taken by the image pickup unit. Calculating means for calculating, storage means for storing a characteristic function for genuineness extraction obtained from the reference characteristic amount of the document serving as a reference calculated by the calculating means, and the document to be discriminated calculated by the calculating means A document authenticity discriminating apparatus including: a genuineness degree extracting unit that extracts the genuineness degree of a document to be discriminated from the feature amount of the above and the characteristic function stored in advance in the storage unit. 2. A document authenticity discriminating apparatus for discriminating the authenticity of a document, comprising a plurality of types of light sources having different wavelengths, and light of the plurality of types of light sources irradiating the transmitted light passing through the document and the document. An image pickup unit that receives both the reflected light and the reflected light, and calculates the feature amount of a plurality of predesignated parts of the document to be discriminated from the imaged image of each light source picked up by the image pickup unit. Calculation means, storage means for storing a feature function for authenticity extraction obtained from the reference feature amount of the reference document calculated using the calculation means, and the document to be discriminated calculated using the calculation means. And a genuineness degree extracting means for extracting the genuineness degree by multiplying the calculated genuineness degree by numerically calculating the genuineness degree for each light source from the feature quantity and the characteristic function stored in advance in the storage means. Document authenticity determination device. 3. The document authenticity discriminating apparatus according to claim 1, wherein the characteristic function is configured by aggregating distributions of reference characteristic amounts obtained from a plurality of different reference documents. 4. The feature amount of the document to be discriminated and the genuineness degree calculated from the feature function stored in advance are compared with a threshold value set from a reference feature amount stored in the storage means. 3. The document authenticity discriminating apparatus according to claim 1 or 2, further comprising an output means for outputting the authenticity discrimination result obtained by the comparison. 5. A document authenticity discrimination method for discriminating the authenticity of a document, the method comprising: illuminating a document with a plurality of types of light sources having different wavelengths; An image capturing step of receiving at least one of the reflected light reflected and an image of the document, and a feature amount of a plurality of predesignated portions of the document to be discriminated is calculated from the captured image of each light source captured in the image capturing step. A calculating step, and a storing step of storing a characteristic function for extracting a genuineness degree obtained from a reference characteristic amount of a document as a reference calculated in the calculating step,
A document authenticity discriminating method comprising: a genuineness degree extracting step of extracting a genuineness degree of a discriminating document from the characteristic amount of the discriminating document calculated in the calculating step and the characteristic function stored in advance in the storing step. 6. A document authenticity discrimination method for discriminating the authenticity of a document, comprising a plurality of types of light sources having different wavelengths, and the light of the plurality of types of light sources irradiating the transmitted light passing through the document and the document. An image capturing step of receiving both reflected light and an image of a document, and a calculation of calculating feature amounts of a plurality of predesignated parts of the document to be discriminated from imaged images of the light sources captured in the image capturing step. Step, a storage step of storing a feature function for authenticity extraction obtained from the reference feature amount of the reference document calculated in the calculating step, a feature amount of the document to be discriminated calculated in the calculating step, and the storing step And a genuineness degree extraction step of extracting the genuineness degree of the document to be discriminated by multiplying the calculated genuineness degree by the light source and the characteristic function stored in advance in each of the light sources. S authenticity discrimination method.