JP2544002B2 - Card identification device - Google Patents

Description

The present invention relates to a card identification device, and more particularly to a card identification device for identifying the card type based on the feature extraction of the card surface of each card of a plurality of types of cards. Is.

[Prior art] Recently, in the card industry, Affinity Card
The concept has been strengthened, and the light emission of various small-quantity cards has increased. Corresponding to this, quick picking is possible in a short time without mistakes, labor saving of extraction, inventory management that can correctly grasp the number of existing cards, reduction of inventory work, etc.
Labor saving of issuing work is required.

To do this, identify the card correctly,
It is necessary to identify the mistake of picking automatically and try to issue the card accurately. Traditionally, cards have been identified by finding a unique mark on the card.

[Problems to be Solved by the Invention] However, as the types of cards have increased, there are various similar cards, such as those having similar marks and those having the same issuing company but different affiliations. Indistinguishability occurs frequently.

The present invention eliminates the above-mentioned conventional drawbacks and provides a card identification device which eliminates misjudgment and inability to identify even multi-similar cards.

In particular, an improvement is made to the card registration method to provide a card identification device that eliminates misjudgment due to similar cards.

[Means for Solving the Problem] In order to solve this problem, the card identification device of the present invention is a range specifying means for specifying a desired range of at least one place on each type of card surface of a plurality of types of cards. A feature extraction means for extracting features of an image within a designated range of each type of card, a feature registration means for registering the features and the range in association with the type of card, and a type for designating a type of card to be identified For comparing the feature of the image within the range corresponding to the designated type of the unspecified card with the designation unit and the feature of the image registered in the feature registration unit corresponding to the designated type. Card identifying means for identifying a card of a designated type based on the above. Further, the apparatus further comprises registration stopping means for comparing the already registered characteristics with the newly extracted characteristics and stopping the registration when there is no difference more than a predetermined value. Also,
When registration is stopped, a display unit is further provided for displaying images of a card to be registered and similar cards in comparison with each other.

Here, in the case where the present invention is applied to a system that has a plurality of pockets in each of which the same type of card is stored and carries out a desired type of card from one of the plurality of pockets, the feature registration means is , The identification code for identifying the pocket is registered in association with the card type. Further, the feature extraction means determines the average value of the density distribution or the median value of the density range from the density distribution of the image as a threshold value,
The appearance frequency of the binarized data of the image based on the threshold value is used as a characteristic parameter. Further, the feature extracting means defines a mean value of the density distribution or a median value of the density range as a threshold value from the density distribution of the image, and accumulates the binarized data of the image by the threshold value in a predetermined direction. , The cumulative histogram is used as a characteristic parameter.

[Operation] In such a configuration, a desired range of at least one place on each type of card surface of a plurality of types of cards is designated, and the range is stored together with image features in correspondence with the type of card. When is designated, the corresponding range and image characteristics are read out, and by identifying the designated card, it is possible to perform accurate identification without identification error or indiscernibility due to the designation of the range. Further, by stopping the registration when there is no difference equal to or more than a predetermined value, a situation where an identification error or an identification failure is likely to occur is avoided.

Embodiments Embodiments of the present invention will be described below with reference to the accompanying drawings.

<Structure of Card Identification Device> FIG. 1 is a block diagram showing the structure of the card identification device of this embodiment.

The image of the card surface in the window set by the window setting unit 20 is input from the card image input unit 10, and the characteristic of the image is extracted by the characteristic parameter extraction unit 30.

At the time of image registration, the characteristic parameter is compared with the characteristic parameter already registered in the characteristic parameter registration unit 50 by the characteristic parameter comparison unit 40, and the card image determination unit
If the image is judged to be similar in 60, do not register. If they are not similar images, they are newly registered in the characteristic parameter registration unit 50.

At the time of card identification, the characteristic parameter extracted by the characteristic parameter extraction unit 30 is compared by the characteristic parameter comparison unit 40 with the characteristic parameter corresponding to the predetermined card already registered in the characteristic parameter registration unit 50, and the card image determination unit 60. It is determined whether or not the features match. The determination signal is transmitted to a system controller such as a card trunk system in which the present apparatus is incorporated, and if they do not match, a predetermined process such as error display is performed.

It should be noted that the determination as to whether the images are similar images or whether the features match each other is determined according to the error tolerance range corresponding to the input accuracy of the image input unit and the extraction method of the feature extraction unit. .

FIG. 2A shows an example of the hardware configuration of the card identification device of this embodiment.

11, 12 are cameras for inputting images on the front and back of the card, 13 is an A / D converter for converting analog data from the cameras 11, 12 into digital data, and 14 is a CPU for arithmetic control for controlling this device. , A flag 14a indicating whether the image is registered or the card is identified. Reference numeral 15 is a ROM that stores a control program for the CPU 14, 16 is a RAM that has a characteristic parameter registration unit 16a that stores auxiliary storage or characteristic parameters,
An input / output interface 17 controls a keyboard 18 having data input and instruction keys, and a CRT 19 for transmitting information to an operator and displaying an image. It is preferable that the characteristic parameter registration unit 16a is provided in an external mass storage such as a floppy disk, and the frequently used parameters are read into the RAM 16 in terms of processing speed and memory capacity.

FIG. 2B is a diagram showing a configuration example of the characteristic parameter registration unit 16a. The application of this device to a card trunk system, which will be outlined below, will be shown in consideration.

21 is the number of the pocket where the card is stored,
It has a column number 21a and a number 21b in the column. The window 22 is stored at 22a and 22b by the upper left coordinate A and the lower right coordinate B of the desired window 101 in the card 100 as shown in FIG. The designated coordinates of the window are not limited to this example. Reference numeral 23 is a threshold value determined according to the density distribution of an image described later, and 24 is an extracted parameter. In this embodiment, two windows, one on the front side and the other on the back side, are shown for one card, but a plurality of windows may be created on one side, which is achieved by a simple modification of the configuration of the characteristic parameter registration unit 16a. It is possible. In this way, more accurate card identification is possible.

<Feature Extraction Method> Here, two examples of the feature extraction method adopted in this embodiment will be given. These methods are examples of those that have the advantage of being able to clarify the differences corresponding to the image of the card with a small amount of data.

1. Feature extraction by appearance frequency Obtain the density distribution of the read image.

The threshold value is determined based on the concentration distribution, and the threshold value is stored. (Note that the threshold value is determined around the center of the concentration range, as shown in FIG. 4A.
This is because feature extraction can be sufficiently performed even if the density distribution of the image is biased to one side. Alternatively, as shown in FIG. 4B, the average value of the concentration distribution is determined. This is to cut the noise in the low density portion. ) Binarized data is obtained by the threshold value determined in.

The frequency of pixels having a value of "1" in the binarized image is obtained and stored as a characteristic parameter.

 An example of the above feature extraction is shown in Figs. 4A and 4B.

2. Feature extraction by a histogram projected in a predetermined direction A binary image is obtained by the same method as (1).

Binary image data is projected in the vertical direction, pixels having a value of "1" are counted, and a histogram is obtained.

Do the same for the horizontal direction.

The histogram obtained for each is stored as a characteristic parameter.

 The above feature extraction examples are shown in Figs. 5A and 5B.

<Example of application system of this device> FIG. 6 is an external view of a card trunk system to which the card identification device of this embodiment is applied. The card trunk system is a system that has a plurality of pockets in which the same type of cards are stored, and by sequentially specifying the type and number of cards, the specified number of cards is automatically taken out from each pocket in order. By the same applicant
An example is disclosed in No. 9135 (published on December 1, 1991).

In the figure, 301 is a personal computer for system control / management, and 322 to 325 are trunk units consisting of two racks each having 32 pockets for storing a maximum of 500 cards. Each trunk unit has its own rack rotation. And control units 322a to 325a for controlling card removal. 350 is an operator panel for local operation, one for each trunk unit, 332 is an accept stacker having four stacker pockets for storing 500 cards, and 332a is a stack for storing cards that have been rejected by image processing. Rejected stacker 340 is a partition card helper that stores partition cards for each card type. The card identification device of this embodiment is set in the card transport path between the trunk unit 322 and the stacker pocket 332 to determine whether the picked card is the desired card.

FIG. 4 is a diagram showing a specific system configuration of the card trunk system.

As the system controller 300, a personal computer (PC) 301 for control, a floppy disk 302 for reading the card carry-out information created in advance according to the processing order into the system, programs, image data, etc. are stored. A hard disk 303 and a printer 304 that outputs a processing result, an inventory management table, and the like are provided. The PC 301 of the system controller 300 exchanges information between the image processing device 311 of the image processing unit 310 via the system bus and the trunk unit controller 321 of the trunk unit unit 320 via the communication cable (RS232CI / F). The former transmits and receives image data, and the latter transmits commands from the system controller 300 to the trunk unit unit 320.

The image processing unit 310 corresponding to this device includes an image processing device 311 that performs image determination, and cameras (CCD) 312 and 313 that input images on both the front and back surfaces of the card. Image processing device 311 of the image processing unit 310
Exchanges information with the trunk unit controller 321 of the trunk unit unit 320 and the stacker controller 331 of the stacker 330 via the parallel bus, and the card that has been rejected by the image judgment is rejected.
Control for storing in 332a is performed. The trunk unit unit 320 includes a trunk unit controller 321 that controls the control units 322a to 325a of each trunk unit via a communication cable (RS422) and the trunk units 322 to 325 having the above-mentioned capacity. The number of trunk units is not limited to four. The stacker unit 330 is composed of a stacker controller 331 and a stacker 332 of four stages each containing 500 sheets.

<Card Registration / Discrimination Procedure> Next, the procedure for card registration and card discrimination according to the present embodiment will be described with reference to the flow charts of FIGS. 8, 9A and 9B.

First, in steps S10 and S40, based on the instruction from the keyboard 18, it is checked whether the processing is card registration or determination based on the content of the flag 14a. If it is not registered and judged,
Other designated processes are performed, but not shown here.

In the case of registration, proceed to step S12, enter coordinates in the desired window as shown in FIG. 3 (preferably input with mouse from the display screen), and enter registration name and pocket of card as registration information in step S14. Enter the column number and so on. If there is no window instruction, a predetermined default value is set. At step S16, several instruction cards are fed, and at step S18, the image in the predetermined window is input. The instruction card may be put in a predetermined pocket, or the pocket to be fed may be designated. Also, the reason for feeding several images is to take the average value of the image. In step S20, the characteristic parameter is extracted as described above, and in step S22, it is checked whether or not there is a similar parameter already registered. If there is no similar parameter, the process proceeds to step S28, and this characteristic parameter is registered together with the card name, pocket position, window and the like.

On the other hand, if there is a similar parameter, the card indicating the similar parameter is fed at step S24, the image thereof is displayed in contrast with the image of the instruction card at step S26, and the process returns to step S12. The display of step S26 makes it easier to create the window.

FIG. 10 shows an outline of a change example of the display screen during the registration operation.

In the case of determination, the process proceeds from step S40 to S42, and the window corresponding to the card to be determined in step S42 is read out from the characteristic parameters registered in step S44. When you feed the instruction card with step S46,
In step S48, the image in the designated window is subjected to the characteristic parameter extraction processing similar to that at the time of registration. In step S50, the extracted characteristic parameters are compared with the registered characteristic parameters. If they are equal, it is determined that the card is correct in step S52, and if they are not equal, the card error is notified in step S54. It should be noted that the criteria for determining whether or not they are equal to each other, as described above, vary depending on the accuracy of image input and feature extraction, and cannot be generally shown.

FIG. 9A is a flow chart showing the procedure of the first characteristic parameter extraction at the time of registration, and FIG. 9B is a flow chart showing the procedure of the second characteristic parameter extraction.

In step S60, the density distribution of the image in the window is created, and in step S62 the threshold value is determined from the density distribution and stored. The threshold value may be set as an average value of the entire distribution, or may be set as a value that can be expressed as its feature from the plurality of images described above.
That is, noise is apt to be picked up at a low density portion and the characteristics of the card are not accurately expressed, so that the portion is cut. In step S64, the binarized data is created based on the above-mentioned threshold value, and in step S66, the appearance frequency of the pixel which becomes "1" is stored as a characteristic parameter in step S68.

The second characteristic parameter extraction is the same as the second characteristic parameter extraction in steps S80 to S84, and therefore the description of this part is omitted.

In step S86, binary data is projected vertically,
Create a pixel histogram of "1". In step S88, a similar horizontal histogram is created, and the histogram created in step S90 is stored as a characteristic parameter. The projection direction is not limited to this example, and the number of histograms is 2
Not limited to one.

Although this example has been described according to the application to the card trunk system, it is clear that the present apparatus is also useful in other systems.

[Effects of the Invention] According to the present invention, it is possible to provide a card identification device that identifies the type of a card by eliminating misidentification and indiscernibility even for various similar cards.

In particular, an improved card registration method has been provided to provide a card identification device for identifying a card type by eliminating identification errors caused by similar cards.

Further, by using this card identification device, it has become possible to realize a card trunk system or the like that can accurately determine a malfunction.

Furthermore, the simple feature extraction technique enables stable feature extraction even if the density distribution of the image is biased, or more strict feature extraction is possible.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of the card identification device of this embodiment, FIG. 2A is a diagram showing an example of the hardware configuration of the card identification device of this embodiment, and FIG. 2B is an example of the configuration of the characteristic parameter registration unit. Fig. 3, Fig. 3 is a diagram showing a window of this embodiment, Figs. 4A and 4B are diagrams showing a first feature extraction example, and Figs. 5 and 5B are a second feature extraction example. 6 and 6 are external views of a card trunk system to which this device is applied, FIG. 7 is a diagram showing a configuration example of a card trunk system to which this device is applied, and FIG. 8 is a card trunk system to which this device is applied. Flow chart showing the processing procedure, FIG. 9A is a flow chart showing the procedure of the first feature extraction, FIG. 9B is a flow chart showing the procedure of the second feature extraction, and FIG. 10 is an example of a display screen at the time of card registration. FIG. In the figure, 10 ... Card explicit input unit, 20 ... Window setting unit, 30 ... Feature parameter extraction unit, 40 ... Feature parameter comparison unit, 50 ... Feature parameter registration unit, 60 ... Card image determination unit, 11,12 …… Camera, 13 …… A / D converter,
14 …… CPU, 14a …… Flag, 15 …… ROM, 16 …… RAM, 16
a: Characteristic parameter registration unit, 17 ... Input / output interface, 18 ... Keyboard, 19 ... CRT.

Claims (6)

(57) [Claims] 1. A range designation means for designating at least one desired range on each type of card surface of a plurality of types of cards, and a feature extraction means for extracting a feature of an image within the designated range of each type of card. A feature registration means for registering the feature and the range in association with the type of card; a type designating means for designating the type of card to be identified; and a range within a range corresponding to the designated type of an unspecified card Card identification means for identifying a card of a specified type based on a comparison between the characteristics of the image and the characteristics of the image registered in the characteristic registration means corresponding to the specified type. Characteristic card identification device. 2. A registration stopping means for comparing the already registered characteristics with the newly extracted characteristics and stopping the registration when there is no difference more than a predetermined value. Card identification device described. 3. The card identification device according to claim 2, further comprising display means for comparing and displaying images of a card to be registered and a similar card when registration is stopped. 4. The feature registration when applied to a system having a plurality of pockets in each of which the same type of card is stored and carrying out a desired type of card from one of the plurality of pockets. The card identification device according to claim 1, wherein the means registers the identification code for identifying the pocket in association with the type of the card. 5. The feature extraction means determines a threshold value from the density distribution of the image and an average value of the density distribution or a median value of the density range.
The card identification device according to claim 1, wherein the appearance frequency of the digitized data is used as a characteristic parameter. 6. The feature extraction means defines a mean value of the density distribution or a median value of the density range as a threshold value from the density distribution of the image, and 2 of the image by the threshold value is set.
2. The card identification device according to claim 1, wherein the digitized data is accumulated in a predetermined direction, and the accumulated histogram is used as a characteristic parameter.