JPH09261462A - Method, device for processing image and preparing method for reference image data to be used for the method or device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and device for discriminating a specified image while using reference data, with which sufficient recognition accuracy can be provided, and an image data preparing method for preparing these reference data. SOLUTION: When preparing the reference data by averaging the density values of respective picture elements expressing the feature part of specified image of a paper money or the like for each prescribed blocks, the block configuration consisting of these reference data (such as the reference data are composed of 5×5 blocks, for example,) is considered. Especially, in the case of preparing the other reference data from the feature part by rotating the specified image, when the block of new reference data prepared by that rotation is compared with the block of non-rotated reference data, the angle of that rotation is decided so that the picture elements contained in the most separated block can be overlapped with the non-rotated image almost for a half even in that block.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method, an apparatus therefor, and a method for creating reference image data used in the method or apparatus, and in particular, for example, a specific image such as a bill or securities from an input image. The present invention relates to an image processing method for recognizing images, an apparatus thereof, and a method of creating reference image data used in the method or apparatus.

[0002]

2. Description of the Related Art With the recent widespread use of color copying machines and color printers, image processing is executed using a high-performance color scanner and computer, so that it can be prepared by a professional printer unless it is a professional printer. High-quality full-color prints, which were not possible, are now easily available. While these devices have become widespread, a technique for preventing counterfeiting of bills and securities using a color copying machine or a color printer is also required. This forgery prevention technology is an image that should not be faithfully reproduced from the digital image information to be printed, such as banknotes or securities, which should not be faithfully reproduced (hereinafter referred to as "specific image"). ) Is applied to the image recognition technology to detect.

When applying such an anti-counterfeiting technique to a color printer, if the image memory for storing all the image data for one screen to be printed is not provided, a host computer (hereinafter, referred to as a host computer) is used. Whether or not the image to be printed includes a specific image such as a banknote or securities using the image data stored in the image memory such as a line memory having a small capacity for storing the image data transmitted serially from It is required to judge

[0004]

Therefore, the method of performing pattern matching with the reference image from the contour extraction of the image, which has been used in the anti-counterfeiting technology such as the color copying machine, expands all the image data in the frame memory. It is difficult to apply because it needs to be done. Furthermore, in printers that have only a small amount of memory such as line memory, the features of bills and securities (characteristic patterns, numbers, etc.)
If we try to determine whether or not this feature is included in the input image, we find that the feature is in a certain reference direction in the input image (for example, in the case of a laser beam printer, the scanning direction of the light beam). It is not possible to predict what angle the image will be in (or the image forming direction), so in this pattern matching, the angle that the characteristic part that may be included in the input image exists. Even if it does, it is required to detect the characteristic part and recognize it as a specific image. In other words, an analysis considering the arrangement of the characteristic parts is required.

When the reference image data (reference data) is created in advance based on the image information of the characteristic portion of a specific image such as a predetermined banknote or securities in consideration of the requirement regarding the arrangement of the characteristic portion. It is conceivable to rotate the specific image so that the reference line of the specific image has some angles from the reference line such as the image forming direction, and create reference data for each angle. That is, a plurality of reference data having angle dependency is created for one specific image.

When preparing such a plurality of reference data, the rotation angle from the reference line or the angular interval of each reference data uses the image of banknotes or securities which is actually a specific image in the apparatus. If it is decided experimentally, it takes a long time to create the ideal reference data, and it becomes difficult to create the reference data in a realistic time. As a result, the reference data that can be created within a realistic time must compromise the recognition accuracy, and the specific image cannot be recognized with sufficient accuracy.

The present invention has been made in view of the above-mentioned conventional example, and an image processing method and an apparatus therefor capable of discriminating a specific image by using reference data (reference image data) with which sufficient recognition accuracy can be obtained. It is an object of the present invention to provide a method of creating reference image data used in the method or apparatus.

[0008]

In order to achieve the above object, the image processing apparatus of the present invention has the following configuration. That is, image processing capable of comparing template image data expressing a characteristic portion of a specific image with input image data to determine whether or not the input image data includes the specific image. In the apparatus, the input means for inputting image data and the template image data are rotated, the rotated template image data is divided into a plurality of blocks at predetermined rotation angles, and each of the plurality of blocks is a representative value. Storing the processed representative value as one unit by collecting the processed representative value by the plurality of blocks and storing the reference image data for each predetermined rotation angle; and a predetermined storage stored in the storage means. The reference image data for each rotation angle and a comparison means for comparing the image data input by the input means are provided, and the predetermined rotation angle is the template. When rotating the image data, an image processing apparatus characterized by being determined by the degree of overlap before and after the rotation of the plurality of blocks constituting the template image data.

According to another aspect of the invention, the template image data representing the characteristic portion of the specific image is compared with the input image data to determine whether the specific image is included in the input image data. In the image processing method, the template image data is rotated, and the rotated template image data is divided into a plurality of blocks at predetermined rotation angles so that each of the plurality of blocks has a representative value. A storage step of processing and storing reference image data in which a unit of the processed representative values is collected by the plurality of blocks for each predetermined rotation angle, and an input step of inputting image data; A comparison step of comparing the reference image data for each predetermined rotation angle stored in the storage medium with the input image data, wherein the predetermined rotation angle is the temperature of the template. When rotating the over preparative image data,
The image processing method is characterized by being determined by a degree of overlap of a plurality of blocks constituting the template image data before and after the rotation.

According to still another aspect of the present invention, there is provided a method of creating reference image data used in an image processing apparatus or image processing method having the above-mentioned configuration, wherein template image data representing a characteristic portion of a specific image is composed of a plurality of blocks. A dividing step of dividing into a plurality of blocks, a generation step of generating a representative value representative of each of the plurality of blocks based on a pixel value included in each of the plurality of blocks, and a reference image of one unit of the plurality of representative values. A step of determining data, a rotation processing step of rotating the template image data by an angle determined based on a ratio of the size of the one block and a template image size corresponding to the template image data, and the rotation processing step. Using the rotated template image data, the dividing step, the generating step, the determining step, the Comprises creating a reference image data; and a control step of controlling so as to repeatedly perform the rolling process.

[0011]

With the above arrangement, the present invention compares the template image data representing the characteristic portion of the specific image with the input image data and determines whether the specific image is included in the input image data. In determining, the template image data is first rotated, and the rotated template image data is divided into a plurality of blocks for each predetermined rotation angle, and each of the plurality of blocks is processed to have a representative value. Then, the reference image data in which the processed representative values are collected as a unit by a plurality of blocks and stored as one unit is stored in the storage medium for each predetermined rotation angle. Next, image data is input, reference image data for each predetermined rotation angle stored in the storage medium is compared with the input image data, and image formation based on the input image data or the formed image is performed according to the comparison result. Control the image output to the recording medium. Here, the above-described predetermined rotation angle is determined by the degree of overlap of the plurality of blocks forming the template image data before and after the rotation when the template image data is rotated.

The degree of overlap is set to about half of the block size of the block having the smallest degree of overlap among the plurality of blocks forming the template image data. The block is a square area whose one side corresponds to a predetermined number of pixels. For example,
The number of pixels is “256”. Further, the above-mentioned predetermined rotation angle is determined according to the combination configuration of blocks included in the reference image data. For example, the reference image data is composed of 5 × 5 blocks as one unit. The predetermined rotation angle at this time is 1/5 radian unit.

The image data is color image density data, and the density data for one page of the recording medium is sequentially input to the apparatus for each color component. In addition, the device has a line memory that temporarily holds the image data. Further, the characteristic portion of the specific image is taken from paper money or securities. Then, the above-mentioned storage medium stores reference image data regarding the characteristic portion of each of the plurality of specific images. In addition, a non-volatile memory can be used as the recording medium. Further, an electrophotographic method can be used for image formation.

As the representative value, an average density value obtained by averaging the densities represented by the pixels included in each of the plurality of blocks can be used. In the comparison between the input image data and the reference image data, the density value of each pixel of the input image data is averaged over the same area size as the block forming the reference image data, and the average density value obtained as a result is
It is determined whether or not the average density value, which is the representative value, is within a predetermined allowable value. The predetermined allowable value is determined in consideration of the horizontal deviation between the reference image data and the input image data in the comparison means.

According to another aspect of the invention, the template image data representing the characteristic portion of the specific image is divided into a plurality of blocks, and each of the plurality of blocks is divided into blocks based on the pixel value contained in each of the plurality of blocks. Representative representative values are generated, these representative values are determined as one unit of reference image data, and a template is formed by an angle determined based on the ratio of the size of one block to the template image size corresponding to the template image data. A reference image used by the image processing apparatus having the above-described configuration by rotating the image data and using the resulting rotated template image data to control to repeatedly execute the division, generation, determination, and rotation described above. Create the data.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 shows a typical embodiment of the present invention, which is 600 dots / inch (dp).
A color laser beam printer (hereinafter, referred to as CLBP or printer) that has the resolution of i) and performs image formation and recording according to an electrophotographic method based on multivalued data in which each pixel of each color component is represented by 8 bits. It is a sectional side view which shows the structure of 1.

In the apparatus shown in FIG. 1, the sheet 102 fed from the sheet feeding section 101 has a gripper 103f at its leading end.
And is held on the outer periphery of the transfer drum 103. At this time, the detector 8 detects the leading end of the paper 102, and a vertical synchronization signal (described later) is generated based on the detection signal. Image carrier (hereinafter referred to as photosensitive drum) 100
Meanwhile, the latent images formed for each color by the optical unit 107 are as follows:
Each color developing device Dy, Dc, Db, Dn develops the image, and the image is transferred a plurality of times to the paper on the outer periphery of the transfer drum to form another color image. Then, the paper 102 is transferred onto the transfer drum 10.
The sheet is separated from the sheet No. 3, is fixed by the fixing unit 104, and is discharged from the sheet discharge unit 105 to the sheet discharge tray unit 106.

Here, the developing devices Dy, Dc, Db, D for the respective colors
n has rotating shafts at both ends thereof, and each is rotatably held by the developing device selecting mechanism unit 108 about the shaft. As a result, each of the developing devices Dy, Dc, Db, and Dn is configured as shown in FIG.
As shown in FIG. 1, the developing device selecting mechanism 1
Even if 08 rotates around the rotation shaft 110, its posture can be maintained constant. After the selected developing device is moved to the developing position, the developing device selection mechanism unit 108 is integrated with the developing device to support the fulcrum 1.
The selection mechanism holding frame 109 is pulled toward the photosensitive drum 100 by the solenoid 109a around the center 09b, and moves toward the photosensitive drum 100.

Next, the color image forming operation of the color laser beam printer having the above structure will be specifically described.
First, the photosensitive drum 1 is uniformly charged to a predetermined polarity by the charger 111, and, for example, a latent image of M (magenta) color is developed on the photosensitive drum 100 by exposure with the laser beam light L of M (magenta) color. Developed by container Dm,
A first toner image of M (magenta) color is formed on the photosensitive drum 100. On the other hand, the transfer paper P
Is fed, and a transfer bias voltage (+1.8 kV) having the opposite polarity (eg, positive polarity) to the toner is applied to the transfer drum 103.
Is applied to the transfer paper P and the first toner image on the photoconductor drum 100 is transferred to the transfer paper P.
3 is electrostatically adsorbed on the surface. Then, the photosensitive drum 100
The residual M (magenta) color toner is removed by the cleaner 112 to prepare for the next latent image formation and development process.

Then, a second latent image of C (cyan) color is formed on the photosensitive drum 100 by the laser beam L.
Next, the second latent image on the photosensitive drum 1 is developed by the developing device Dc for C (cyan) to form a second toner image for C (cyan). And the second color of C (cyan)
Is transferred to the transfer paper P in accordance with the position of the first M (magenta) toner image previously transferred to the transfer paper P. In the transfer of the second color toner image, the transfer drum 103 is charged with ++ just before the transfer paper P reaches the transfer portion.
A 2.1 kV bias voltage is applied.

Similarly, third and fourth latent images of Y (yellow) color and Bk (black) color are sequentially formed on the photosensitive drum 100, and are sequentially developed by the developing devices Dy and Db. The Y (yellow) color and the Bk (black) color third and fourth toner images are sequentially transferred in alignment with the toner image previously transferred to the transfer paper P. In this way, the four color toner images are formed on the transfer paper P in an overlapping state. In the transfer of the third and fourth color toner images, bias voltages of +2.5 kV and +3.0 kV are applied to the transfer drum 103 immediately before the transfer paper reaches the transfer portion.

The reason why the transfer bias voltage is increased each time the toner image of each color is transferred is to prevent the transfer efficiency from decreasing. The main cause of this decrease in transfer efficiency is that when the transfer paper separates from the photosensitive drum 100 after the transfer, the surface of the transfer paper is charged to a polarity opposite to the transfer bias voltage by air discharge (the transfer paper carrying the transfer paper). The drum surface is also slightly charged), and this charged charge is accumulated for each transfer, and the transfer electric field is reduced for each transfer if the transfer bias voltage is constant.

During the transfer of the fourth color, when the leading edge of the transfer paper reaches the transfer start position (including immediately before and after), the fourth toner image of the fourth toner image is generated at an effective AC voltage of 5.5 kV (frequency is 500 Hz). A DC bias voltage +3.0 kV having the same polarity and the same potential as the transfer bias applied during transfer is superimposed and applied to the charger 111. In this way, in the transfer of the fourth color, when the leading edge of the transfer paper reaches the transfer start position, the charger 1
The operation of 11 is for preventing transfer unevenness. In particular, in the transfer of a full-color image, even if slight transfer unevenness occurs, it is conspicuous as a difference in color. Therefore, it is necessary to apply the required bias voltage to the charger 111 to perform the discharge operation as described above. .

After that, when the leading end of the transfer paper P on which the four color toner images are superposed and transferred approaches the separation position, the separation claw 113 is formed.
Approach, the leading end thereof comes into contact with the surface of the transfer drum 103, and the transfer paper P is separated from the transfer drum 103. The tip of the separation claw 113 keeps contact with the transfer drum surface,
After that, it separates from the transfer drum 103 and returns to the original position. As described above, the charger 111 operates from the time when the leading end of the transfer paper reaches the transfer start position of the final color (the fourth color) to the time when the rear end of the transfer paper leaves the transfer drum 111, and the accumulated charges (on the transfer paper). (The polarity opposite to that of the toner) is removed, thereby facilitating separation of the transfer paper by the separation claw 113 and reducing aerial discharge during separation. When the rear end of the transfer sheet reaches the transfer end position (the exit of the nip formed by the photosensitive drum 100 and the transfer drum 103), the transfer bias voltage applied to the transfer drum 103 is turned off (ground potential). I do. At the same time, the bias voltage applied to the charger 111 is turned off. Next, the separated transfer paper P is transported to the fixing device 104, where the toner image on the transfer paper is fixed, and is discharged onto the paper discharge tray 106.

Next, an image forming operation by laser beam scanning will be described. In FIG. 1, 107 is an optical unit, which is composed of a detector 9, a semiconductor laser 120, a polygon mirror 121, a scanner motor 122, a lens 123, and a mirror 125. When the recording paper P is fed and the leading end of the recording paper P is conveyed to the transfer drum, an image signal VDO for one page is output to the semiconductor laser 120 in synchronization therewith, and the light beam L modulated by the image signal VDO is generated. , The light beam L emitted toward the polygon mirror 121 rotated by the scanner motor 122, and the emitted light beam L by the lens 123 and the mirror 125.
00 is led. When the light beam L is emitted, the light beam L is detected by the detector 9 disposed on the main scanning axis, and a BD (beam detection) signal serving as a horizontal synchronization signal is output. As a result, the photosensitive drum 100 is scanned and exposed by the light beam L in synchronization with the BD signal, and an electrostatic latent image is formed.

The color laser beam printer of the present embodiment outputs an image with a resolution of 600 dots / inch (dpi) through the image forming process as described above. As input data of this apparatus, color image data (for example, Y,
Density image data represented by M, C, Bk components) and the like are considered. Therefore, as shown in FIG. 1, the apparatus includes a printer controller 2 that receives image information and a command (1005) for image formation from a host to generate image data, and a signal that processes the image data. A processing unit 4 is provided.

In the embodiment described here, color image data sent from the host is considered as input data. FIG. 2 is a block diagram showing a functional configuration of the printer 1 according to this embodiment. In FIG. 2, the printer 1 receives image information sent from a host computer (hereinafter referred to as a host) 1000 as density image data for each color component or as PDL command data. YMC composed of bits (D0 to D7)
Printer controller 2 for outputting as Bk image signal 6
And the printer engine 3. Therefore, the value of each pixel of each color component takes a value of 0 to 255.

Various signals other than the image signal 6 are transmitted and received between the printer controller 2 and the printer engine 3 in the form of serial communication. These signals include a page (sub scanning direction) synchronization signal (PSYNC) sent from the printer engine 3 to the printer controller 2, a main scanning direction synchronization signal (LSYNC), and a data transfer clock (VCLK).
The printer controller 2 controls the 8 of each color component of the image signal 6.
The bit signal is output in synchronization with the data transfer clock (VCLK).

FIG. 3 is a block diagram showing the functional arrangement of the printer engine 3 according to this embodiment. In FIG. 3, the reference clock from the reference oscillator 10 included in the optical unit 107 is divided by the frequency divider 11, and the phase difference between the divided clock and the feedback signal from the scanner motor 122 is set to a predetermined phase difference. Scanner motor 1
22 is rotated at a constant speed by the motor control circuit 12 (which includes a known phase control circuit (not shown)). Then, the rotation of the scanner motor 122 is transmitted to the polygon mirror 121 to rotate the polygon mirror 121 at a constant speed.

On the other hand, the transfer drum 103 is rotated at a constant speed by a drive motor (not shown), the leading edge of the recording paper P on the transfer drum 103 is detected by the detector 8, and the vertical synchronization signal (V
SYNC) is output to the signal processing unit 4. Then, the vertical sync signal (VSYNC) defines the image front end of each color. After the vertical synchronizing signal (VSYNC) is output, the BD signal generated by the detector 9 is used as the horizontal synchronizing signal (HSYNC) and
The image signals (VDO) are sequentially sent to the semiconductor laser 120 in synchronization with the D signal.

Further, the CPU 14 built in the signal processing unit 4
Performs serial communication with the printer controller 2,
The control signals are exchanged to synchronize the operations of the printer controller 2 and the printer engine 3. The timings of the above-described vertical synchronization signal (VSYNC), horizontal synchronization signal (BD), and image signals (VDO) of four density color components (YMCBk) in the image forming process are as shown in FIG.

FIG. 5 is a block diagram showing the configuration of the signal processing unit 4. The signal processing unit 4 is roughly divided into a line memory 20, a pattern recognition unit 21, and a PWM halftone processing unit. The line memory 20 stores the multi-valued image data (D0 to D7) sent from the printer controller 2 at the data transfer clock (VCLK) and then reads it out at the image clock (PCLK) of the printer engine 3.

The PWM halftone processing unit is composed of a γ correction unit 22, a D / A conversion unit 23, a comparator 24, and a triangular wave generation unit 25. Then, the multivalued image data from the line memory 20 is γ-corrected by the γ-correction unit 22, converted into an analog signal by the D / A conversion unit 23, and then input to the positive input terminal (+) of the comparator 24. To be done. On the other hand, the negative input terminal (-) of the comparator 23 is supplied with the output signal of the triangular wave generator 25 that generates a triangular wave signal based on the clock of the image clock (PCLK).

Then, the comparator 23 compares these two signals and generates a signal having a pulse width corresponding to the multi-valued image signal. The resolution from the comparator 23 is 600 dp
The PWM signal for forming the image of i is the image signal (VD
O) is sent to the semiconductor laser 121. Now, the pattern recognition section 21 reads out reference data stored in a non-volatile memory 26 such as an EEPROM through a signal line 28 based on a control signal 27 from the CPU 14, and reads out this data and the input multi-valued image data. To see if the input multi-valued image contains a particular image.

FIG. 6 is a block diagram showing the structure of the pattern recognition section 21. As shown in FIG. 6, the pattern recognition unit 21
Is composed of a blocking unit 21a that blocks input multi-valued image data as described below, and a comparison unit 21b that compares the blocked image data with reference data. In this embodiment, Y (yellow), M (magenta), C (cyan), and Bk are sequentially sent from the printer controller 2 for each page of recording paper.
Of the density data of the four color components of (black), only one color component, for example, the density data of the M (magenta) component having the highest luminance is compared with the reference data. As a result, the recognition determination process of the specific image is performed at a higher speed. Therefore, the reference data of one color component is stored in the memory 26. However, it goes without saying that which of the four color components is to be compared with the reference data is not limited to the M (magenta) component described here.

First, the blocking unit 21a inputs multi-valued image data from the line memory 20. This input is performed so that the density data for one page of the recording paper is sequentially received for each color component. Therefore, in the comparison between the input image data and the reference data, the order of the input color components is Y → M → C → Bk, and if the comparison is performed for the M component, the second input color component Only done on data.

Next, the blocking unit 21a converts the input image data into blocks of a predetermined size (here, one side has 256 blocks).
Each pixel (square of a pixel) is converted into data quantized with an average density, and the quantized data is output to the comparison unit 21b. Since the number of pixels corresponding to the size of one page of recording paper is determined by the recording resolution and the paper size,
In this block formation, which part of the image data is included in each block can be known by counting the number of pixels of the input image data for each line.

The comparison unit 21b compares the reference data read from the memory 26 with the input quantized data, and outputs a determination result as to whether the input image data includes the specific image pattern registered in the memory 26. .
Next, the block formation of the input image data will be described in detail. FIG. 7 is a diagram showing an outline of block processing executed in the blocking unit 21a in FIG. Figure 7
As shown in, the average value of the densities of the pixels in each block is calculated based on the image data input from the line memory 20 for each square block having 256 pixels (pixels) on a side, and this is calculated for one block of the blocked image. The value.

Next, the reference data registered in the memory 26 will be described in detail. FIG. 8 shows the memory 2
6 is a diagram showing a format of one reference data stored in No. 6 and used for discrimination of a specific image. FIG. This reference data is obtained by, for example, extracting a characteristic portion from one bill and rotating the portion at a predetermined angular interval to obtain image data corresponding to each image obtained by the blocking unit 2
Blocking is performed in the same manner as executed in 1a to obtain quantized data, and the value of each of these blocks is given a width of ± (allowable width). In this way, the quantized data of each block has a certain width (allowable width) to facilitate recognition of the specific image. In this embodiment, quantized data for 5 (vertical) × 5 (horizontal) blocks is used as one reference data with a value after blocking.

In the example of the reference data shown in FIG. 8, "20" is added above and below the value of the obtained quantized data.
Has a width (allowable value). Now, the comparison unit 21b
In the pattern recognition in, when the block data of the input image data is within the allowable range of the block of the reference data, the block is considered to match, and all the 5 × 5 blocks that match are the input image. If there is, the input image is considered to include the pattern of the specific image represented by the reference data. This result is D / as a discrimination result signal (DSR).
It is output to the A conversion unit 23. In this embodiment, one color component (for example, M (magenta)) is provided for each page of the recording sheet.
Component density image data, the discrimination result signal (DS
R) is output, and if DSR = "1", the specific image is detected, and if DSR = "0", the specific image is not detected.

With respect to the DSR, if DSR = "1", the CPU 14 forms an image of the entire page with maximum density in the image formation of the color component.
3 is controlled. Then, the subsequent image formation of color components is stopped regardless of the reception of the image data even if the image data is sent from the host. In addition, DSR = "1"
In addition to the control as described above, the image forming control when the above is, for example, control such that when DSR = "1" is output, the subsequent image forming is stopped and the image is not output. The black component image density may be maximized so that black toner is output over the entire page, or a specific image pattern is added to the image in a specific color. You may do it.

Next, the relationship between the reference data and the rotation angle will be described in detail. The angle of a circular arc of radius r of length 1 is 1 / r radian [rad]. When the reference data composed of 5 × 5 blocks is rotated around a certain block, the minimum angle at which the block farthest from the center of rotation does not overlap with the block before the rotation becomes the smallest is shown in FIG. As shown, this is the case where the end block of the reference data is selected as the center of rotation. For example, as shown in FIG. 9, it is assumed that each block of reference data is a square having a unit length (“1”), and the reference data is counterclockwise ⅕ around the lower left end of the lower left block. If you rotate only rad], the lower right block will almost not overlap with that before rotation. The radius gyration (r) at this time is r = 5.

On the other hand, when the reference data similarly composed of 5 × 5 blocks is rotated around a certain block, the minimum angle at which the block farthest from the center of rotation does not overlap with the block before rotation is the most. It becomes large when the center block of the reference data is selected as the rotation center, as shown in FIG. For example, as in the case shown in FIG. 10, the reference data is 2 / clockwise counterclockwise around the center of the center block.
When it is rotated by 5 [rad], the upper left block almost does not overlap with that before rotation. The radius of gyration (r) at this time is r = 5/2.

Now, as a reference data used for discriminating a specific image, a case where a characteristic part of the specific image which may be included in the image data input from the host has an angle with the reference data is considered. Then, it is necessary to prepare a plurality of reference data based on the image expressing the characteristic portion obtained at several rotation angles when the specific image is rotated once. The rotation angle (angle interval) at this time is based on the configuration of the reference data and the angle between the characteristic portion of the specific image represented by the reference data and the characteristic portion of the specific image that may be included in the input image data. It can be said that a value between 1 / r [rad] and 2 / r [rad] is suitable when the size of one side of the reference data is “r”.

Which angle is taken between these values depends on how much tolerance is given to the average density value obtained for each block of reference data. However, from the viewpoint of improving recognition accuracy, it is most desirable to make the allowable width as small as possible. Therefore, in order to reduce the allowable width, it is necessary to reduce the rotation angle, in other words, reduce the angle interval. For this reason, in this embodiment, the rotation angle is 1 / r.
[rad] is adopted to improve the recognition accuracy.

FIG. 11 is a diagram showing a state in which the reference data is rotated counterclockwise by 1/5 [rad] with the center of the center block of the reference data as the center of rotation. In this case, the blocks at the end of the reference data are overlapped with each other in a half area of one block. Thus, in this embodiment, the rotation angle is 1 / r [r
ad]], the allowable width given to the average density value obtained for each block unit of the reference data is taken into consideration, taking into account that about half the area of one block will overlap the same block before rotation. I have decided.

FIG. 12 is a diagram showing how to determine the permissible width of the average density value of each block constituting the reference data. The average density of each block of the reference data is obtained by averaging the pixel density values corresponding to the characteristic portion of the specific image such as a bill, which is the original image for creating the reference data, every 256 (vertical) × 256 (horizontal) pixels. can get. However, in actual pattern matching, the feature part expressed by the reference data and the feature part of the specific image that may be included in the input image data do not completely match, and normally, both of them exist in some horizontal direction. To do.

If each of these blocks is ½ block or more, the comparison in the pattern matching may be performed between the adjacent blocks, and one reference data is composed of 5 × 5 blocks. This horizontal deviation is at most 1/2 block in the vertical and horizontal directions,
That is, it is sufficient to consider a shift of 128 pixels. Therefore, in this embodiment, as shown in FIG. 12, the average density value obtained from the template 30 and the template 30 within the search region 31 in which the template 30 is divided into up, down, left, and right by a maximum of 128 pixels (1/2 block).
The maximum value of the difference from the average density value obtained when not disturbing is set as the allowable value of the average density value of each block of the reference data.

Therefore, according to the embodiment described above, the reference data is provided in the number according to the rotation angle determined from the number of blocks forming the reference data, and the average density of each block forming each reference data is provided. The allowable value is determined in consideration of the maximum horizontal deviation between the input image data and the reference data. As a result, it is possible to recognize the specific image using the reference data in consideration of both the horizontal deviation and the rotation deviation between the input image data and the reference data, and thus it is possible to recognize the specific image with higher accuracy. Become.

In the example described above, a plurality of reference data are prepared by paying attention to the average density of blocked pixels, but the present invention is not limited to this. For example, when performing pattern matching by paying attention to the spatial frequencies of the images of the input image and the reference data, a plurality of reference data focusing on the spatial frequency of the image data expressing the specific image are prepared, and It is also possible to recognize the specific image using the reference data.

Further, although the comparison between the input image data and the reference data in the above-described embodiment is performed only for one color component of the input image data, the present invention is not limited to this. For example, the memory may be provided with reference data for all four density color components, and this comparison may be performed for all four density color components. And for all four color components
If it is determined that the specific image is detected, it may be determined that the specific image is included in the input image data, or 4
It may be determined that the specific image is included in the input image data when it is determined that the specific image is detected for any one of the four color components, and further, among the four color components, It may be determined that the specific image is included in the input image data when it is determined that the specific image has been detected for some predetermined color components.

Furthermore, in the above embodiments, density image data was assumed as input image data from the host, but the present invention is not limited to this. For example, a configuration may be adopted in which a color conversion circuit that receives brightness (RGB image data) from the host and performs brightness → density conversion in the printer controller or printer engine is provided. In this case, the comparison between the input image data and the reference data may be performed, for example, by paying attention to the G (green) component that is the component having the highest brightness, or as described above, regarding all the color components. It goes without saying that comparison may be performed.

The printer used in the above embodiments has a line memory, but the present invention is not limited to this. For example, a frame memory capable of storing image data for one page of recording paper. It can also be applied to a printer having a configuration having. Further, the printer used in the above embodiments has a configuration for performing image formation and recording according to an electrophotographic method, but the present invention is not limited to this, and for example, a configuration for performing image formation and recording according to an inkjet method. Can also be applied to other printers.

The present invention can be applied to a system composed of a plurality of devices such as "a host computer, an interface, a printer" and to an apparatus composed of a single device such as "a copying machine". good. Needless to say, the present invention can be applied to a case where the present invention is achieved by supplying a program to a system or an apparatus. In this case, the system or device enjoys the effects of the present invention by reading the program into the system or device from a storage medium storing a program represented by software for achieving the present invention. It becomes possible.

[0055]

As described above, according to the present invention, the template image data representing the characteristic portion of the specific image is divided into a plurality of blocks, and based on the pixel values contained in each of the plurality of blocks, A representative value representative of each of the plurality of blocks is generated, these representative values are determined as one unit of reference image data, and further, based on the ratio between the size of one block and the template image size corresponding to the template image data. Rotate the template image data by a specified angle, and use the resulting rotated template image data to divide, generate, determine,
Since the rotation is repeatedly executed, there is an effect that a plurality of reference image data having angle dependency can be easily created within a realistic time with respect to the characteristic portion of one specific image without trial and error.

Further, in the reference image data thus created, the overlap when the input image data and the reference image data exist at a certain angle is taken into consideration.
Since a plurality of reference image data having angle dependence is provided for the characteristic portion of one specific image, the use of these reference image data has an effect that the specific image can be recognized with higher accuracy. .

[Brief description of drawings]

FIG. 1 is a side sectional view showing a configuration of a color laser beam printer which is a typical embodiment of the present invention.

FIG. 2 is a block diagram showing a functional configuration of the printer 1.

FIG. 3 is a block diagram showing a functional configuration of a printer engine 3.

FIG. 4 shows a vertical synchronizing signal (VS) in an image forming process.
YNC), horizontal sync signal (BD), and image signal (V
It is a figure which shows the timing of (DO).

5 is a block diagram showing an internal configuration of a signal processing unit 4. FIG.

FIG. 6 is a block diagram showing a configuration of a pattern recognition unit 21.

FIG. 7 is a diagram showing an outline of block processing executed in a blocking unit 21a in FIG.

FIG. 8 is a diagram showing an example of reference data.

[Fig. 9] Centering on the lower left corner of the lower left block of reference data, the reference data is rotated counterclockwise by 1/5 [r.
It is a figure which shows a mode when only [ad] is rotated.

FIG. 10 is a diagram showing a state when the reference data is rotated counterclockwise by 2/5 [rad] with the center of the center block of the reference data as the center of rotation.

FIG. 11 is a diagram showing a state in which the reference data is rotated counterclockwise by ⅕ [rad] with the center of the center block of the reference data as the rotation center.

FIG. 12 is a diagram showing a range of a characteristic portion of a specific image to be taken into consideration when determining an allowable width of an average density value of each block forming reference data.

[Explanation of reference numerals] 20 line memory 21 pattern recognition section 21a blocking section 22b comparison section 26 memory 30 template 31 search area

Claims (19)

[Claims] 1. It is possible to determine whether or not the input image data includes the specific image by comparing the template image data expressing the characteristic portion of the specific image with the input image data. An image processing apparatus, comprising: input means for inputting image data; rotating the template image data, dividing the rotated template image data into a plurality of blocks at predetermined rotation angles, and dividing each of the plurality of blocks into a plurality of blocks. Storing the processed representative value as a unit by collecting the processed representative value by the plurality of blocks for each predetermined rotation angle, and storing the reference image data in the storage means. The reference image data for each predetermined rotation angle, and a comparison unit for comparing the image data input by the input unit, the predetermined rotation angle, When rotating the PLATES image data, the image processing apparatus characterized by being determined by the degree of overlap before and after the rotation of the plurality of blocks constituting the template image data. 2. The degree of overlap is about half of the block size even for the block having the least degree of overlap among a plurality of blocks forming the template image data. The image processing device described. 3. The image processing apparatus according to claim 1, wherein the block is a square area whose one side has a predetermined number of pixels. 4. The image processing apparatus according to claim 3, wherein the predetermined number of pixels is “256”. 5. The image processing apparatus according to claim 1, wherein the predetermined rotation angle is determined according to a combination configuration of blocks included in the reference image data. 6. The image processing apparatus according to claim 5, wherein the reference image data is composed of n × n blocks as one unit. 7. The image processing apparatus according to claim 6, wherein the predetermined rotation angle is in a unit of 1 / n radian. 8. The image data is color image density data, and the input unit sequentially inputs the density data for one page of the recording medium for each color component.
An image processing apparatus according to claim 1. 9. The image processing apparatus according to claim 1, wherein the input unit has a line memory that temporarily holds the image data. 10. The image processing apparatus according to claim 1, wherein the characteristic portion of the specific image is a banknote, a securities, or the like. 11. The image processing apparatus according to claim 1, wherein the storage unit stores reference image data regarding a characteristic portion of each of the plurality of specific images. 12. The image processing apparatus according to claim 1, wherein the storage unit is a non-volatile memory. 13. The image processing apparatus according to claim 1, further comprising an image forming unit that forms an image in accordance with a comparison result by the comparison unit. 14. The image processing apparatus according to claim 13, wherein the image forming unit forms an image according to an electrophotographic method. 15. The image processing apparatus according to claim 1, wherein the representative value is an average density value obtained by averaging densities represented by pixel pixels included in each of the plurality of blocks. 16. The comparing unit averages the density values of the pixels of the input image data over the same area size as the block, and the average density value obtained by the blocking unit. 16. The image processing apparatus according to claim 15, further comprising: a determining unit that determines whether the average density value that is the representative value is within a predetermined allowable value. 17. The predetermined allowable value is determined in consideration of a horizontal shift between the reference image data and the input image data by the comparing means. The image processing device described. 18. Comparing template image data representing a characteristic portion of a specific image with input image data,
An image processing method for determining whether or not the input image data includes the specific image, wherein the template image data is rotated, and the rotated template image data is divided into a plurality of blocks for each predetermined rotation angle. Divided into a plurality of blocks, processed so that each of the plurality of blocks has a representative value, and collecting the processed representative values by the number of the plurality of blocks to form one unit of reference image data for each of the predetermined rotation angles. Storing the image data, inputting the image data, and comparing the input image data with the reference image data for each predetermined rotation angle stored in the storage medium. The rotation angle is such that, when the template image data is rotated, a plurality of blocks forming the template image data are overlapped before and after the rotation. An image processing method characterized by being determined by the degree of. 19. A method of creating reference image data used in the image processing apparatus according to claim 1 or the image processing method according to claim 18, wherein a plurality of template image data representing a characteristic portion of a specific image are included. A dividing step of dividing into a plurality of blocks, a generating step of generating a representative value representative of each of the plurality of blocks based on a pixel value included in each of the plurality of blocks, and a plurality of representative values of one unit A step of determining the reference image data; a rotation processing step of rotating the template image data by an angle determined based on a ratio between the size of the one block and a template image size corresponding to the template image data; Using the template image data rotated in the step, the dividing step, the generating step, the determining step, The method of creating the reference image data; and a control step of controlling to execute repeatedly a serial tumbling process.