JP2000341512A - Image reader - Google Patents

Abstract

PROBLEM TO BE SOLVED: To prevent an image reader from being affected by color tone fluctuation by identifying whether or not a specified image is included in an input color image on the basis of a detected edge quantity. SOLUTION: A second solution converting part 38 converts inputted image data to a prescribed resolution. The converted image data are temporarily stored in a memory 54. An edge quantity detecting circuit 55 detects an edge quantity caused by a secondary differentiation operator from the value of a predetermined feature color and generates feature vector data. A template selecting part 56 compares the generated feature vector data with prepared plural templates and selects the template of a shortest Eucledean distance. A buffer 58 stores the edge quantity detected by the edge quantity detecting circuit 55, the number of the template selected by the template selecting part 57, the feature vector data and the Eucledean distance of the template. Then, a CPU 59 for recognition recognizes whether the specified image is included in the image.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a copier, scanner, facsimile,
The present invention relates to an image reading device such as a currency exchange machine, a vending machine, and a banknote pay-in / pay-out device.

[0002]

2. Description of the Related Art Conventionally, in an image reading apparatus for discriminating the authenticity of banknotes and securities, etc., a characteristic pattern of a banknote, a securities, etc. having a specific shape or a specific color distribution is read, and this characteristic pattern is stored. By comparing a reference pattern registered in advance with a reference pattern, the authenticity of bills, securities, and the like is identified.

[0003]

However, in the above-described conventional image reading apparatus, only the characteristic pattern is compared with the basic pattern. There is a problem that it is difficult to identify a specific image.

In this image reading apparatus, it is required that the identification of a specific image is not affected by a change in tint.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an image reading apparatus which is not affected by a change in tint by identifying whether or not a specific image is based on the edge amounts of a plurality of characteristic colors.

[0006]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, an image reading apparatus according to the present invention is an image reading apparatus for judging the authenticity of an original such as a bill or a securities, and an input color image having a predetermined size. An image division unit that divides the image into blocks, an edge amount detection unit that detects an edge amount by counting edge pixels for each block divided by the image division unit, and an edge amount detected by the edge amount detection unit. And an identification unit for identifying whether or not the specific image is included in the input color image.

[0007] Thus, an image reading apparatus which is not affected by a change in tint can be obtained by identifying whether or not a specific image is based on the edge amounts of a plurality of characteristic colors.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An image reading apparatus according to a first aspect of the present invention is an image reading apparatus for determining the authenticity of a document such as a bill or a security, and converts an input color image into blocks of a predetermined size. An image dividing unit to divide, an edge amount detecting unit that detects an edge amount by counting edge pixels for each block divided by the image dividing unit, and an input color based on the edge amount detected by the edge amount detecting unit. And an identification unit for identifying whether or not the image includes a specific image.

With this configuration, since the specific image is identified based on the detected edge amount, the specific image can be identified without being affected by the variation in tint.

According to a second aspect of the present invention, in the image reading apparatus of the first aspect, the image reading apparatus further includes an extraction unit for extracting a plurality of characteristic colors, and the edge amount detection unit includes a plurality of the edge amounts detected by the extraction unit. In this case, the edge amount is detected for the characteristic color.

With this configuration, it is possible to perform edge detection using a characteristic color corresponding to a specific image, and to improve the identification accuracy.

An image reading apparatus according to a third aspect of the present invention is an image reading apparatus for determining the authenticity of a document such as a bill or a security, and an image dividing unit for dividing an input color image into blocks of a predetermined size. An edge amount detection unit that detects a sum of edge amounts for each block divided by the image division unit;
An identification unit that identifies whether or not a specific image is included in the input color image based on the sum of the edge amounts detected by the edge amount detection unit.

According to this configuration, since the specific image is identified based on the detected edge amount, the specific image can be identified without being affected by the variation in tint.

According to a fourth aspect of the present invention, in the image reading apparatus of the third aspect, the image reading apparatus further comprises an extracting unit for extracting a plurality of characteristic colors, and the edge amount detecting unit comprises a plurality of edge colors extracted by the extracting unit. In this case, the edge amount is detected for the characteristic color.

With this configuration, it is possible to detect an edge using a characteristic color corresponding to a specific image, and to improve the accuracy of identification.

Hereinafter, an embodiment of the present invention will be described with reference to FIG.
This will be described with reference to FIG.

(Embodiment 1) An image reading apparatus of a copying apparatus will be described as an example of an image reading apparatus according to an embodiment (Embodiments 1 and 2) of the present invention. Machines, vending machines,
Any device having a function of reading an image, such as a banknote pay-in / pay-out device, is applicable.

FIG. 1 is a configuration diagram showing an image copying system having an image reading apparatus according to Embodiment 1 of the present invention.

In FIG. 1, reference numeral 1 denotes an image reading device;
Denotes an image recording device, 3 denotes a host computer, and 4 denotes a cable. The image reading device 1 reads a document and outputs digital color image data to an external device, the image recording device 2 forms a color image based on image data transferred from the outside, and the host computer 3 controls the image reading device 1 , A plurality of types of commands are output to obtain image data, or image data is output to the image recording apparatus 2. The cable 4 connects the image reading device 1, the image recording device 2, and the host computer 3 to each other, and image data and commands are bidirectionally communicated between the devices by the cable 4.

In the present embodiment, the image reading device 1, the image recording device 2, and the host computer 3 are connected to the SCSI (S
mall Computer System Inte
The image recording device 2 obtains image data from the image reading device 1 by issuing a plurality of commands to the image reading device 1 without the intervention of the host computer 3. An image can also be formed based on image data.

FIG. 2 is a schematic sectional view showing the structure of the image reading device 1 in the image copying system of FIG.

In FIG. 2, 5 is an image reading apparatus main body, 6 is a document glass on which a document to be read is placed, 7 is a carriage for scanning and reading the document, 8 is a stepping motor as a driving source for driving the carriage, 9 Is a drive pulley, 10 is a timing belt, 11 is a belt, 12
Is a driven pulley, 13 is a document, 14 is a document cover, 15 is a support section, and 16 is a reference acquisition position. The carriage 7
It is supported by a support member (not shown) such as a shaft and a rail, and the movement direction is restricted to one direction. Drive source 8
Is transmitted to the drive pulley 9 by the timing belt 10. The belt 11 is stretched between the driving pulley 9 and the driven pulley 12, and moves the carriage 7 in the direction d1 and the opposite direction with the rotation of the driving pulley 9.
Further, the original 13 is read line by line by the movement of the carriage 7. The document cover 14 is supported by a support unit 15 so as to be openable and closable. A white reference plate is stuck on the document glass 6 at the reference acquisition position 16. po
Reference numeral 1 denotes a home position of the carriage 7, and when the image reading apparatus 1 is on standby, the carriage 7 is always located at the home position po1.

FIG. 3 shows a carriage 7 of the image reading apparatus 1.
It is a schematic sectional drawing which shows the internal structure of.

In FIG. 3, reference numeral 17 denotes a lamp for illuminating the original 13, reference numeral 18 denotes an aperture for substantially specifying an image reading position, reference numerals 19-1 and 19-2 denote reflecting mirrors for reflecting light reflected from the original 13, and reference numeral 20. Denotes an image sensor (extraction unit) that converts optical information into an electric signal, and 21 denotes an imaging lens that forms an image on the image sensor 20. The image sensor 20 is fixed inside the carriage 7, is reflected from the document 13, and is reflected by the reflection mirror 19-1 and the reflection mirror 19-1.
The optical information reduced and imaged by the imaging lens 9-2 and the imaging lens 21 is read in a one-to-one relationship with the document surface.

The image reading apparatus 1 configured as described above
The operation will be described below with reference to FIGS. 2 and 3.

When the power of the image reading apparatus 1 is turned on, the carriage 7 is moved regardless of the position before the power is turned on.
It returns to the home position po1. Thereafter, the aperture 18 moves to the reference acquisition position 16 immediately below the reference plate, turns on the lamp 17 to actually read the reference plate, determines the amplification factor for the analog signal output from the image sensor 20, and determines the black and white level. Correction (shading correction) and the like are performed. After that, the home position po1 again
Returns to the standby state.

Next, a single reading operation of the image reading apparatus 1 will be described.

After setting the reading resolution, reading range, and the like from an external host such as the host computer 3 shown in FIG. 1 and then issuing a document reading command, the lamp 17 is turned on and the driving source 8 is turned on. To transmit the driving force to the carriage 7 via the timing belt 10, the driving pulley 9, the belt 11, and the driven pulley 12, and move the carriage 7 in the direction d1. This direction d1 is called a sub-scanning direction. Immediately before the carriage 7 reaches the head of the area corresponding to the reading range set by the host computer 3, the driving speed is changed from the host computer 3 to a speed corresponding to a reading resolution set in advance,
The reading of the document 13 placed on the document glass 6 is started. The original 13 is illuminated by a lamp 17 through the original glass 6, and reflected light from the original 13 is reflected by a reflection mirror 19.
The light is reflected by -1 and 19-2, is reduced and imaged on the image sensor 20 by the imaging lens 21, and is converted into an electric signal. When the reading operation for the designated reading range is completed, the carriage 7 is moved in the direction opposite to the direction d1 to return to the home position po1.

FIG. 4 is a perspective view showing details of the optical system of the image reading apparatus 1. In FIG. 4, to make the drawing easier to see,
The reflection mirrors 19-1 and 19-2 are represented by lines.
In FIG. 4, 22R is a line sensor array for reading a Red signal, 22G is a line sensor array for reading a Green signal, and 22B is a line sensor array for reading a Blue signal. Each line sensor array 22
A color filter corresponding to the color to be read is mounted on the surface of the. As described above, in the present embodiment, an image is read using a so-called three-line color sensor.
The direction of the line sensor array is referred to as a main scanning direction.

Reference numeral 23R denotes a reading line indicating a position on the original glass 6 read by the line sensor array, 23G denotes a reading line indicating a position on the original glass 6 read by the line sensor array, and 23B
Is a reading line indicating a position on the original glass 6 to be read by the line sensor array. Since the position of the line sensor array for reading each color is different in the three-line color sensor, one position (line) of the document cannot be read at the same time. For this reason, as described later, means for delaying the obtained image data by a predetermined amount is required.

FIG. 5 is a block diagram showing each block for performing image data processing in the image reading apparatus 1.

In FIG. 5, reference numeral 8 denotes a stepping motor 8 similar to that of FIG. 2, and 20 denotes an image sensor similar to that of FIG.
4 is an amplifier / A / D converter, 25 is a shading correction unit, 26 is a line correction unit, 27 is a first resolution conversion unit, 2
8 is a color processing unit, 29 is a buffer, 30 is an interface, 31 is another device connected to the image reading device 1,
32 is a CPU, 33 is a motor control unit, 37 is a specific image recognition unit, 38 is a second resolution conversion unit, 39 is a recognition unit, 40 is a serial communication line, and 34 to 36 are control signals.

As described above, the image sensor 20 is composed of a three-line sensor array, and outputs analog image information for each line of R, G, and B colors. Amplifier A
The / D converter 24 amplifies the analog image information output from the image sensor 20 with a predetermined gain,
The analog signal amplified by the A / D converter is converted into a digital signal. The shading correction unit 25 normalizes the input digital image signal to a previously acquired dynamic range of black and white. The line correction unit 26 corrects the difference between the line sensor array positions of the respective colors described above, and makes the R, G, and B lines equivalent to reading the same document position (line). The operation of the line correction unit 26 will be described later in detail. The first resolution conversion unit 27 converts the resolution of the image data output from the line correction unit 26 based on parameters specified by the host computer 3 and the image recording device 2.
The operation of the first resolution converter 27 will also be described later in detail. The color processing unit 28 enables a vivid color reproduction by reducing the influence of unnecessary absorption bands on the spectral spectrum existing in the color filters on the line sensor array. The buffer 29 temporarily stores the image data processed in the above process. This is a means for absorbing a difference in communication speed with the outside and outputting image data to an external device at a higher speed. The other device 31 substantially refers to the host computer 3 or the image recording device 2 in the present embodiment. In this embodiment, the image reading device 1 and the other device 31 are connected by SCSI, and the image reading device 1 outputs image data to the other device 31 via SCSI, and outputs the image data to the other device 31. From 31, reading parameters such as a reading range and a reading resolution can be obtained.

The CPU 32 controls the operation sequence of the image reading apparatus 1 and the like, and the motor control unit 33 sends a drive signal (more precisely, an excitation signal for the stepping motor) to the motor 8 for moving the carriage 7 of the image reading apparatus 1. Output. The CPU 32 controls the operation of the line correction unit 26 by the control signal 34 and
Controls the operation content of the first resolution conversion unit 27, and controls the rotation speed of the motor 8 via the motor control unit 33 by the control signal 36.

The specific image recognition section 37 detects whether a specific image exists in the read image data. The second resolution converter 38 converts the read image data into a predetermined resolution, for example, 75 dpi (dot per inch). The recognition unit 39 recognizes a specific image based on the image data converted to a predetermined resolution by the second resolution conversion unit 38. The serial communication line 40 is connected to the recognition unit 39 and the CP.
U32 is connected, and the recognition unit 39 and the CPU 32 can exchange information by performing bidirectional communication. Regarding the configuration and operation of the specific image recognition unit 37,
Details will be described later.

Next, the optical system of the image reading apparatus will be described in detail with reference to FIG. FIG. 6 is a schematic view of the carriage 7 of the image reading device 1 as viewed from the side. For simplicity, the lamp 17 and aperture 18 shown in FIG. 3 are omitted.

The line sensor array 22R arranged on the image sensor 20 reads the red image information. The position of the read line on the original glass 6 is PR. The line sensor array 22G reads the image information of Green, and the position of the reading line on the original glass 6 is PG. Also, the line sensor array 22B
Reads the image information of Blue, but the position of the reading line on the original glass 6 is PB.

Assuming now that an image is being read, the carriage 7 is moving in the sub-scanning direction (d1 direction), and the position of the PB becomes a reading line with respect to the document 13 first, and then the PG of the PG. The position is finally the reading line at the position of PR. In other words, the same position (line) on the document
, First, Blue image data is obtained, then Green, and finally, Red image data is obtained. The blue image data obtained first and the green image data obtained next are held for a predetermined number of lines,
When the image data of
And output the image data of Green, the R, G, and B line positions can be aligned and output.

Next, the configuration of the image sensor 20 alone will be described. FIG. 7 is an array diagram of the image sensor 20 viewed from the line sensor array side. The line sensor arrays of each color are arranged in a line in the main scanning direction, and there are intervals of L1 and L2 between the line sensor arrays of each color in the sub-scanning direction. By the way, in FIG.
Indicates individual pixels of the line sensor array. For simplicity, □ indicates the size of one pixel at 600 dpi of the image reading apparatus.

In a general image sensor, L1 and L2 are equal, and L1 and L2 each have a value that is an integral multiple of the read pixel size. For example, in this embodiment,
L1 and L2 correspond to eight lines when converted to a line of 600 dpi. That is, the line sensor array of each color is 600 dpi.
/ 8 = 75 dpi. As described above, the same position (line) cannot be read simultaneously with an image sensor having such a structure, and the line correction unit 26 corrects this as described above.

Next, the operation of the line correction section 26 will be described in detail with reference to FIG. FIG. 8 is an operation explanatory diagram for explaining the operation principle of the line correction unit 26. In FIG. 8, reference numeral 50 denotes a memory area for storing Green image data in line units, and reference numeral 51 denotes a memory area for storing Blue image data in line units.

Image reading device 1 according to this embodiment
Are Blue, Green for the same line of the original
The data is read in the order of n and Red. Since the interval between each line sensor array is equivalent to 8 lines of 600 dpi,
When reading an image at 600 dpi, the image data of 8 lines is read for Green image data, and
The image data of 16 lines is stored for the image data of ue, and when the image data of Red is read, the image data of 8 lines before is read for the image data of Green, and the image data of 16 lines is stored for the image data of Blue.
If the image data before the line is output, it is the same as reading at the same position on the document.

In this way, the image is once read at 600 dpi in the sub-scanning direction, converted to a lower resolution after performing the above-described line correction, and if the resolution is lower than 600 dpi, the image can be read at all resolutions. Can be. However, in this case, there is a premise that an image is read once at 600 dpi, so that the reading speed cannot be increased. This problem can be dealt with by reading the image while moving the carriage 7 in the sub-scanning direction at a high speed, and changing the setting of the line correction unit 26.

FIG. 9 is an operation explanatory diagram for explaining the operation of the line correction section 26 when an image is read at a resolution of 300 dpi in the sub-scanning direction. Assuming that the moving speed of the carriage 7 when reading a document at 600 dpi, that is, the moving speed in the sub-scanning direction d1, is V, the moving speed of the carriage when reading a document at 300 dpi is set to 2V. That is, the moving speed of the carriage is set to be twice that at the time of reading at 600 dpi. The carriage moving speed Vx at an arbitrary reading resolution is, for example, assuming that the reference reading resolution is 600 dpi, the carriage moving speed in reading at 600 dpi is V, and the actual reading resolution is X (dpi), Vx = (600 / X) × V ... (1)

In the case where an image is read at 300 dpi, the moving speed of the carriage is twice that of 600 dpi, so that the moving distance per unit time is also twice. Since the distance between the line sensor arrays of the respective colors does not always change, if the moving speed of the carriage is doubled, the distance that the image reading device moves when reading one line of image data is also doubled and stored. The number of lines of the image data to be stored may be １ ／. That is, as shown in FIG. 9, the interval between the line sensor arrays is equivalent to eight lines of 600 dpi, ie, 3 lines.
When reading an image at 300 dpi, image data for 4 lines is stored for Green image data, and image data for 8 lines is stored for Blue image data.
When the image data of Red is read, the image data of four lines before is read for Green image data, and
If the image data of eight lines before is output with respect to the lue image data, it is the same as reading at the same position on the document. Table 1 shows a generalization of the above.

[0046]

[Table 1]

That is, in the present embodiment, as shown in (Table 1), the reading resolution is set to an integral multiple N based on 75 dpi. At this time, the Green memory area 5
The number of delay lines of Green image data stored in 0 can be generalized to N, and the number of delay lines of Blue image data stored in the Blue memory area 51 can be generalized to 2N. These settings are made to the line correction unit 26 by the control signal 34 from the CPU 32 in FIG. Further, the carriage moving speed Vx at each resolution is given by equation (1). This setting is made to the motor control unit 33 by the control signal 36 from the CPU 32 in FIG.

As described above, the difference in the reading position caused by the difference in the position of the line sensor array of the image sensor 20 is corrected.
Is in the same state as when the same line of the document is read.

As described above, the line correction unit 26
Is relative to the carriage movement direction, that is, the sub-scanning direction.
Correct the difference in the reading position of each color. At this time, the reading resolution is specified as a discrete value. However, the actual image reading apparatus 1 receives the reading resolution specification in units of 1 dpi from the host computer 3 or the image recording apparatus 2 and corrects the image data. Output. The processing performed by the line correction unit 26 is alignment in the sub-scanning direction, and no conversion is performed on image data in the main scanning direction.

The first resolution converter 27 performs these processes. Hereinafter, the processing in the first resolution conversion unit 27 will be described in detail.

First, a description will be given with reference to FIG. For the sake of simplicity, it is assumed that the image reading apparatus 1 is externally designated to read at 200 dpi. When reading at 200 dpi is designated, the CPU 32
For 3, the carriage moving speed for the reading resolution of 225 dpi is set. According to (Table 1), this is 2.7 times the carriage moving speed V at 600 dpi. Next, the CPU 32 similarly sets the reading resolution of 225 dpi for the line correction unit 26. That is, the Green memory area 5
The delay amount at 0 is set to three lines, and the delay amount at the Blue memory area 51 is set to six lines (FIG. 8).
Or see FIG. 9). When the image is read with these settings, the line correction unit 26 outputs image data converted to 225 dpi in the sub-scanning direction.

Here, as an example, a case in which an image is read at a resolution of 225 dpi when a resolution of 200 dpi is designated has been described. However, the designated value of the reading resolution for the image reading apparatus 1 in the present embodiment and the motor Table 2 shows the relationship between the setting contents for the control unit 33 and the line correction unit 26, that is, the actual reading resolution.

[0053]

[Table 2]

FIG. 10 is an algorithm diagram showing a resolution conversion algorithm.

First, the resolution conversion algorithm in the main scanning direction will be described in detail with reference to FIG. FIG.
At 0, 53 indicates one pixel of 600 dpi. However, for the sake of simplicity, the actual pixel size is ignored and 6
The center position of one pixel of 00 dpi is shown. 600d
P600_0, P600
The numbers 600_1, P600_2,..., P600_6 are given, and these are codes indicating the positions of the pixels. For convenience, the values of the pixels for these positions are
For example, the pixel value corresponding to the position of P600_0 is * P60
It is represented as 0_0. That is, the concept of the pointer in the C language is used.

First, image information of 600 dpi is converted to 200d.
The case of conversion to pi will be described. The position of the first pixel after conversion is always the first pixel of 600 dpi, that is, P600
It shall be aligned to the position of _0. Therefore, the first pixel position at 200 dpi is P200_0, which is the same as P600_0. Since the location is the same, the pixel value adopts the same value as P600_0, that is, * P600_0.

The next pixel position is P200_1. To obtain this pixel value, the position of P200_1 is set to 600d.
Consider that the pixel position is represented by pi. Since (600/200) × 1 = 3 using a simple proportional expression, P20
0_1 = P600_3. Therefore, the pixel value at the position of P200_1 is * P200_1 = * P600_3.
Similarly, * P200_2 = * P600_6 can also be obtained.

Next, the image information of 600 dpi is converted to 300d.
The case of conversion to pi will be described. The position of the first pixel after conversion is always the first pixel of 600 dpi, that is, P600
It shall be aligned to the position of _0. Since the head pixel position of 300 dpi is the same as P600_0, the pixel value is P
The same value as 600_0, that is, * P600_0 is adopted.

The next pixel position is P300_1. To obtain this pixel value, the position of P300_1 is set to 600d.
Consider that the pixel position is represented by pi. Since (600/300) × 1 = 2 using a simple proportional expression, P30
0_1 = P600_2. Therefore, the pixel value at the position of P300_1 is * P300_1 = * P600_2.
Similarly, * P300_2 = * P600_4, and further *
P300_3 = * P600_6 can be obtained.

Next, the image information of 600 dpi is converted to 400d.
The case of conversion to pi will be described. The position of the first pixel after conversion is always the first pixel of 600 dpi, that is, P600
It shall be aligned to the position of _0. Since the head pixel position of 400 dpi is the same as P600_0, the pixel value is also P600
The same value as 600_0, that is, * P600_0 is adopted.

The next pixel position is P400_1. To obtain this pixel value, the position of P400_1 is set to 600d.
Consider that the pixel position is represented by pi. When calculated using a simple proportional expression, (600/400) × 1 = 1.5, which indicates that P400_1 exists between P600_1 and P600_2. Therefore, the pixel value of P400_1 is calculated as in the following equation (2) using the position information of 1.5.

* P400_1 = (1.5-1) × (* P600_1) + (2-1.5) × (* P600_2) (2) This is the position where the pixel after resolution conversion exists. Is 600dp
i is determined based on the pixel position of i, and the adjacent 600 dpi
By performing a weighting operation based on the distance to the pixel of
This is nothing but finding pixel values after resolution conversion.

When the above concept is applied to P400_2, (600/400) × 2 = 3.
_2 exists at the position of P600_3. Therefore, * P400_2 = * P600_3. Further P4
When the above concept is applied to 00_3, (600
/400)×3=4.5, and P400 — 3 is P60
It can be seen that it exists between 0_4 and P600_5. Therefore, the pixel value of P400_3 is calculated as in the following equation (3) using the position information of 4.5.

* P400_3 = (4.5-3) × (* P600_4) + (5-4.5) × (* P600_5) (3) The pixel values of the subsequent pixels are similarly set. You can ask.

The resolution conversion to 500 dpi can be processed in exactly the same way.

As described above, the conversion from 600 dpi to another resolution has been described as the resolution conversion processing in the main scanning direction.
It is not an arithmetic method applied only to, but a method applicable to any case as long as the original resolution and the converted resolution are known. For example, 225 dp by line correction
Image data in the sub-scanning direction output at the resolution of i can be converted to, for example, 200 dpi in the same manner. In the present embodiment, the method described above is used for resolution conversion in the sub-scanning direction.

Next, the specific image recognition unit 37 will be described in detail with reference to FIGS. FIG. 11 is a block diagram showing the specific image recognition unit 37 in detail.

In FIG. 11, reference numeral 30 denotes an interface similar to that of FIG. 5, 31 denotes another device similar to that of FIG.
5, a second resolution converter similar to that of FIG. 5, 40 a serial communication line similar to that of FIG. 5, 54 a memory, 55 an edge amount detection circuit, 56 a template selector, 57 a template storage A memory, 58 is a buffer, 59 is a recognition CPU, 60 is a main / sub pixel counter (image division unit), 61 is a ROM, and 64 is a work RAM. Here, the template selection unit 56, the template storage memory 57, and the recognition CPU 59 constitute an identification unit for identifying whether or not the image is a specific image.

In FIG. 11, the second resolution converter 38
Converts input image data into a predetermined resolution.
The image data converted to a predetermined resolution by the second resolution converter 38 is temporarily stored in the memory 54. The edge amount detection circuit 55 detects an edge amount by a second derivative operator (described later with reference to FIG. 12) based on a predetermined characteristic color value, and generates feature vector data according to the detected edge amount. The template selection unit 56 compares the feature vector data generated by the edge amount detection circuit 55 with a plurality of templates prepared in advance, and selects a template having the closest Euclidean distance. Template storage memory 5
7 stores a plurality of templates used by the template selection unit 56 for comparison with the feature vector. The buffer 58 stores the edge amount detected by the edge amount detection circuit 55, the template number selected by the template selection unit 57, the feature vector data, and the Euclidean distance of the template. The recognition CPU 59 recognizes whether or not a specific image is included in the image. The main / sub-pixel counter 60 counts the number of input image data in the main scanning direction and the sub-scanning direction, and outputs an interrupt signal 63 to the recognition CPU 59 every time a predetermined count is reached.
The ROM 61 stores a recognition program and the like. The recognition CPU 59 notifies the second resolution converter 38 of resolution conversion parameters and the like.

The input to the second resolution converter 38 is performed before the first resolution converter 27. The reason will be described below. As described above, the image data output from the line correction unit 26 corrects the distance between the RGB lines in the sub-scanning direction due to the different positions of the line sensor arrays for each color. At this time, the resolution in the main scanning direction remains output from the image sensor, and no processing is performed. That is, in the configuration described above, the line correction unit 26 outputs 600 dpi in the main scanning direction.
Is output.

As described above, at the time of output from the line correction section 26, the resolution in the main scanning direction is always 600 dp regardless of the designation of the reading resolution by the other device 31.
i, it is fixed to a predetermined resolution, for example, 75
Conversion to dpi can be performed by a single, parameter-invariant processing system. If the output of the first resolution converter 27 is to be converted to a predetermined resolution, for example, 75 dpi, image data of various resolutions must be handled, and the hardware becomes complicated.

In the sub-scanning direction, the line data output from the line correction section 26 is 75 dpi, 150 dpi, 225 dpi, 300 dpi as shown in Table 2.
dpi, 375 dpi, 450 dpi, 525 dpi,
600 dpi. Most importantly, they are all multiples of 75 dpi.
It is very easy to convert these data to the above-mentioned predetermined resolution and 75 dpi.

Now, the recognition CPU of the specific image recognition unit 37
Reference numeral 59 denotes a serial communication line connected to the CPU 32 of the image reading apparatus. The CPU 32 obtains the image reading conditions transferred from the other device 31 via the interface 30, and based on the obtained conditions, the image reading device 1
The control of the line correction unit 26, the first resolution conversion unit 27, and the motor control unit 33 is as described above.
The U 32 also notifies the recognition CPU 59 of the reading conditions regarding these resolutions via the serial communication line 40. Thereby, the recognition CPU 59 can know the resolution of the image to be read. Based on this information, the recognition CPU 59 outputs the second
For the resolution conversion unit 38, the processing in the sub-scanning direction, more specifically, the thinning rate for all lines is specified. Of course, since the main scanning direction is constant regardless of the reading resolution, the pixel thinning rate in the line is fixed. (Table 3) shows the setting contents of the thinning rate for the second resolution converter 38.

[0074]

[Table 3]

As described above, by setting the fixed pixel thinning rate to 2 in the main scanning direction, image data of 600 dpi is always converted to 300 dpi. By performing the thinning process in this manner, the amount of image data to be processed thereafter can be significantly reduced. In the sub-scanning direction, the line thinning rate is changed according to the reading resolution. Thereby, as shown in the column of resolution after thinning, the resolution of main scanning × sub scanning is 3
00 dpi x 75 dpi or 300 dpi x 150d
pi.

Next, the image data obtained by the thinning-out process is averaged by 75% in both the main scanning and sub-scanning directions.
Convert to dpi. Hereinafter, this 75 dpi is referred to as a predetermined resolution. First, 300dp in the main scanning direction by thinning processing
When the image is converted to i.times.75 dpi in the sub-scanning direction, the value of 4.times.1 pixel is averaged using four pixels in the main scanning direction and pixels for one line in the sub-scanning direction. In addition, 300d in the main scanning direction × 150d in the sub-scanning direction by the thinning process.
When converted to pi, 4 × 2 pixel values are averaged using four pixels in the main scanning direction and pixels for two lines in the sub-scanning direction.

Through the above processing, image data of a predetermined resolution of 75 dpi is obtained in both the main scanning and sub-scanning directions.
However, the predetermined resolution is not limited to 75 dpi.

Next, an outline of a specific image recognition algorithm in the image recognition apparatus according to the present invention will be described with reference to FIG.

The RGB image signal converted to a predetermined resolution by the second resolution converter 38 is temporarily stored in the memory 54. The RGB image signal stored in the memory 54 is cut out in block units of a predetermined size,
The signal is sent to the edge amount detection circuit 55 as a point B sequential signal.
The size of the block is, for example, 50 × 50 pixels (250
0 pixels).

The edge amount detection circuit 55 receives the inputted RGB
Edge detection is performed on the image signal on the basis of the RGB values determined in advance as the characteristic colors, and the one block (50 × 50) is detected.
The edge amount within the pixel is detected. When the detection of the edge amount of one block is completed, the result is transferred to the template selecting unit 56 and written into the buffer 58. The output from the edge amount detection circuit 55 is obtained by counting the edge amounts for a plurality of characteristic colors existing in a 50 × 50 pixel block. Since the number of characteristic colors is 3, this can be regarded as outputting a three-dimensional characteristic vector. That is, the edge amount detection circuit 55 generates a feature vector. The template selection unit 56 stores the feature vector generated by the edge amount detection circuit 55 and the template storage memory 5
7 is compared with a plurality of templates stored in advance based on the three-dimensional Euclidean distance, the closest template is selected, and the template number and the three-dimensional Euclidean distance are stored in the buffer 58. The nearest neighbor template number and the three-dimensional Euclidean distance serve as indices indicating the similarity between the input image data and the specific image.

The total number of pixels processed by the edge amount detecting circuit 55 is counted and managed by the main / sub-pixel counter 60. When the counted number of pixels processed reaches a predetermined amount, the main pixel count is counted. The sub-pixel counter 60 is a CPU 5 for recognition.
9 to generate an interrupt signal 63. Upon receiving the interrupt signal 63, the CPU 59 reads the buffer 58,
The edge amount detection result, the nearest template number, and the three-dimensional Euclidean distance stored here are obtained. CP
U59 selects the similarity based on the nearest template number read from the buffer 58 and the similarity based on the three-dimensional Euclidean distance for a plurality of blocks according to a certain rule, calculates the sum of them, and outputs a determination result according to the sum. The determination result is transmitted from the recognition CPU 59 to the CPU 32 via the serial communication line 40. The CPU 32 outputs the result to the interface 30, and this determination result is output to another device 31 such as the host computer 3 or the image recording device 2 by SCSI.

Next, the edge amount detection circuit 55 will be described in detail. FIG. 12 is a block diagram showing the edge amount detection circuit 55.

In FIG. 12, reference numeral 55 denotes an edge amount detection circuit similar to that of FIG. 11, reference numeral 56 denotes a template selection unit similar to that of FIG. 11, reference numeral 57 denotes a buffer similar to that of FIG.
0_C1 and 70_C2 are second derivative operators, 71 is an absolute value circuit, 72 is a comparator, 73 is a counter, and 74 is a count buffer. Second derivative operator 70_C0,
70_C1 and 70_C2 detect edges with respect to independent and predetermined characteristic color values, respectively.
In the present embodiment, since the number of edge pixels is detected for each of the three characteristic colors, it has three secondary differential operators. In the present embodiment, an input RGB signal is divided into three characteristic colors, and Red and Gre are used.
Although the colors of en and Blue are used, the characteristic color i can be obtained as in the following equation (4) using the coefficients ri, gi, and bi (i = 0, 1, 2).

Characteristic color i = ri × R + gi × G + bi × B (i = 0 to 2) (4) Second derivative operator 70_C0, 70 for each characteristic color
Since _C1 and 70_C2 have no structural difference, only one color is shown in detail. Second derivative operator 70_C
0 has a window of 3 pixels × 3 pixels, and has a coefficient of “8” for the center pixel and “−1” for all eight pixels other than the center. The edge amount is calculated by multiplying the coefficient in the window of the second derivative operator 70_C0 by the value of the characteristic color corresponding to the position of the window, and adding the respective multiplied values. The absolute value circuit 71 calculates the absolute value of the edge amount obtained by the second derivative operator 70_C0. The comparator 72 compares the data output from the absolute value circuit 71 with a predetermined threshold value TH, and detects whether the pixel of interest is an edge pixel. The data output from the absolute value circuit 71 is set to a predetermined threshold value TH.
In the case of (e.g., 50) or more, the target pixel is determined to be an edge pixel, and the output of the comparator 72 is set to "1". The data output from the absolute value circuit 71 is a predetermined threshold T
If less than H (for example, 50), it is determined that the target pixel is not an edge pixel, and the output of the comparator 72 is set to “0”. The counter 73 counts the number of times the output of the comparator has become “1”. The count buffer 74 accumulates the count result of the counter 73.

The counting is performed on a block basis. Here, the block is obtained by dividing the read image in units of a plurality of pixels in the main scanning direction and the sub-scanning direction. In this case, the pixels of a predetermined resolution converted by the second resolution conversion unit 38 are set in units of 50 pixels. A rectangle of 50 × 50 pixels is defined as one block. Therefore, the counter 7
Reference numeral 3 indicates that the count result is stored in the count buffer 74 for each input of 50 pixels and reset. Count buffer 74
Exist in the number of blocks in the main scanning direction, and one block of data is recorded in the sub-scanning direction. At the time of recording from the counter 73 to the count buffer 74, the addition result to the data already written in the count buffer 74 is always written, that is, by performing the read-modify-write operation, the characteristic color of one block in the sub-scanning direction is obtained. The number of pixels is accumulated. When the data input for one block is completed in the sub-scanning direction, the contents of the count buffer 74, that is, the result of counting the characteristic colors obtained for each block are stored in the buffer 58 and passed to the template selection unit 56.

The above operation is performed by the second derivative operator 70_C
The first and second derivative operators 70_C2 also perform the operation in parallel, and the edge pixels are counted with respect to the predetermined characteristic color signal, and the count result is 3
Buffer 58 as a dimensional vector or feature vector
And passed to the template selection unit 56.

FIG. 13 is a data structure diagram showing the structure of data stored in the buffer 58. In the figure, a thick solid line indicates the break of each block, and C0 (n), C1
(N) and C2 (n) indicate the results of edge pixel counting counted in the n-th block, respectively, and indicate that one block feature data is composed of three edge pixels.

Next, the template selecting section 56 will be described in detail. FIG. 14 is a flowchart showing the operation of the template selection unit 56. In the following description, FIG.
And FIG.

When a feature vector composed of three edge amounts is passed from the edge amount detection circuit 55 to the template selection unit 56 for each block, the feature vector is compared with the template. First, a feature vector for each block is obtained (S1). The acquired data is Cn = (C0 (n), C1
(N), C2 (n)) (where n is the block number). It is determined whether the magnitude | Cn | of Cn is equal to or greater than a predetermined value (S2). If the number is equal to or more than a certain value, a template closest to Cn is searched from the templates stored in the template storage memory 57. The template in the template storage memory 57 is Tm = (TC0
(M), TC1 (m), TC2 (m)) (where m is a reference data number and m = 1 to M), and the distance Dnm = | Cn−Tm | (3 When the Euclidean distance of the dimensional vector is minimized, Dnm (D
min) and detects the template number m and the distance data D
min is stored in the buffer 58 (S3). If | Cn | does not exceed a certain value in step 2, the template search in step 3 is not performed, and a template number (eg, M + 1) for which no template is defined and a value Dmax equal to or more than the maximum value that Dnm can take are used. Buffer 5
8 (S4).

Here, the template stored in the template storage memory 57 will be described in detail. The template storage memory 57 is a RAM, and the ROM 61
CPU 5 for recognizing template data stored in advance in
9 is copied. The template is obtained in advance from the specific image to be processed, and is stored in advance.

FIGS. 15A to 15D are relationship diagrams showing the relationship between the template and the specific image. The template is
As shown in FIG. 15, when the target specific image is placed at the horizontal position as a reference (FIG. 15A), when the target specific image is rotated from the horizontal position by a minute angle unit (FIG. 15). (B)), (FIG. 15 (c)), and the characteristics of each block when the positional relationship between the block and the specific image is shifted by several pixels in the horizontal and vertical directions (FIG. 15 (d)). A template obtained by calculating an edge amount for a color is used as a template. However, since the number of templates obtained as described above is enormous, a representative one is extracted by using a clustering technique such as vector quantization and stored in the ROM 61.

The specific image recognition process will be described first with reference to FIG.
This will be described with reference to FIG. The output of the template selection unit 56 is temporarily stored in the buffer 58. When these processes have been completed for a predetermined number of blocks, the main / sub-pixel counter 60 generates an interrupt signal 63 and requests the recognition CPU 59 to acquire the contents of the buffer 58. Recognition CPU5
9 reads the contents of the buffer 58, and
To be stored.

FIG. 16 is a data structure diagram showing a data structure in the working RAM 64. In the figure, TN (n) is a template number closest to the feature vector obtained for each block. D (n) is the distance Dmin or Dmax between each block and the template closest to the feature vector.

FIGS. 17A to 17C show TN (n) and D (n) given to each block of an actual specific image.
It is an image figure showing the image of. FIG. 17A shows an image including a specific image, and FIG. 17B shows the value of TN (n) for each block. Here, the template number is 254 at maximum, and is defined as a template. Unassigned numbers are 254 + 1 = 255. FIG. 17 (c) shows the value of D (n) for each block, and the portion of 00 in the figure shows Dmax or a value close to Dmax, which is not quite similar to the edge amount distribution in the specific image. It means that it is an image. Also, 02 in the figure is Dmin, which means an image in which the value of the distance between the feature vector and the template is 0 or a value close thereto, that is, an image very similar to the edge amount distribution in the specific image.
Also, the 01 part in the figure indicates an intermediate value, that is, an ambiguous image.

The recognition CPU 59 determines the presence or absence of a specific image based on the distribution state of TN (n) and D (n) developed on the work RAM 64 and a frame mask described later.
First, the frame determination processing performed using D (n) will be described. In the frame determination process, a group of a plurality of adjacent blocks is regarded as one frame, and the frame is processed while shifting its center position from the upper left of the input image horizontally and vertically in units of one block.

Next, the frame mask will be described with reference to FIG. FIGS. 18A to 18D are frame mask structure diagrams showing the structure of the frame mask. The frame mask is used to mask blocks constituting a frame, and a plurality of masks having different mask angles are prepared as shown in FIG. In FIG. 18, a hatched square indicates a mask block, a white square indicates a non-mask block, and the former is represented by a binary code of 0 and the latter is represented by a binary code of 1 in the ROM 6 of FIG.
Keep it in 1.

FIG. 19 is a flowchart showing the processing contents for one frame in the frame processing.

In FIG. 19, first, for the block at the center of the frame, the distance D (n) between the feature vector and the selected template is read and compared with the threshold Th1 (S
11) If it is larger than the threshold value Th1 (the distance is far), it is determined that there is no specific image for this frame, and the process moves to the next frame. If the distance D (n) is equal to the threshold Th1
If it is below (the distance is short), one of the frame masks is obtained from the ROM 61 (S12). The acquired frame masks are sequentially checked for each block, the following processing is skipped for the mask blocks, and D (n) of the corresponding blocks are acquired from the work RAM 64 for the non-mask blocks (S13). , S14). The acquired D (n) is sequentially added to Dsum (S15), and a counter value Bnu for counting the number of blocks subjected to processing is added.
m is incremented (S16). Step 13
Steps 16 to 16 are performed until the blocks constituting the frame are completed (S17). After the processing of all the constituent blocks is completed, the block number counter value Bnu
The average distance Dmean is obtained from m and the sum Dsum of the distances (S18), and is compared with the threshold Th2 (S1).
9). When Dmean ≦ Th2, it is determined that there is a high possibility that the specific image is included in the image. Also, Dmea
If n> Th2, it is determined that there is no specific image, and steps 12 to 19 are repeated with the frame mask changed (S
20).

When it is determined that there is a high possibility that the specific image is included in the image being processed in the above-described frame determination processing, a final determination is made. FIG. 2 shows the final determination process thereafter.
This will be described with reference to FIG.

FIGS. 20A and 20B are rotation angle correction diagrams showing the rotation angle correction in the final judgment. For the final decision,
TN (n) in which a template number closest to the feature vector of each image block is described is used. However, the image cut out by the frame determination processing may include a specific image that is not in the normal arrangement.

Step 14 of the above-described frame determination process
In, depending on the type of frame mask used at the time when it is determined that the specific image is likely to be included in the image,
At the stage of frame determination, it is possible to give an idea of the angle at which the specific image is arranged. TN based on this information
(N) is relocated to the normal position. FIG. 20 shows the situation, in which FIG. 20 (a) shows the situation before rotation correction, and FIG. 20 (b) shows the situation after rotation correction. In each of the figures, the positions of ○, □, and Δ correspond to each other.

Next, the final judgment processing will be described with reference to FIG. FIG. 21 is a recognition process relationship diagram showing the relationship between the frame, the block, and the recognition process in the final determination, and assumes that the above-described rotation correction has been performed. ○ in the figure
The marks, □, and Δ correspond to those in FIG. FIG.
In (b), the block after the normal placement is partly stepped, but in the final determination, as shown in FIG. 21, the processing is performed assuming that the specific document is placed correctly.

In FIG. 21, reference numeral 75 denotes a set of blocks in which TN (n) has a value other than 255 (= template undefined) for a frame determined to have a high possibility that a specific image is present in the frame determination. , 76 are blocks included in the block set 75, and a template is described for each block as a histogram for ease of explanation. Reference numeral 77 denotes a final determination unit formed of a neural network, reference numeral 78 denotes an input layer of the neural network, reference numeral 79 denotes an intermediate layer of the neural network, reference numeral 80 denotes an output layer of the neural network, and reference numeral 81 denotes comparison means. The comparing means 81 compares the two outputs output from the output layer 80 and selects the larger one. Note that the final determination unit 77 may be configured by either software or hardware.

Now, the blocks converted into the normal arrangement positions by the rotation correction have template numbers TN respectively.
(N). Since the template number is actually the feature vector itself included in the specific image, these can be represented by a histogram as shown in a block 76 in FIG. The frequency of this histogram is input to the input layer 78 of the final determination unit 77. The input layer 78 has three nodes for one input unit to correspond to the three-dimensional information of the histogram, and the frequency of the histogram is input to each node. The frequency is input to the nodes corresponding to all the blocks. In the neural network, the operation is performed in the intermediate layer 79 by the weighting operation previously learned, and the output layer 80
Then, the degree of the specific image and the degree of the non-specific image are output. Finally, the comparing means 81 selects and outputs the one having a higher degree of output. Therefore, the comparison means 8
The output of 1 is a binary output indicating whether or not the input image is a specific image. With the above operation, the specific image can be detected.

Now, it is determined whether or not a specific image exists in the image read by the image reading apparatus as described above. However, this type of recognition always involves an erroneous determination. In particular, in the case of a device that detects a specific image for which copying is prohibited, if a general image is recognized as a specific image, there is a problem that an image for which copying is not prohibited is not copied. Hereinafter, this solution will be described with reference to FIG.

As shown in FIG. 11, the specific image recognizing unit 37 has a memory 54, and the RGB image signal converted to a predetermined resolution by the second resolution converting unit 38 is temporarily stored in the memory 5
4 is stored. By the way, in step 19 of the frame determination processing, the type of the frame mask used when it is determined that the specific image is likely to be included in the image is used to determine the angle at which the specific image is arranged in the frame determination stage. Can be attached. In addition, at the stage of applying a frame mask by performing a shift or the like on an image, it is possible to give an estimate of the coordinate information of the specific image.
The hardware for the recognition CPU 59 is configured so as to be able to access the memory 54. If the specific image is recognized in the final determination, the recognition CPU 59 determines the corresponding image in the memory 54 based on the position and rotation angle information. Address, and for specific parts of specific images,
It is possible to obtain structural information on a printing net or the like, more detailed color information, and the like. By performing the re-determination based on these pieces of information after the detailed determination, the possibility that an erroneous determination occurs can be greatly reduced.

It is determined whether or not a specific image is included in the image read as described above.
The operation after the determination result is output will be described using FIG.

The recognition CPU 59 is connected to the serial communication line 4
The determination result is transferred to the CPU 32 using 0. CPU3
2 controls the interface 30 to notify another device 31 of the determination result. In the case of a general-purpose interface such as SCSI, the image reading device 1 sends an Abo
rt may be output to forcibly receive the result, or a command for another device 31 to receive the result may be output to the image reading device 1.

The determination operation is started when the reading operation of the image reading device 1 is completed. At the time when the determination is started, the image data read by the image reading device 1 is stored in the host computer 3 or the image recording device 2. Has already been transferred to another device 31. The other devices 31 stop processing or recording the image until receiving the determination result. As described above, the determination result of the specific image is notified from the image reading device 1, and the other devices 31 determine, for example, whether to shift to the image recording operation in accordance with the notification.

As described above, according to the present embodiment, the image dividing unit 60 that divides an input color image into blocks of a predetermined size, and the edge pixels are counted for each block divided by the image dividing unit 60. An edge amount detection circuit 55 for detecting an edge amount, and identification units 56, 57, and 59 for identifying whether or not a specific image is included in the input color image based on the edge amount detected by the edge amount detection circuit 55. With the provision, the specific image can be identified based on the detected edge amount, so that the specific image can be identified without being affected by the tint fluctuation. 3
The image sensor 20 is provided as an extraction unit for extracting two characteristic colors (RGB).
By detecting the edge amounts for the three characteristic colors extracted by 0, it becomes possible to detect the edge by the characteristic color corresponding to the specific image, and it is possible to improve the identification accuracy.

(Embodiment 2) The configuration of an image reading apparatus according to Embodiment 2 of the present invention is the same as that of Embodiment 1. This embodiment is different from the first embodiment in the processing method in the edge amount detection circuit 55 in FIG. 11, and this point will be described with reference to the drawings.

FIG. 22 is a block diagram showing an edge amount detection circuit 55 constituting the image reading apparatus according to the present embodiment.

In FIG. 22, reference numeral 56 denotes a template selection unit similar to that of FIG. 11, reference numeral 58 denotes a buffer similar to that of FIG.
2_C0, 82_C1, 82_C2 are second derivative operators, 83 is an absolute value circuit, 84 is an adder, and 85 is a buffer. Second derivative operators 82_C0, 82_C1,
82_C2 is for detecting an edge with respect to an independent predetermined characteristic color value. In the present embodiment, since the edge amount is detected for each of the three characteristic colors, it has three secondary differential operators.
In the present embodiment, an input RGB signal is divided into three characteristic colors, and Red, Green, and Blue are divided.
Colors are used. The second derivative operators 82_C0, 82_C1, and 82_C2 for each characteristic color have no structural difference, and only one color is shown in detail. The second derivative operator 82_C0 is composed of a window of 3 pixels × 3 pixels, and has a coefficient of “8” in the center pixel and “−1” in all eight pixels other than the center. An edge amount is calculated by multiplying the coefficient in the window of the second derivative operator 82_C0 by the value of the characteristic color corresponding to the position of the window, and adding the respective multiplied values. The absolute value circuit 83 calculates the absolute value of the edge amount obtained by the second derivative operator 82_C0. The adder 84 adds the data output from the absolute value circuit 83 by the number of pixels in one block. The buffer 85 accumulates the addition result of the adder 84.

The addition process is performed on a block basis. Here, the block is a block in which the read image is divided into a plurality of pixels in the main scanning direction and the sub-scanning direction. In this case, the pixels of the predetermined resolution converted by the second resolution conversion unit 38 are set to 50 pixels. A rectangle of 50 × 50 pixels is defined as one block. Therefore, the adder 84 is 5
The addition result is stored in the buffer 85 for each input of 0 pixel and reset. The buffer 85 exists for the number of blocks in the main scanning direction, and records data for one block in the sub-scanning direction. When recording from the adder 84 to the buffer 85,
The characteristic color edge amount of one block in the sub-scanning direction is accumulated by always writing the addition result to the data already written on the buffer 85, that is, by performing the read-modify-write operation. When data input for one block is completed in the sub-scanning direction, the contents of the buffer 85, that is, the result of counting the characteristic colors obtained for each block are stored in the buffer 58 and passed to the template selecting unit 56.

The above operation is performed by the second derivative operator 82_C.
The first and second derivative operators 82_C2 also perform the operations in parallel, and detect the edge amounts for the predetermined characteristic color signals, respectively.
It is stored in the buffer 58 as a dimension vector, that is, a feature vector, and is also passed to the template selection unit 56.

As described above, according to the present embodiment, the image division unit 60 that divides an input color image into blocks of a predetermined size, and the sum of edge amounts is detected for each block divided by the image division unit 60 By providing an edge amount detection circuit 55 and identification units 56, 57, and 59 for identifying whether or not a specific image is included in an input color image based on the sum of the edge amounts detected by the edge amount detection unit. Since the specific image can be identified based on the detected edge amount, the specific image can be identified without being affected by the tint fluctuation.

[0117]

As described above, according to the image reading apparatus of the first aspect of the present invention, there is provided an image reading apparatus for judging the authenticity of a document such as a banknote or a security, wherein the input color image is reproduced. An image dividing unit that divides the image into blocks of a predetermined size, an edge amount detecting unit that counts edge pixels for each block divided by the image dividing unit to detect an edge amount, and an edge amount detected by the edge amount detecting unit And an identification unit that identifies whether a specific image is included in the input color image based on the input image, the specific image can be identified based on the detected edge amount. This has the advantageous effect that the specific image can be identified without any need.

According to the image reading apparatus of the second aspect, the image reading apparatus of the first aspect further includes an extracting section for extracting a plurality of characteristic colors, and the edge amount detecting section includes:
By detecting the edge amount for a plurality of characteristic colors extracted by the extraction unit, it is possible to perform edge detection using a characteristic color corresponding to a specific image, and to obtain an advantageous effect of improving identification accuracy.

According to the image reading apparatus of the third aspect, the image reading apparatus determines the authenticity of a document such as a bill or a securities, and the image dividing unit divides an input color image into blocks of a predetermined size. And an edge amount detection unit that detects the sum of the edge amounts for each block divided by the image division unit, and whether the specific image is included in the input color image based on the sum of the edge amounts detected by the edge amount detection unit. Since the specific image can be identified based on the detected edge amount by having the identification unit for identifying whether or not the specific image is included, it is advantageous that the specific image can be identified without being affected by the tint fluctuation. Effects can be obtained.

According to the image reading device of the fourth aspect, the image reading device of the third aspect includes an extraction unit for extracting a plurality of characteristic colors, and the edge amount detection unit includes:
By detecting the edge amount for a plurality of characteristic colors extracted by the extraction unit, it is possible to perform edge detection using a characteristic color corresponding to a specific image, and to obtain an advantageous effect of improving identification accuracy.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an image copying system having an image reading device according to a first embodiment of the present invention;

FIG. 2 is a schematic sectional view showing the structure of an image reading device in the image copying system of FIG. 1;

FIG. 3 is a schematic sectional view showing an internal structure of a carriage of the image reading apparatus.

FIG. 4 is a perspective view showing details of an optical system of the image reading apparatus.

FIG. 5 is a block diagram showing each block for performing image data processing in the image reading apparatus.

FIG. 6 is a schematic view of a carriage of the image reading apparatus viewed from a side.

FIG. 7 is an array diagram of the image sensor viewed from the line sensor array side.

FIG. 8 is an operation explanatory diagram for explaining the operation principle of the line correction unit;

FIG. 9 is an operation explanatory diagram for explaining the operation of the line correction unit when reading an image at a resolution of 300 dpi in the sub-scanning direction.

FIG. 10 is an algorithm diagram showing a resolution conversion algorithm.

FIG. 11 is a block diagram showing a specific image recognition unit in detail.

FIG. 12 is a block diagram illustrating an edge amount detection unit.

FIG. 13 is a data configuration diagram showing a configuration of data stored in a buffer.

FIG. 14 is a flowchart showing the operation of a template selection unit.

15A is a relationship diagram showing a relationship between a template and a specific image. FIG. 15B is a relationship diagram showing a relationship between a template and a specific image. FIG. 15C is a relationship diagram showing a relationship between a template and a specific image. Diagram showing the relationship between the image and the specific image

FIG. 16 is a data configuration diagram showing a data configuration in a working RAM;

17A is an image diagram showing an image of TN (n) and D (n) given to each block of an actual specific image. FIG. 17B is a diagram showing TN (TN) given to each block of an actual specific image. (c) Image diagram showing images of TN (n) and D (n) given to each block of an actual specific image

18A is a frame mask structure diagram showing the structure of a frame mask. FIG. 18B is a frame mask structure diagram showing the structure of a frame mask. FIG. 18C is a frame mask structure diagram showing the structure of a frame mask.

FIG. 19 is a flowchart showing processing contents for one frame in frame processing;

20A is a rotation angle correction diagram showing a rotation angle correction in a final determination, and FIG. 20B is a rotation angle correction diagram showing a rotation angle correction in a final determination.

FIG. 21 is a recognition process relationship diagram showing a relationship between a frame, a block, and a recognition process in a final determination.

FIG. 22 is a block diagram illustrating an edge amount detection unit included in the image reading device according to the second embodiment of the present invention;

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Image reading device 2 Image recording device 3 Host computer 4 Cable 5 Image reading device main body 6 Document glass 7 Carriage 8 Drive source (stepping motor) 9 Drive pulley 10 Timing belt 11 Belt 12 Follower pulley 13 Document 14 Document cover 15 Supporting part 16 Reference acquisition position 17 Lamp 18 Aperture 19-1 and 19-2 Reflection mirror 20 Image sensor (extraction unit) 21 Imaging lens 22R, 22G, 22B Line sensor array 23R, 23G, 23B Read line 24 Amplifier / A / D converter Reference Signs List 25 shading correction unit 26 line correction unit 27 first resolution conversion unit 28 color processing unit 29 buffer 30 interface 31 other device 32 CPU 33 motor control unit 37 specific image recognition unit 38 second resolution conversion unit 39 Recognition unit 40 serial communication line 50, 51 memory area 53 pixel 54 memory 55 edge amount detection circuit 56 template selection unit 57 template storage memory 58, 85 buffer 59 recognition CPU 60 main / sub pixel counter (image division unit) 61 ROM 64 Working RAM 70_C0, 70_C1, 70_C2, 82_C0, 8
2_C1, 82_C 22nd derivative operator 71, 83 Absolute value circuit 72 Comparator 73 Counter 74 Count buffer 75 Set of blocks 76 Block 77 Final decision unit 78 Input layer 79 Intermediate layer 80 Output layer 81 Comparison means 84 Adder

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G06F 15/70 335Z 9A001 F term (Reference) 3E041 AA01 AA02 AA03 BA11 BA12 BB03 BC03 CA01 CB02 CB09 DB01 EA03 5B047 AA01 AB04 BA02 BC05 BC09 DA04 DB01 5B057 AA11 BA02 CA01 CA08 CA12 CA16 CD05 DA03 DB02 DB06 DC16 DC25 5C077 LL14 MM08 MP07 MP08 PP06 PP20 PP32 TT06 5L096 AA02 BA03 BA07 BA18 CA02 EA03 FA06 GA03 GA12 GA19 GA38 JA03 JA09 9A001

Claims (4)

[Claims] An image reading apparatus for determining the authenticity of a document such as a bill or a security, comprising: an image dividing unit for dividing an input color image into blocks of a predetermined size; and a block divided by the image dividing unit. An edge amount detection unit that detects an edge amount by counting edge pixels in units; and identifies whether a specific image is included in the input color image based on the edge amount detected by the edge amount detection unit. An image reading device comprising: an identification unit. 2. The image processing apparatus according to claim 1, further comprising an extraction unit for extracting a plurality of characteristic colors, wherein the edge amount detection unit detects an edge amount for the plurality of characteristic colors extracted by the extraction unit. 2. The image reading device according to 1. 3. An image reading apparatus for judging the authenticity of a document such as a bill or a security, comprising: an image dividing section for dividing an input color image into blocks of a predetermined size; and a block divided by the image dividing section. An edge amount detection unit that detects the sum of the edge amounts in units, and an identification unit that identifies whether a specific image is included in the input color image based on the sum of the edge amounts detected by the edge amount detection unit. An image reading device comprising: 4. An apparatus according to claim 1, further comprising an extraction unit for extracting a plurality of characteristic colors, wherein said edge amount detection unit detects an edge amount for the plurality of characteristic colors extracted by said extraction unit. 4. The image reading device according to 3.